Genomics of diffusion-imaging integrating GWAS, exome data and single-cell sequencing unravels lifespan determinants of cerebral small vessel disease

Genomics of diffusion-imaging integrating GWAS, exome data and single-cell sequencing unravels lifespan determinants of cerebral small vessel disease | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Genomics of diffusion-imaging integrating GWAS, exome data and single-cell sequencing unravels lifespan determinants of cerebral small vessel disease Stephanie Debette, Yasaman Saba, Hideaki Suzuki, Pauline Maillard, and 95 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5926137/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Peak width of skeletonized mean diffusivity (PSMD) is an emerging automated diffusion imaging marker showing clinically relevant changes in cerebral small vessel disease (cSVD), a leading cause of stroke and dementia with no mechanism-based treatment. We conducted a genome-wide association study of PSMD in 58,403 participants from 24 population-based cohorts (89% European, 10% East-Asian, 1% African-American), identifying 31 independent common variant associations. Additionally, a whole-exome sequencing analysis in 32,957 participants yielded associations of PSMD with single and burden of rare coding variants in four novel genes. Mendelian randomization supported causal association of higher blood pressure with larger PSMD values, and of larger PSMD with an increased risk of stroke, especially intracerebral hemorrhage. Strikingly, genetic susceptibility to white matter hyperintensities, an established MRI-marker of cSVD, was associated with higher PSMD from early childhood to older age, with prominent lifespan effects for VCAN and SMG6 . Leveraging unique brain single-cell sequencing resources we showed temporal changes in the cell-type specificity of these genes in the developing brain and overall enrichment of PSMD risk loci in genes expressed in fetal brain endothelial cells. Finally, through extensive integration with multi-omics resources, we provide precious leads for gene prioritization to accelerate drug discovery for cSVD. Health sciences/Diseases/Neurological disorders/Cerebrovascular disorders Biological sciences/Genetics/Genetic association study/Genome-wide association studies Health sciences/Diseases/Neurological disorders/Neurovascular disorders Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 INTRODUCTION Cerebral small vessel disease (cSVD) is a leading cause of stroke, cognitive decline, and dementia, and also a major source of postural balance, gait, and mood disturbances in older age. 1 , 2 In addition, large population-based cohort studies have shown that covert cSVD, detectable with brain imaging in the absence of positive neurological history, is extremely common in the general population with increasing age, portending a considerably increased risk of stroke, dementia, and disability. Covert cSVD could therefore represent a major target to prevent stroke and dementia and promote healthier brain aging. However, there is no mechanism-based treatment to date for cSVD. Imaging features most commonly used to define cSVD include volume of white matter hyperintensities (WMH), lacunes, microbleeds, and perivascular space burden. 3 – 5 Most of these imaging markers represent advanced stages of the disease, reflecting consequences of alterations of small vessel structure and function on the brain parenchyma. Their quantification is still in great part based on labor-intensive visual reading of brain scans or heterogeneous automated software, subject to variability and bias. 4 Diffusion tensor imaging (DTI) is the most commonly used magnetic resonance imaging (MRI) technique to study variations in white-matter microstructure in cSVD, 6 detectable long before the occurrence of the aforementioned macrostructural lesions on conventional MRI. The typical pattern of DTI in cSVD is a reduction in directionality as captured by fractional anisotropy (FA), and a higher magnitude of diffusion as captured by mean diffusivity (MD). 6 Recently, a robust and fully automated DTI measure known as the peak width of skeletonized mean diffusivity (PSMD), which is highly sensitive to change in longitudinal analyses and reproducible across different MRI scanners, was proposed as a novel marker of cSVD. 7 – 9 PSMD is the width of the distribution of MD in core white matter regions, thereby reflecting variability of MD. PSMD increases steadily with age across the adult lifespan, in contrast with MD and FA, which have a U-shaped pattern. 8 PSMD is associated with WMH volume, lacunes, total brain volume, global cognition, executive function, and processing speed. 7 – 9 Its association with processing speed was shown to be stronger than for WMH volume, lacunes, and total brain volume. 7 , 8 Therefore, PSMD might represent a useful phenotype to identify genetic susceptibility to cSVD across the lifespan. In recent years, large collaborative genome-wide association studies (GWAS), mostly conducted in older individuals aged 65 years on average, have identified over 70 genetic risk loci for cSVD. 10 , 11 Interestingly, genetic risk variants for WMH volume detected in older age already showed association with DTI metrics at age 20, in a direction compatible with variations preceding WMH occurrence. 12 Several GWAS of DTI-based indices of white-matter microstructure have also been published, mostly on UK Biobank. 13 – 17 The largest of these (N = 43,802 participants, mean age 54.2 years), identified 121 loci associated with region-specific FA and/or MD. 16 To our knowledge, no genomic association study of the summary measure PSMD, offering a more stable measure of diffusivity, that better represents microstructural damage across white matter tracts than traditional, region-specific DTI metrics, has been conducted so far. Here we performed a meta-GWAS of PSMD in 58,403 individuals from 24 population-based cohort studies (89% European [EUR], 10% East-Asian [EAS], and 1% African ancestry [AFR]), and a whole-exome association study of PSMD in 32,957 population-based participants leveraging whole-exome and whole-genome sequencing data. Our main objective was to identify novel and early molecular signatures of cSVD risk. We further examined shared genetic determinants between PSMD and other MRI-markers of cSVD, putative vascular risk factors, and common neurological diseases. Finally, by integrating our genomic findings with tissue and cell-specific transcriptomic and proteomic data, we sought to detect putative causal genes and mechanisms to be prioritized for experimental follow-up and drug discovery. RESULTS Genetic discovery through GWAS Our study population for the GWAS meta-analyses of PSMD comprised 58,403 participants from 24 population-based cohorts (mean age 58.46 ± 6.19 years, range 18–100 years, 48.20% women, Supplementary Table 1 ). We tested association of PSMD with ~ 7.5 million common single nucleotide polymorphisms (SNPs) with minor allele frequency (MAF) > 0.01 (see Methods and Supplementary Table 2 for genotyping and imputation methods). Participants were of European (N = 51,921, 89%), East-Asian (N = 5,691, 10%), and African-American (N = 521, 1%) ancestry. PSMD, which corresponds to the dispersion of MD values across white matter tracts, was calculated using the same fully automated, publicly available script in all cohorts ( Methods and Supplementary Table 3 for MRI protocols in each cohort). GWAS of PSMD were conducted using linear regression with ln(PSMD*10000) as the dependent variable to correct for skewness, adjusting for age, sex, total intracranial volume (TIV), principal components of population stratification, study site, and familial relationships when applicable. Inverse-variance-weighted (IVW) meta-analyses were conducted using METAL, within each ancestry, followed by a meta-analysis across ancestries. To identify age-specific associations we also conducted secondary GWAS stratified on age: ≤35 years (N = 3,411), 36–65 years (N = 23,782), and > 65 years (N = 24,654, Methods, Extended Data Fig. 1 ). There was no evidence for systematic inflation of association statistics at the cohort or meta-analysis level ( Supplementary Table 4 ). In the European-only GWAS, 16 independent genome-wide significant loci (P < 5x10 − 8 ) were identified across all PSMD models (full model and age-stratified analyses, Table 1 , Fig. 1 , Extended Data Fig. 2 , Supplementary Table 5 ). In sensitivity analyses without adjustment for TIV, associations were substantially unchanged ( Supplementary Table 6 ). Using GCTA-COJO to detect independent associations within individual loci by conditioning on the lead variant, 18 we identified five additional independent signals, one at chr10q24.33 ( SH3PXD2A ), one at chr6p21.33 ( MUC21 ), and three at chr6p22.1 (HLA-G , HLA-H , LOC401242, TRIM27 , Table 1 , Supplementary Table 7, Extended Data Fig. 3 ), consistent with LD clumping. In the cross-ancestry IVW GWAS meta-analyses, two additional loci were identified at chr1p31.3 ( JAK1 ) and chr12q13.13 ( ATP5G2 ) (Table 1 , Supplementary Table 8 ). In addition, multi-ancestry meta-regression (MR-MEGA) 19 was performed for loci showing heterogeneity in allelic effect across ancestries ( P-het < 0.01, Methods , Extended Data Fig. 1 ), identifying two additional loci at chr9q22.31 ( IARS ) and chr6p24.3 (low-frequency variant near TFAP2A ), leading to a total of 25 independent loci associated with PSMD ( Supplementary Table 9 ). Per-allele effect sizes showed moderate correlation (r = 0.57) between European and East-Asian participants, the two largest contributing ancestries (Fig. 2 , Supplementary Table 10 ). Despite the smaller sample size of the Japanese cohort study by Tohoku Medical Megabank Organization (ToMMo, N = 5,961 vs. 55,060 in the EUR GWAS meta-analysis), a weighted genetic risk score (wGRS) of European PSMD lead SNPs or best proxies in East-Asians ( Methods ) was significantly associated with PSMD in the ToMMo cohort (beta = 0.077 ± 0.017; P = 6.82x10 -6 ). To identify probable causal variants at genome-wide significant risk loci, we performed multiple-causal-variant fine-mapping using SuSiE 20 in Europeans ( Methods ). We identified 23 95% credible set (CS-trait) pairs, overall and across age strata, at 15 PSMD risk loci, each having a 95% posterior probability of containing a causal variant with multiple CS identified. None of these loci had a single credible SNP, but 2 loci had only 2 credible SNPs (rs12521212 and rs7728421 at chr5q23.2, and rs62434144 and rs275350 at chr6q25.1, Supplementary Table 11 ). Of note, rs275350 is a cis protein quantitative trait locus ( cis -pQTL, P = 8.56x10 − 5 ) in plasma for LRP11, 21 involved in lipid metabolism, response to extreme cold, 22 , 23 and binge eating 24 ( Supplementary Table 12 ). Furthermore, to enhance statistical power, and given a significant genetic correlation (r g =0.56, P = 7.24x10 − 30 ) between PSMD and WMH volume, we ran multitrait genome-wide association analyses using MTAG, 25 to increase statistical power by including summary statistics from GWAS of both PSMD and WMH volume. 25 We identified 6 additional genome-wide significant loci for PSMD ( Supplementary Table 13, Extended Data Fig. 2 ), at AMZ2P1 , VTA1 , FOXF2 , DEPDC1B , APOE , and RAPGEF4 , which were not associated at genome-wide significance with either PSMD or WMH volume 12 in single trait GWAS. Thus, across all methods (IVW, GCTA-COJO, MR-MEGA, and MTAG), we identified 31 loci associated with PSMD at genome-wide significance (Fig. 1 , Table 1 ). In gene-based analyses, we tested the combined association of variants within genes with PSMD in European ancestry participants ( Methods, Supplementary Table 14 ). Using MAGMA 26 we identified 33 gene-wide significant associations ( P 65 group: GPSM3/VWA7/PBX2/AGER/RNF5 , APOC1 ); using VEGAS 27 we identified 6 additional gene-wide significant associations at 5 loci ( WDR12/ICA1L, TTL, TFPI (same locus as CALCRL ), MOG , SLC54A4 ) for the full model, and one additional gene, APOE , for the age > 65 stratum (same locus as APOC1 ), all in suggestive GWAS loci (5x10 − 8 <P < 5x10 − 6 , Supplementary Table 15 ). Genetic discovery through whole-exome association study (WEAS) We systematically tested associations of rare and common exonic variants with PSMD using the whole-exome sequencing (WES) resource from UK Biobank (N = 29,938 with PSMD data), using similar association models as for GWAS; significant associations were followed up in the BRIDGET (BRain Imaging, cognition, Dementia and next generation Genomics) whole genome sequencing resource (N = 1,647, Methods , Supplementary Methods ). The single variant association analysis identified 4 loci associated with PSMD in the full model (all ages combined) in UK Biobank at P < 5x10 − 8 , including common synonymous variants at two PSMD GWAS loci ( TRIM47 and EFEMP1 ) and two novel rare risk variants, namely a synonymous variant (rs148195895, MAF = 1.22×10 − 3 ) in ST3GAL5 and a 20kb deletion (10:88910872:TATTGAAAATCCCACTAATCA:T, MAF = 2.54×10 − 4 ) in the intron of STAMBPL1 ( Supplementary Table 16 ). In participants aged > 65 years, we additionally identified a significant association of a common intronic variant in NOTCH4 at chr6p21.33, another PSMD GWAS locus (Table 1 , Supplementary Table 1 ). The common variant associations at TRIM47 and EFEMP1 showed the same direction of effect in the much smaller BRIDGET dataset and became more significant in the meta-analysis of UK Biobank and BRIDGET than in UK Biobank alone. Rare variants in ST3GAL5 and STAMBPL1 were monomorphic in BRIDGET and the meta-analysis of all single variant association statistics in UK Biobank and BRIDGET did not yield additional PSMD risk loci ( Supplementary Table 16) . Next, we performed a gene-based burden analysis, testing associations of PSMD with 310,705 variant sets representing 18,190 genes in UK Biobank, and reporting associations at gene-wide (p < 2.75×10 − 6 ) and set-wide (p < 1.31×10 − 7 ) significance thresholds. Analyses were conducted with five different variant masks and four MAF thresholds ( Methods ). We identified gene-wide significant associations with variant burden in five genes in the full model ( Supplementary Table 17A ), of which two ( KIF13B , CABYR ) also reached set-wide significance. In participants aged > 65 years, we additionally identified a gene-wide significant association with the burden of singleton variants in ITPKC ( Supplementary Table 17A , Fig. 3 ). Most gene-based burden associations were discovered for sets of singletons, rare (MAF < 0.01), and very rare (MAF < 0.001) variants. Given the smaller sample size and underrepresentation of rare and very rare alleles, we were unable to replicate these gene-based burden associations in BRIDGET ( Supplementary Table 17A ). In secondary analyses using effect-agnostic gene-based tests with SKAT and ACATV, we identified significant associations of PSMD with burden of rare variants in ST3GAL5 ( Supplementary Table 17B ), driven by the aforementioned synonymous variant ‘rs148195895’. We further tested associations of PSMD with burden of rare variants in eight familial cSVD and Alzheimer's disease (AD) genes, representing 160 variant sets. We identified set-wide significant (p < 3.13×10 − 4 ) associations of PSMD with burden of loss-of-function (LOF) and missense very rare variants (MAF < 0.001) in HTRA1 (p = 5.49×10 − 5 ), gene-wide significant (p < 6.25×10 − 3 ) associations with burden of singletons in NOTCH3 (p = 5.65×10 − 3 ) and close to gene-wide significant associations with burden of very rare variants in PSEN2 (p = 7.10×10 − 3 , Supplementary Table 17C ). Lifespan exploration of PSMD and relationship with genetically determined WMH To identify age-specific effects we conducted an age-stratified meta-analysis of GWAS and used the p-value for heterogeneity as a proxy of p-value for interaction. No significant evidence for interaction with age was observed for any of the 16 genome-wide significant risk loci in the European ancestry IVW meta-analysis (p 65 respectively ( Supplementary Table 18 ). We found strong genetic correlation between PSMD measured in the middle- (36–65) and older-age (> 65) group (r g =0.57 ± 0.11, P = 3.31x10 − 7 , Supplementary Table 18 ), while the youngest age group was too small to establish reliable genetic correlation. A wGRS of PSMD combining independent genome-wide significant variants from the European ancestry PSMD GWAS (full model) was significantly associated with PSMD already in children in an independent sample (ABCD [N = 3,769, mean age 9.96 ± 0.63 years], p = 1.78x10 − 5 , Supplementary Table 19 ). Individually, one PSMD risk locus (common variant in SMG6 ) was significantly associated with PSMD in all age groups (P = 2.21x10 − 4 in children). Next, to explore whether PSMD is also associated with genetic susceptibility to established cSVD features in different age groups throughout the lifetime, we used a wGRS of WMH volume derived from the largest European-ancestry WMH GWAS and tested its association with PSMD in independent European-ancestry participants in four age strata ( 65 years - N = 3,356, Methods ). The WMH wGRS was significantly associated with PSMD across all four age strata: effect sizes were similar across the adult lifetime, and an order of magnitude smaller in children (Fig. 4 ). Of the 25 independent lead WMH genetic risk variants, 4 were associated with PSMD at p < 0.002 (0.05/25), at VCAN and COL4A2 at age 18–35, and VCAN , CARF1 , and RASL12 at age < 18 ( Supplementary Table 20 ). Imaging and clinical correlates of PSMD We explored shared genetic variation of PSMD with 22 independent traits (i) traditional MRI-markers of cSVD or other DTI metrics, (ii) vascular risk factors, and (iii) the most common neurological diseases associated with cSVD, i.e. stroke, its subtypes, and AD, using the largest published GWAS ( Methods , Supplementary Methods, Supplementary Table 21A ). Of the 21 European-ancestry PSMD GWAS loci (lead or proxy variants with r 2 > 0.90), after accounting for the number of independent variants and traits tested (p < 1.08x10 − 4 ), 10 (48%) were associated with at least one traditional MRI-marker of cSVD (WMH [38%], PVS [24%]), and 5 (24%) with other DTI metrics (FA or MD)). Moreover, 9 loci (43%) were associated with at least one vascular risk factor, mostly blood pressure; and 4 (19%) with risk of stroke (Fig. 5 , Supplementary Table 21B ). Systematic screening for PSMD risk variants in the GWAS catalog highlighted shared genetic variants with additional complex traits, especially regional volumes and cortical phenotypes on brain MRI ( Supplementary Table 22 ). Using LD-score regression, 28 accounting for 22 traits explored (p < 7.58x10 − 4 ), we found significant genetic correlations of PSMD with WMH, WM-PVS, MD, and FA (|r g | ≥ 0.6 for WMH, MD and FA), with diastolic and systolic blood pressure (DBP, SBP), any stroke and ischemic stroke, and additionally with intracerebral hemorrhage in participants > 65 years (r g >0.6, Supplementary Table 23 , Fig. 6 , Extended Data Fig. 4 ). Finally, we used two-sample Mendelian randomization to explore the causal association of vascular risk factors with PSMD, and of PSMD with neurological diseases, using generalized summary-data-based Mendelian randomization (GSMR) 29 , and confirming significant associations (p < 7.58x10 − 4 ) with TwoSampleMR 30 ( Methods, Supplementary Table 24 ). We used large published GWAS for instruments and outcomes after removing overlapping samples (mainly through contributions of UK Biobank, Methods ). After multiple testing correction, genetically determined higher SBP and DBP were significantly associated with higher PSMD in participants aged > 65 years, while genetically predicted higher PSMD was significantly associated with increased risk of intracerebral hemorrhage ( Supplementary Table 24 ). There was no evidence for reverse causation (MR-Steiger) 31 or horizontal pleiotropy (weighted mode, MR-Egger intercept). Functional exploration of identified PSMD loci Using MAGMA 26 we identified significant enrichment of PSMD GWAS loci in the “superoxide_generating_nadph_oxidase_activator_activity” oxidative stress-related pathway (full model) and the “negative regulation of cytosolic calcium ion concentration” pathway (> 65 years, Supplementary Tables 25 ). We used VEGAS2Pathway 32 to explore pathway enrichment for genes identified the WEAS (gene-based burden results): the most significant pathway, although not significant after multiple testing correction, was related to ubiquitination (“protein_K11-linked_ubiquitination”, Supplementary Tables 26 ), previously described as a central pathway in verbal memory 33 . Next, to generate hypotheses of putative causal genes and directions of effect, we conducted transcriptome-wide association studies (TWAS) using TWAS-Fusion 34 leveraging the PSMD GWAS summary statistics (European-only) and GTEx v8 multi-tissue expression quantitative trait loci (eQTL) from brain, vascular, and blood tissue. We identified 62 transcriptome-wide expression-trait associations for PSMD that were significant in colocalization analyses (TWAS-COLOC), suggesting evidence for a shared causal variant between the corresponding gene expression and PSMD ( Supplementary Table 27 , Fig. 7 ). Overall, 34 genes showed transcriptome-wide significant associations with PSMD: 16 were in 6 genome-wide significant risk loci, while 18 were in 13 loci that did not reach genome-wide significance in the PSMD GWAS ( JAK1 , ICA1L / NBEAL1 / FAM117B / WDR12 , CALCRL, FAM107B , DPP3 , ATP5MC2 , UPF3A , FAM47E , NDUFAF2 , CASC1 , DEPDC1B , ANO1 , and AC007993.3 / MPP2 ). Furthermore, in proteome-wide association studies (PWAS) lower cis -regulated protein abundance of ICA1L and TRIM47 in the dorsolateral prefrontal cortex (DLPFC) was associated with higher PSMD, with evidence for colocalization (posterior probability of hypothesis 4 for one shared SNP [PP4] > 0.93, full model and age > 65, Supplementary Table 28, Extended Data Fig. 5 ). Applying the STEAP pipeline to several publicly available brain single-cell sequencing resources in humans and mice ( Methods, Supplementary Table 29 ), we found significant enrichment in brain vascular endothelial cells for PSMD (full model, age 36–65 and > 65), including markedly in human fetal brain vascular endothelial cells. Enrichment in immune-response-related cell types and pericytes was also observed ( Supplementary Table 30 ). Two genome-wide significant loci for PSMD and WMH, at SMG6 and VCAN , 13 both showed early significant associations with PSMD, starting in childhood ( Supplementary Tables 19–20 ), with reduced expression of SMG6 and VCAN in brain tissue and fibroblasts being associated with higher PSMD and WMH values in TWAS ( Supplementary Table 27 ). 13 We therefore explored expression patterns of these two genes in the human brain across the lifespan ( Methods ), 35 and found that both had maximal expression in the prenatal period, suggesting developmental effects ( Extended Data Fig. 6 a-b). Next, we utilized an unpublished single-nuclei RNA sequencing dataset across the human lifespan in the DLPFC to decipher cell-type specific expression patterns ( Methods ). VCAN expression showed an increasing trend in oligodendrocyte progenitor cells (OPCs) and a decreasing trend in mature oligodendrocytes across the lifecourse, with sharply decreasing expression in endothelial cells until early childhood, supporting developmental involvement, consistent with trends previously described in cSVD rat models 37 – 39 . SMG6 exhibited dominant expression in vascular and leptomeningeal cells (VLMCs) throughout the lifespan and a trend towards decreasing expression in OPCs and neurons during development and increasing expression in adulthood ( Extended Data Fig. 6 e-f ). To further substantiate these findings we used single-cell RNA sequencing data from a mouse developmental time course ( Methods ). 39 We observed that VCAN was highly expressed in oligodendrocytes across brain regions, with highest expression at P14 decreasing slowly, thereafter, consistent with patterns seen in humans. SMG6 was expressed strongly in neurons with increasing expression during postnatal development, as in humans, and weaker expression in progenitor cells, vascular cells, and oligodendrocytes ( Extended Data Fig. 6 c-d). Finally, among genes showing transcriptome-wide significant association with PSMD with colocalization evidence, we prioritized putative drug targets using the genetic priority score (GPS) browser ( Methods ). 40 We found evidence of very high GPS (> 2.1) for SMG6 and CARF with ischemic heart disease, conferring a 9.7-fold increased likelihood of having a drug indication in Open Targets ( Supplementary Table 31 for all PSMD-associated genes with a GPS > 1.5). DISCUSSION In a first, cross-ancestry GWAS of PSMD, a novel fully automated diffusion imaging metric associated with cSVD, in up to 58,403 participants from 24 population-based cohort studies, we identified 31 independent genome-wide significant risk loci, over half of which had not been identified before in GWAS for other MRI-markers of cSVD. In addition, a whole-exome association study in 29,938 participants identified associations of PSMD with rare variants in ST3GAL5 and STAMBPL1; and burden of rare LoF and singleton variants in KIF13B and CABYR . Genetically determined larger WMH volume, the most studied MRI-marker of cSVD, was strongly associated with PSMD across the lifetime, starting in childhood. Mendelian randomization supported causal associations of high blood pressure with PSMD and of PSMD with stroke, especially intracerebral hemorrhage. Using TWAS and PWAS we provided evidence for causal implication of 34 genes predominantly through genetically regulated gene expression and protein levels in vascular and brain tissue. Integration with single-cell sequencing data in humans and mice displayed significant enrichment of PSMD loci in genes expressed in fetal brain vascular endothelial cells and revealed early developmental changes in cell-type specificity for genes with strong lifespan effects ( VCAN and SMG6 ). 13 Our findings provide strong genetic evidence that PSMD is a relevant MRI-marker for cSVD, PSMD showed significant genetic correlation with WMH (r g =0.63) and to a lesser extent PVS (r g =0.26), more prominently so in participants aged > 65 years (r g =0.72 and 0.43). Several genome-wide significant PSMD loci were shared with other MRI-markers of cSVD, especially WMH (at KCNK2-CENPF, SH3PXD2A-STN1, TRIM47, NBEAL1-ICA1L, CALCRL , PLEKHG1 , LOC100505841 ) 12 , 13 , 41 , 42 . The WEAS also revealed significant associations of PSMD with the burden of very rare variants in genes causing monogenic cSVD, especially HTRA1 . 43 Furthermore, corroborating a recent observational study 44 and in line with other MRI-markers of cSVD 10 , 43% of PSMD loci were associated with blood pressure, with evidence for a putative causal association of high blood pressure with higher PSMD values. The association of higher genetically determined PSMD with increased risk of stroke, especially intracerebral hemorrhage, is consistent with known clinical complications of cSVD. An interesting feature of PSMD is that it can be measured in a standardized automated fashion at any time in life, allowing to capture subtle changes in the overall white matter microstructure prior to the occurrence of traditional cSVD markers such as WMH. Strikingly, genetically determined larger WMH volume was significantly associated with higher PSMD values across the full lifespan, 36 , 37 with effect sizes an order of magnitude smaller in children. Although the cross-sectional nature of our analysis prompts important caution, we speculate that, rather than reflecting early subtle features of cSVD, this association could perhaps correspond, at least in part, to increased vulnerability, or lesser resilience of the brain white matter to vascular insults occurring later in life. As for PVS, 11 we saw strong enrichment of PSMD risk loci in genes expressed in fetal brain vascular endothelial cells, notably in all age strata, supporting an important role of vascular developmental factors. Particularly marked lifespan associations were observed for two loci, with evidence for a causal involvement of VCAN 13 and SMG6 in TWAS, and demonstration of maximal brain expression of these prenatally. We have previously shown an association of common variants in VCAN , a robust risk locus for cSVD, 12 , 13 with lower neurite density index in young adults, specifically in regions harboring the highest frequency of WMH in older age. 45 VCAN is an extracellular matrix proteoglycan involved in development, inflammation, and remyelination, and has been suggested as a potential drug target for multiple sclerosis and possibly cSVD. 45 , 46 Observed changes in the cell-type specificity of its expression throughout the lifespan provide precious insights for the design of future experiments. SMG6 is involved in non-sense-mediated mRNA decay, 47 which regulates axonal guidance during neurodevelopment, and has been implicated in several neurodevelopmental and neurodegenerative diseases. 48 Interestingly, SMG6 was also predicted to have a very high likelihood of being a suitable drug target using GREP. Published GWAS for MRI-markers of cSVD have been conducted in populations of nearly exclusively European ancestry so far (> 95%), 11 although cSVD is even more prevalent in other ancestry groups, especially East-Asians. 49 Our study included 11% of non-European participants (10% East-Asian). While the overall correlation of effect sizes between European and East-Asian ancestry participants was moderate to good, some variants showed large discrepancies in effect size, especially those with important differences in allele frequency (> 30%, at JAK1 , LOC100505841 , TRIM47 ). Enhancing non-European contributions to cSVD GWAS will be crucial to enhance the discovery of novel risk variants and optimize transportability of genetic risk prediction and genomics-driven drug discovery. The most significant association with PSMD by far, consistent across age strata and ancestries, was observed for the chr1q41 locus near KCNK2 and CENPF , previously shown to be associated with WMH, PVS, 11 , 12 cortical thickness and surface. 50 – 52 KCNK2 encodes a voltage-gated potassium channel, involved in neuronal migration, blood-brain barrier function, 53 and mechanosensing. 54 CENPF encodes centromere protein F, a mitotic protein involved in cellular differentiation, vesicle transport, cortical neurogenesis, 55 and endothelial cell proliferation, particularly within the brain, where its upregulation in endothelial cells has been associated with ischemic stroke in mice. 56 – 58 At chr17q25, the leading WMH risk locus 12 , 59 was also associated at genome-wide significance with PSMD in the full model and older adults (> 65 years). Interestingly, higher PSMD values were associated with lower expression levels of TRIM47 in brain tissues and cultured fibroblasts (TWAS) 60 and with lower TRIM47 protein levels in the DLPFC, supporting a putative causal involvement of TRIM47 with LoF effects, in line with recent experimental evidence. 61 The present GWAS also highlights novel genome-wide significant loci not previously associated with other MRI-markers of cSVD, and TWAS and PWAS provide evidence for putative causal genes to be prioritized for functional follow-up, e.g. SMG6 , TRIM47 , and ICA1L . Additionally, several genes showing transcriptome-wide significant associations with PSMD and significant colocalization, located outside of genome-wide significant risk loci, warrant further explorations. These include for instance JAK1 (Janus Kinase 1), involved in interferon-mediated signaling, and Mendelian diseases associated with immune dysregulation (OMIM 618999), DEPDC1B (DEP Domain Containing 1B), contributing to positive regulation of the Wnt signaling pathway, itself central in brain-specific angiogenesis and blood-brain barrier integrity 62 , 63 , or ATP5MC2 (ATP Synthase Membrane Subunit C Locus 2), involved in energy metabolism and mitochondrial function. Overall, PSMD risk loci were enriched in genes involved in oxidative stress pathways, converging with recent experimental findings in TRIM47 knock-out mice. 61 Our whole exome association study identified novel associations of PSMD with single rare variants in ST3GAL5 and STAMBPL1 and burden of rare variants in KIF13B and CABYR . ST3GAL5 encodes the GM3 synthase, deficiency of which causes salt and pepper developmental regression syndrome, an autosomal recessive neurocutaneous disorder characterized by recurrent seizures and impaired brain development 64 , 65 . Mutations in STAMBPL1 cause microcephaly–capillary malformation syndrome 66 . KIF13B encodes a polarized transporter of VEGF-A receptor to the plasma membrane of endothelial cells 67 , playing a key role in VEGF-A-induced neovascularization and angiogenesis, especially under pathological conditions 68 , 69 . Specific isoforms of CABYR have been suggested to play a role in brain development 70 . We acknowledge limitations. Images were acquired on different scanners and eras with varying diffusion parameters and sensitivity, however, participating cohorts used standardized image acquisition protocols and identical PSMD quantification algorithms. Furthermore, PSMD was previously shown to have very good inter-scanner reproducibility. 71 The non-European contribution to the study sample is still limited, although substantially larger than previous genomic studies for MRI-markers of cSVD. In summary, in this first large cross-ancestry GWAS and WEAS of PSMD, an emerging, fully automated imaging marker of cSVD, we describe numerous novel genetic risk loci comprising both common and rare coding variants. Through extensive integration with multi-omics resources, including at single-cell resolution and across the lifespan, our findings offer multiple lines of evidence suggesting a lifetime process with developmental factors contributing to cSVD susceptibility. They further provide precious leads for gene prioritization towards the identification of novel therapeutic targets. Declarations Acknowledgments Detailed acknowledgments are included in the Supplementary method. We thank all the participating cohorts for contributing to this study. Author contributions L.L., H.S., V.Gudnason, P.M.M., M.M.B., A.M., and S.D. jointly supervised research. Y.S., H.Suzuki, P.M., R.M.T, M.Luciano, I.C., and N.J. contributed equally. Y.S. and S.D. designed and conceived the study. Y.S., M.D., F.B., W.B., M.G., Q.L.G, A.T, H.L., M.Sargurupremraj, M.G.D, H.H.H.A., H.J.A., K.A., N.J.A., N.R.B., M.E.B., A.S.B., D.A.B. R.R.B., G.B., H.B., S.Cichon, A.C., I.J.D., C.E., L.F., P.G., R.F.G., V.G., M.H., T.H., E.H., J.H., M.A.Ikram, M.A.I., T.I., J.J., T.K., K.K., M.J.K., A.K., Y.K., M.Lathrop, S.E.L., F.M., N.M., T.H.M., I.N.M., S.M., I.M.N., T.M.N., Y.P., J.G.D.P., J.R.R., P.S.S., C.S., M.S., K.N.S., J.S., S.Sigurdsson, A.Thalamuthu, J.N.T., A.Tsuchida, A.V., J.M.W., W.W., J.Y., Q.Y., M.Z., A.H.Z., T.W.M., K.A.M., R.D., Z.P., P.L.D.J., F.C., S.C., V.A.W., C.T., H.T., N.S., G.R., T.P., S.S., M.F, and C.D. generated the PSMD phenotype, genomic data, conducted cohort-wise GWAS analyses, and provided funding. Y.S., Q.L.G., A.M., A.C., M-G.D., I.C., R.D, M.G., T.H., H.T., N.S., M.Z., A.M., N.S., H.U.T., and G.R. contributed to statistical, bioinformatics and functional analyses. Y.S. and S.D. wrote the manuscript. All of the authors provided critical revision. Ethic declerations Competing interests H.U.T has presented at user meetings of 10xGenomics, Oxford Nanopore Technologies and Pacific Biosciences, which in some cases has involved payment for travel, accommodation and food. Other authors declared no potential conflicts of interest with respect to research, authorship, and/or publication of this article. Data Availability Summary statistics for the GWAS meta-analysis of the European and cross-ancestry are deposited in the GWAS catalog. Individual cohort data are restricted for privacy and legal reasons (national and European restrictions, including GDPR), just like other meta-analyses of GWAS or sequencing data. This is applicable to all cohorts involved (the cohorts included in the meta-analyses). Access to UKB data was conducted under Application Numbers 18545, 23509, and 94113. We used publicly available data analyses described in the manuscript, including data from GTEx (https://gtexportal.org/home/), the Gusev laboratory (http://gusevlab.org/projects/fusion/), RNA sequencing datasets: PsychENCODE DER-22 (www.ncbi.nlm.nih.gov/geo/, accession code GSE97942), GSE67835 (www.ncbi.nlm.nih.gov/geo/, accession code GSE67835), GSE101601, Allen Brain Atlas (http://portal.brain-map.org/), Mousebrain (http://mousebrain.org/), Tabula Muris (https://tabula-muris.ds.czbiohub.org/). All other data supporting the findings of this study are available within the article, the supplementary information, or the supplementary data files. Code Availability We used publicly available data from METAL (https://github.com/statgen/METAL), GCTA-cojo (https://yanglab.wetlake.edu.cn/software/gcta/#COJO), FUMA (https://fuma.ctglab.nl), MAGMA (https://ctg.cncr.nl/software/magma), LD-Score Regression (https://github.com/ldsc/), MTAG (https://github/JonJala/mtag), coloc (https://chr1swallace.github.io/coloc/) R package, TwoSampleMR (https:/mrcieu.github.io/TwoSampleMR), STEAP (https://github.com/erwinerdem/STEAP), CELLECT (https://github.com/perslab/CELLECT), S-LDSC (https://github.com/bulik/ldsc), H-MAGMA (https://github.com/thewonlab/H-MAGMA), HBT (https://hbatlas.org), and susieR (https://stephenslab.github.io/susieR/index.html) R packages. References Pasi M, Cordonnier C (2020) Clinical Relevance of Cerebral Small Vessel Diseases. Stroke 51:47–53 Godin O et al (2008) White matter lesions as a predictor of depression in the elderly: the 3C-Dijon study. Biol Psychiatry 63:663–669 Duering M et al (2023) Neuroimaging standards for research into small vessel disease-advances since 2013. Lancet Neurol 22:602–618 Wardlaw JM et al (2013) Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol 12:822–838 Wardlaw JM et al (2020) Perivascular spaces in the brain: anatomy, physiology and pathology. Nat Rev Neurol 16:137–153 Nucifora PGP, Verma R, Lee S-K, Melhem ER (2007) Diffusion-tensor MR imaging and tractography: exploring brain microstructure and connectivity. Radiology 245:367–384 Baykara E et al (2016) A Novel Imaging Marker for Small Vessel Disease Based on Skeletonization of White Matter Tracts and Diffusion Histograms. Ann Neurol 80:581–592 Beaudet G et al (2020) Age-Related Changes of Peak Width Skeletonized Mean Diffusivity (PSMD) Across the Adult Lifespan: A Multi-Cohort Study. Front Psychiatry 11:342 Zanon Zotin MC et al (2023) Peak Width of Skeletonized Mean Diffusivity: A Neuroimaging Marker for White Matter Injury. Radiology 306:e212780 Bordes C, Sargurupremraj M, Mishra A, Debette S (2022) Genetics of common cerebral small vessel disease. Nat Rev Neurol 18:84–101 Duperron M-G et al (2023) Genomics of perivascular space burden unravels early mechanisms of cerebral small vessel disease. Nat Med 29:950–962 Sargurupremraj M et al (2020) Cerebral small vessel disease genomics and its implications across the lifespan. Nat Commun 11:6285 Persyn E et al (2020) Genome-wide association study of MRI markers of cerebral small vessel disease in 42,310 participants. Nat Commun 11:2175 Smith SM et al (2021) An expanded set of genome-wide association studies of brain imaging phenotypes in UK Biobank. Nat Neurosci 24:737–745 Zhao B et al (2021) Large-scale GWAS reveals genetic architecture of brain white matter microstructure and genetic overlap with cognitive and mental health traits (n = 17,706). Mol Psychiatry 26:3943–3955 Zhao B et al (2021) Common genetic variation influencing human white matter microstructure. Science 372:eabf3736 Rutten-Jacobs LCA et al (2018) Genetic Study of White Matter Integrity in UK Biobank (N = 8448) and the Overlap With Stroke, Depression, and Dementia. Stroke 49:1340–1347 Yang J et al (2012) Conditional and joint multiple-SNP analysis of GWAS summary statistics identifies additional variants influencing complex traits. Nat Genet 44:369–375 Mägi R et al (2017) Trans-ethnic meta-regression of genome-wide association studies accounting for ancestry increases power for discovery and improves fine-mapping resolution. Hum Mol Genet 26:3639–3650 Wang G, Sarkar A, Carbonetto P, Stephens M (2020) A simple new approach to variable selection in regression, with application to genetic fine mapping. J R Stat Soc Ser B Stat Methodol 82:1273–1300 Ferkingstad E et al (2021) Large-scale integration of the plasma proteome with genetics and disease. Nat Genet 53:1712–1721 Yang X et al (2022) Transcriptome analysis provides new insights into cold adaptation of corsac fox ( Vulpes Corsac ). Ecol Evol 12:e8866 Lyu Y et al (2024) Recent selection and introgression facilitated high-altitude adaptation in cattle. Sci Bull 69:3415–3424 Burstein D et al (2023) Genome-wide analysis of a model-derived binge eating disorder phenotype identifies risk loci and implicates iron metabolism. Nat Genet 55:1462–1470 Turley P et al (2018) Multi-trait analysis of genome-wide association summary statistics using MTAG. Nat Genet 50:229–237 de Leeuw CA, Mooij JM, Heskes T, Posthuma D (2015) MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol 11:e1004219 Mishra A, Macgregor S (2015) VEGAS2: Software for More Flexible Gene-Based Testing. Twin Res Hum Genet 18:86–91 Bulik-Sullivan BK et al (2015) LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat Genet 47:291–295 Zhu Z et al (2018) Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun 9:224 Hemani G et al (2018) The MR-Base platform supports systematic causal inference across the human phenome. Elife 7:e34408 Hemani G, Tilling K (2017) Davey Smith, G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet 13:e1007081 Mishra A, MacGregor SA (2017) Novel Approach for Pathway Analysis of GWAS Data Highlights Role of BMP Signaling and Muscle Cell Differentiation in Colorectal Cancer Susceptibility. Twin Res Hum Genet 20:1–9 Debette S et al (2015) Genome-wide studies of verbal declarative memory in nondemented older people: the Cohorts for Heart and Aging Research in Genomic Epidemiology consortium. Biol Psychiatry 77:749–763 Gusev A et al (2016) Integrative approaches for large-scale transcriptome-wide association studies. Nat Genet 48:245–252 Kang HJ et al (2011) Spatio-temporal transcriptome of the human brain. Nature 478:483–489 Quick S et al (2022) Loss of the heterogeneous expression of flippase ATP11B leads to cerebral small vessel disease in a normotensive rat model. Acta Neuropathol 144:283–303 Rajani RM et al (2018) Reversal of endothelial dysfunction reduces white matter vulnerability in cerebral small vessel disease in rats. Sci Transl Med 10:eaam9507 Lopez LM et al (2015) Genes From a Translational Analysis Support a Multifactorial Nature of White Matter Hyperintensities. Stroke 46:341–347 Joglekar A et al (2024) Single-cell long-read sequencing-based mapping reveals specialized splicing patterns in developing and adult mouse and human brain. Nat Neurosci 27:1051–1063 Duffy Á et al (2024) Development of a human genetics-guided priority score for 19,365 genes and 399 drug indications. Nat Genet 56:51–59 Traylor M et al (2019) Genetic variation in PLEKHG1 is associated with white matter hyperintensities (n = 11,226). Neurology 92:e749–e757 Traylor M et al (2021) Genetic basis of lacunar stroke: a pooled analysis of individual patient data and genome-wide association studies. Lancet Neurol 20:351–361 Debette S, Markus HS (2022) Stroke Genetics: Discovery, Insight Into Mechanisms, and Clinical Perspectives. Circ Res 130:1095–1111 Petrea RE et al (2024) Hypertension Trends and White Matter Brain Injury in the Offspring Framingham Heart Study Cohort. Hypertension 81:87–95 Le Grand Q et al (2024) Diffusion imaging genomics provides novel insight into early mechanisms of cerebral small vessel disease. Mol Psychiatry. 10.1038/s41380-024-02604-7 Ghorbani S et al (2022) Versican promotes T helper 17 cytotoxic inflammation and impedes oligodendrocyte precursor cell remyelination. Nat Commun 13:2445 Eberle AB, Lykke-Andersen S, Mühlemann O, Jensen TH (2009) SMG6 promotes endonucleolytic cleavage of nonsense mRNA in human cells. Nat Struct Mol Biol 16:49–55 Petrić Howe M, Patani R (2023) Nonsense-mediated mRNA decay in neuronal physiology and neurodegeneration. Trends Neurosci 46:879–892 Mok V et al (2014) Race-ethnicity and cerebral small vessel disease–comparison between Chinese and White populations. Int J Stroke 9:36–42 van der Meer D et al (2020) Understanding the genetic determinants of the brain with MOSTest. Nat Commun 11:3512 Shadrin AA et al (2021) Vertex-wise multivariate genome-wide association study identifies 780 unique genetic loci associated with cortical morphology. NeuroImage 244:118603 Hofer E et al (2020) Genetic correlations and genome-wide associations of cortical structure in general population samples of 22,824 adults. Nat Commun 11:4796 Bando Y, Hirano T, Tagawa Y (2014) Dysfunction of KCNK potassium channels impairs neuronal migration in the developing mouse cerebral cortex. Cereb Cortex 24:1017–1029 Lengyel M, Enyedi P, Czirják G (2021) Negative Influence by the Force: Mechanically Induced Hyperpolarization via K2P Background Potassium Channels. Int J Mol Sci 22:9062 Waters AM et al (2015) The kinetochore protein, CENPF , is mutated in human ciliopathy and microcephaly phenotypes. J Med Genet 52:147–156 Mahmoud AD et al (2019) The Human-Specific and Smooth Muscle Cell-Enriched LncRNA SMILR Promotes Proliferation by Regulating Mitotic CENPF mRNA and Drives Cell-Cycle Progression Which Can Be Targeted to Limit Vascular Remodeling. Circul Res 125:535–551 Wang F et al (2022) Endothelial cell heterogeneity and microglia regulons revealed by a pig cell landscape at single-cell level. Nat Commun 13:3620 Yan Y et al (2023) Identification of brain endothelial cell-specific genes and pathways in ischemic stroke by integrated bioinformatical analysis. Brain Circulation 9:228–239 Fornage M et al (2011) Genome-wide association studies of cerebral white matter lesion burden: the CHARGE consortium. Ann Neurol 69:928–939 Mishra A et al (2022) Gene-mapping study of extremes of cerebral small vessel disease reveals TRIM47 as a strong candidate. Brain 145:1992–2007 Duplàa C et al (2024) Cerebral Small Vessel Disease genetic determinant TRIM47 controls brain homeostasis via the NRF2 antioxidant system. Preprint https://doi. org/10.1101/2024.10.08.616723 Schevenels G et al (2024) A brain-specific angiogenic mechanism enabled by tip cell specialization. Nature 628:863–871 Bats M-L et al (2022) Wnt/frizzled Signaling in Endothelium: A Major Player in Blood-Retinal- and Blood-Brain-Barrier Integrity. Cold Spring Harb Perspect Med 12:a041219 Fragaki K et al (2013) Refractory epilepsy and mitochondrial dysfunction due to GM3 synthase deficiency. Eur J Hum Genet 21:528–534 Bowser LE et al (2019) Recessive GM3 synthase deficiency: Natural history, biochemistry, and therapeutic frontier. Mol Genet Metab 126:475–488 McDonell LM et al (2013) Mutations in STAMBP, encoding a deubiquitinating enzyme, cause microcephaly-capillary malformation syndrome. Nat Genet 45:556–562 Yamada KH et al (2014) KIF13B regulates angiogenesis through Golgi to plasma membrane trafficking of VEGFR2. J Cell Sci 127:4518–4530 Waters SB et al (2021) VEGFR2 Trafficking by KIF13B Is a Novel Therapeutic Target for Wet Age-Related Macular Degeneration. Invest Ophthalmol Vis Sci 62:5 Waters SB et al (2022) KIF13B-mediated VEGFR2 trafficking is essential for vascular leakage and metastasis in vivo. Life Sci Alliance 5:e202101170 Hsu H-C et al (2005) Characterization of two non-testis-specific CABYR variants that bind to GSK3beta with a proline-rich extensin-like domain. Biochem Biophys Res Commun 329:1108–1117 Maillard P et al (2022) Instrumental validation of free water, peak-width of skeletonized mean diffusivity, and white matter hyperintensities: MarkVCID neuroimaging kits. Alzheimers Dement (Amst) 14:e12261 Methods Ethics statement This study complies with all relevant ethical regulations, and all participants gave written, informed consent. Institutional review boards that approved each contributing study are described in Supplementary Table 1 . Study design In total, 62,172 participants were included in this study, of whom 55,690 of European (EUR), 5,961 of East-Asian (EAS), and 521 of African-American (AFR) ancestry. Of these 4,269 participants aged 18 years and older (3,411 EUR, 858 EAS) were included in the GWAS meta-analysis, while data from 3,769 children was used to explore lifespan associations with PSMD. Individuals included in this project participated in 25 cohorts (Extended Data Fig. 1 ) from the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortium, 72 UK Biobank (UKB), the Brain Imaging, cognition, Dementia, and Next generation GEnomics (BRIDGET) consortium, and the Enhancing Neuroimaging Genetics through meta-analysis (ENIGMA) consortium. 73 Characteristics of study participants for each cohort are provided in Supplementary Tables 1-3 . All participants gave written informed consent, and institutional review boards approved individual studies ( Supplementary Methods , Supplementary Table 1 ). Peak width of Skeletonized Mean Diffusivity (PSMD) Various metrics were derived from the DTI data using a script developed by Baykara et al (http://www.psmd-marker.com), 7 and adapted by Beaudet et al. 8 The fully automated calculation of PSMD, corresponding to the dispersion of MD values across white matter tracts, was conducted in two steps: 1) skeletonization of the DTI data using the FA map and 2) histogram analysis of the MD within the white matter skeleton after application of a custom mask as described in Baykara et al. 7 PSMD was computed as the difference between the 95 th and the 5 th percentiles of the skeletonized and masked MD volume voxel value distribution. 7,8 MRI scans were acquired using scanners with field strengths ranging from 1.5 to 3.0 Tesla. MRI protocols of each cohort can be found in the Supplementary Methods and Supplementary Table 2 . Due to a skewed distribution of PSMD in several studies, we used the natural logarithmic transformation of the variable with the following formula ln(PSMD*10000). Genotyping and Imputation Information on genotyping platforms, quality control (QC), and imputation methods for each participating cohort are provided in Supplementary Table 3 . Each cohort conducted genotyping using different commercially available genotyping platforms. Genotyped data were mostly imputed to the Haplotype Reference Consortium (HRC) reference panel, while a few studies used the 1000 Genome reference panel (Phase I version 3), in particular African ancestry samples. Genome-wide association analyses Each cohort performed a linear regression model to determine association between ln(PSMD*10000) and allele dosages of SNPs using additive genetic effects adjusted for age, sex, TIV, and principal components of population stratification, study site, and familial relationships, where relevant. Furthermore, each cohort performed secondary association analyses in three age strata: 18-35, 36-65, and >65. QC of summary statistics shared by each cohort was conducted following recommendations of Winkler et al. 74 using EasyQC R package ( Supplementary Methods ). Rare variants with a minor allele frequency (MAF) < 0.01, with low imputation quality (R 2 < 0.5), or a product of MAF, R 2 , and sample size less than 10 were excluded. First, a fixed-effects Inverse-Variance Weighted (IVW) meta-analysis was performed using METAL 75 within each ancestry group (EUR, EAS, and AFR) followed by a meta-analysis across ancestries. Genomic control was applied to each study-specific GWAS with a genomic inflation factor greater than 1 (λ GC > 1.00) to correct for any residual population stratification. After meta-analysis, only SNPs represented in more than four of the participating studies and/or more than one-third of the sample size and with no evidence of between-study heterogeneity (P Het > 1x10 -4 ) were considered. Quantile-Quantile (Q-Q) plots, Manhattan plots, and genomic inflation factors for all meta-analyses are presented in Extended Data Fig. 2 and Supplementary Table 4 . Across all meta-analysis models, we considered variants reaching a P-value < 5x10 -8 as genome-wide significant. To identify independent SNPs within genome-wide risk loci, we used the PLINK clumping function, considering linkage disequilibrium (LD) with both an LD threshold of r 2 > 0.1 and a physical distance of ± 1 Mb from the index SNPs of a given locus. In order to identify variants that were independently associated with PSMD within 1 MB of lead SNPs, the step-wise conditional regression and joint analysis (GCTA-COJO) 18 was performed using Genome-wide Complex Trait Analysis (GCTA, version 1.26.0). 76 LD patterns were selected based on HRC imputed data of 1,862 participants from the i-Share study ( Supplementary Methods ). Next, fixed-effects IVW meta-analysis was performed to conduct cross-ancestry GWAS of PSMD using the METAL software. We used the Cochran Q test as implemented in METAL to test for heterogeneity of effects between ancestries. Furthermore, we conducted a multi-ancestry meta-analysis using the MR-MEGA software 19 , which uses meta-regression to model allelic effects including axes of genetic variation as covariates in the model. Exploration of age-specific genetic associations and heritability estimates The SNP-based heritability estimates were calculated using LD score regression (LDSC package https://github.com/bulik/ldsc/) 28 and the European LD-score files calculated from the 1000G reference panel provided by the developers, overall and in each age stratum. Gene-based analysis We performed gene-based analyses on European PSMD GWAS meta-analyses, using the Multi-marker Analysis of Genomic Annotation (MAGMA) 26 software implemented in FUMA 77 (19,090 protein coding genes) and VEGAS2 27 software (18,432 protein coding, autosomal genes). Both methods considered variants in the gene or within 10 kb on either side of a gene’s transcription site to compute a gene-based P-value. We performed MAGMA gene-based tests using the default parameters, whereas the VEGAS2 analyses were performed using the “—top 10” parameter that tests enrichment of the top 10% variants assigned to a gene accounting for the linkage disequilibrium between variants and the total number of variants within a gene. Gene-wide significance was defined at P-value < 2.62x10 -6 for gene-based tests using MAGMA and P-value < 2.71x10 -6 for VEGAS2. Genes were considered in the same locus if they were < 200 kb apart from each other. Multi-Trait Analysis of GWAS with white matter hyperintensity volume We applied Multi-Trait Analysis of GWAS (MTAG), 25 performing a joint analysis of summary statistics from the EUR PSMD GWAS with the largest EUR WMH GWAS (N=48,454), 12 to uncover additional genetic risk loci for PSMD, as PSMD and WMH were strongly correlated. MTAG estimates per SNP effect size for each trait by incorporating information contained in other correlated traits. The effect size re-estimated by MTAG is a generalized estimate of IVW meta-analysis by integrating GWAS summary statistics from different traits, where the P-value is derived from the re-estimated effect size. 25 We prioritized associations fulfilling the following conditions: (1) MTAG P-value for PSMD < 5x10 -8 , (2) univariate GWAS P-value for PSMD < 0.05; and (3) MTAG P-value for PSMD was lower the univariate GWAS P-value for WMH. PSMD whole exome association study We performed a whole exome association study (WEAS) to identify (rare) exonic variants associated with PSMD. We conducted discovery analyses on processed population level OQFE (an update Functional Equivalence (FE) protocol) UK biobank whole exome sequencing (WES) data 78 with PSMD information (N=29,938) accessible through the UKB-RAP platform. The REGENIE software 79 (v3.1.1) was used to perform single variant association tests, gene-based burden test, gene-based Sequencing Kernel Association Test (SKAT), and Aggregated Cauchy Association Test with Variance component (ACATV) tests. The REGENIE software employs a two-step approach – step 1 is applied on a small set of directly genotyped variants that captures a good fraction of the phenotype variance, and step 2 is applied on full WES data testing different association models (e.g. single variant, gene-based, etc.). Prior to REGENIE Step 1, we performed quality control (QC) the directly genotyped array data removing variants with minor allele frequency (MAF) <0.01, minor allele count (MAC) <20, Hardy Weinberg equilibrium (HWE) p-value 10%, and samples with >10% missingness. Prior to running REGENIE Step 2, we performed QC of WES data removing variants with HWE p-value10%. For single variant association testing, we also removed variants with minor allele count (MAC) <10. We used the variant annotation provided in the helper files of population level OQFE UKB WES data generated using SnpEff software. 80 Using this annotation we created the following variant masks for gene-based burden tests: M1: Loss of function (LoF) variants only, M2: LoF and missense variants predicted to be deleterious by five predictor software (LRT, 81 PolyPhen-2 HDIV, PolyPhen-2 HVAR, 82 SIFT, 83 and MutationTaster 84 ), M3: LoF and missense variants predicted to be deleterious by at least one of the five predictor software, M4: LoF and all missense variants, M5: LoF, all missense, and all synonymous variants. Additionally, we considered four strata based on maximum MAF of variants namely: singletons, MAF<0.001, MAF<0.01, and MAF<0.05. Variant sets with MAC<5 were removed from gene-based burden tests. Overall 310,705 variant sets of 18,190 genes were tested for gene-based burden analyses, leading to a gene-wide and set-wide significance threshold of p<2.75×10 -6 and p<1.31×10 -7 respectively. We also performed secondary effect-agnostic gene-based tests using SKAT 85 and ACATV. 86 For this we considered the default strategy in REGENIE considering all variants in each mask category irrespective of their allele frequency. REGENIE collapses ultra-rare variants (MAC≤10) into a burden mask to include them in these tests. To avoid reporting ACATV/SKAT associations that are driven by the burden masks of only one or two singleton variants (burden set MAC0.1. Correcting for 62,355 variant sets considering all variants in respective masks without allele frequency stratification the significant threshold was set at p<8.02×10 -7 . Discovered associations were followed-up in independent BRIDGET (BRain Imaging, cognition, Dementia and next generation Genomics) whole genome sequencing data (N=1,647). Details on the BRIDGET WGS data processing and QC is described in the Supplementary Methods . BRIDGET data was also analyzed using the REGENIE software, employing similar QC and association parameters as described for UK biobank. The BRIDGET data was annotated using the SnpEff software and the same five variant masks (M1-M5) were created as for gene-based association tests in UK Biobank. Due to small sample size, the gene-based burden test was performed only for variant set with maximum MAF<0.05. Association of genetically determined WMH with PSMD across the lifespan We explored the association of genetically determined weighted genetic risk score (wGRS) WMH with PSMD across the adult age strata and in younger age stratum in a non-overlapping sample between WMH, and PSMD (18-35, N=3,265; 36-65, N=4,159; >65, N=3,356). Associations were tested using linear mixed models adjusted for age, sex, total intracranial volume, and the first four principal components of populations stratifications ( Supplementary Methods ). For associations with individual SNPS the significant threshold was set at P<2x10 -3 (0.05/25). The aggregate effect of 25 WMH risk variants with DTI metrics was estimated by using the “gtx” package in R. In secondary analyses, we searched for an association of PSMD genome-wide significant loci identified in the GWAS with PSMD in 3,769 children aged 9.9±0.63 years, participating in ABCD study, using the same approach as described above. Association of genetically determined PSMD with PSMD across ancestries We tested the association of a European PSMD weighted genetic risk score (wGRS) with PSMD in the Japanese cohort using the R-package “gtx” (http://cran.r-project.org/web/packages/gtx/). We used a LD reference panel including 7,062 unrelated individuals from the Nagahama study. 87 Among the 16 PSMD lead SNPs, 5 were not present in the Nagahama reference panel, so we used as tag SNPs the SNP in LD r 2 >0.7, window size = 1 Mb (European HRC-imputed 3C-Dijon study or 1000Gp3 reference panel) which was present in the Nagahama reference panel and had the most significant p-value in the PSMD European meta-analysis. Two SNPs (chr6:31807540 and chr6:1365883) were rare variants in EAS. The 14 SNPs included in the GRS were independent using Nagahama reference panel (r 2 0.7, window size = 1 Mb, using Nagahama reference panel) in the Tohoku Megabank cohort which was present in the European PSMD meta-analysis. SNPs were weighted by the SNP effect sizes in the European GWAS meta-analysis. Shared genetic variation of PSMD with related vascular, neurological, and psychiatric traits We first assessed (in European ancestry participants) whether PSMD risk loci were associated with: (i) putative risk factors (SBP, 88 DBP, 88 pulse pressure (PP), 88 body mass index (BMI), 89 high density lipoprotein (HDL) cholesterol, 90 low density lipoprotein (LDL) cholesterol, 90 triglycerides, 90 type 2 diabetes, 91 (ii) other MRI-markers of brain aging (WMH burden, 12 FA, 13 MD, 13 PVS in WM, 11 BG, 11 and HIP 11 ); (iii) stroke (any stroke, 92 any ischemic stroke, 92 large artery stroke, 92 cardio-embolic stroke, 92 small vessel stroke, 92 intracerebral hemorrhage [ICH] 93 ), and AD. 94 P-value < 3.98x10 -5 correcting for 22 independent phenotypes, the 3 PSMD GWAS models, and 19 independent genome-wide significant PSMD risk loci (EUR IVW analysis ( Supplementary Methods ). We then used LD-score regression (LDSC package https://github.com/bulik/ldsc) 28 to assess the genetic correlation between PSMD and the aforementioned traits, using summary statistics from the largest publicly available GWASs. 11–13,88–93,95 To decrease the potential bias due to poor imputation quality, the summary statistics were filtered to the subset of HapMap3 SNPs for each trait. A P-value ≤ 7.58x10 -4 correcting for 22 phenotypes and for the 3 tested models (overall and by age strata) was considered significant. Additionally, we used the Functional Mapping and Annotation of Genome-wide Association studies (FUMA) 77 to obtain extensive functional annotation for genome-wide significant SNPs, and to identify SNPs associated with any other trait from the GWAS catalog at genome-wide significant level. 96 Mendelian randomization Mendelian randomization (MR) was used to seek evidence for a causal relation of putative vascular risk factors (SBP, DBP, PP, BMI, LDL- and HDL-cholesterol, triglyceride, type 2 diabetes) with PSMD, and of PSMD with the most common neurological diseases associated cSVD, stroke, and AD. We used the following two-sample MR:approaches: the Generalized Summary-data-based Mendelian Randomization (MR) method implemented in GCTA (version 1.93.2beta) software (GCTA-GSMR), 29 and the TwoSampleMR software. 30 To build our instruments for MR, we used genetic risk variants for the aforementioned traits (exposures). Only independent SNPs (LD-r 2 <0.05 for GSMR, and LD-r 2 <0.01 reaching genome-wide significance (P-value < 5x10 -8 ) were included. 97 To select these SNPs, we clumped the summary statistics of the vascular risk factors (window: 1000kb, r 2 <0.01) after filtering the SNPs, excluding variants with MAF < 0.01, ambiguous alleles, and non-matching alleles between the exposure and PSMD summary statistics and variants with an average imputation score < 0.9 in the PSMD GWAS. In GSMR, we removed SNPs that have pleiotropic effects on both exposure and outcome by using the heterogeneity in independent instrument (HEIDI)-outlier method (pHEIDI < 0.01) and ran GSMR, based on a two-step least square approach, to estimate the effects of the exposures on the outcomes. 29 In TwosampleMR, 30 we harmonized data between exposure and outcome using the default parameters. F-statistic, which should be over 10 to avoid weak instrument bias, was calculated to confirm the attainment of the relevance assumption. We used IVW for the primary analyses, and sensitivity analyses including weighted median, and MR-Egger 98 methods to confirm the observed effects. We applied MR-Egger intercept to assess the horizontal pleiotropy (P≥0.05), In addition, we confirmed the directionality of observed associations with Steiger test 31 ( Supplementary Methods ). A P-value ≤ 7.58x10 -4 correcting for 22 independent phenotypes and 3 tested models was considered significant. Pathway analyses We conducted pathway analyses with MAGMA gene set analyses 26 implemented in FUMA, on EUR PSMD GWAS summary statistics, using the 1000G phase3 reference panel. 10,678 gene sets (curated gene sets: 4,761, Go terms: 5,917) from MsigdB v6.2 were used, and a P-value < 3.8x10 -5 correcting for 1,320 independent gene sets was considered significant. We also used the VEGAS2Pathway approach, which aggregates association strengths of individual variants into pre-specified biological pathways using VEGAS-derived gene association P-values for PSMD. The empirical significance threshold for VEGAS2Pathway was 1x10 -5 accounting for 6,213 correlated pathways. 32 Transcriptome-wide association study We performed transcriptome-wide association studies (TWAS) of PSMD using TWAS-Fusion 34 leveraging the association statistics from the EUR PSMD GWAS (Full model, Age 36-65, Age > 65) and precomputed gene expression weights from GTEx v8 gene expression reference panels. 99 TWAS Z score (association statistic between predicted expression and PSMD) was derived from the integration of expression reference panels (SNP-expression weights), GWAS summary statistics (SNP-PSMD effect estimates), and linkage disequilibrium reference panels (SNP correlation matrix). 34 Transcriptome-wide significance genes (eGenes) and corresponding eQTLs were determined using Bonferroni correction based on the average number of features (6140.5 genes) tested across tissues considering all three independent models tested (p<2.7 x 10 -6 ). Identified genes were then tested in conditional analysis as implemented in Fusion software. Colocalization analysis was then conducted on the conditionally significant genes (p<0.05) using COLOC 100 to estimate the posterior probability of a shared causal variant between gene expression and PSMD (PP4≥0.70). eGene regions with eQTL not in LD (r 2 <0.01) with the lead SNP for genome-wide significant PSMD risk loci were considered as novel. ( Supplementary Methods ) . Cell type enrichment analysis We conducted a cell-type enrichment analysis using S ingle cell T ype E nrichment A nalysis for P henotypes (STEAP; https://github.com/erwinerdem/STEAP/). This is an extension to CELLECT and uses S-LDSC, 101 MAGMA, 26 , and H-MAGMA 102 for enrichment analysis (Supplementary Methods) . PSMD GWAS summary statistics were first munged. Then, expression specificity profiles were calculated using human and mouse single cell RNA-seq databases ( Supplementary Table 27 ). Cell-type enrichment was calculated with three models: MAGMA, H-MAGMA (incorporating chromatin interaction profiles from human brain tissues in MAGMA), and stratified LD score regression. P-values were corrected for the number of independent cell types in each database (Bonferroni correction). Lifetime brain gene expression profile To look for developmental processes, we examined the lifetime expression of genes in loci reaching genome-wide significance with PSMD Phenotype in the children (ABCD) study and the younger age stratum 18-35, as well as WMH genome-wide significant genes which significantly associate with PSMD. We used a public database (https://hbatlas.org/) 35 that contained genome-wide exon-level transcriptome data from 1,340 tissue samples from 16 brain regions of 57 postmortem human brains, spanning from embryonic development to late adulthood. Next, we used single-cell RNA sequencing data from a mouse developmental time course (postnatal days 14, 21, 28, and 56) for visual cortex and hippocampus as well as further P56 data from cerebellum, thalamus, and striatum 39 and calculated the average gene expression and percentage of cells that express SMG6 and VCAN using the Seurat software. 103 Furthermore, we analyzed unpublished single-nuclei RNA sequencing data from the dorsolateral prefrontal cortex (DLPFC) across the human lifespan, covering prenatal and postnatal development as well as aging, to investigate the temporal dynamic gene expression pattern. Raw counts for each cell type were aggregated across 137 libraries (from 114 donors) as pseudo-bulked counts and scaled to log2(CPM+1). The expression dynamics were smoothed by loess function with a span of 0.8, and a 95% confidence interval, implemented in R version 4.3.2. 104 Prioritizing drug targets for enhanced drug discovery To find evidence of drug target genes, we searched for identified TWAS genes associated with PSMD using browser of Genetic Priority Score for Therapeutic Targets (https://rstudio-connect.hpc.mssm.edu/geneticpriorityscore/). Genes with a high negative GPS-DOE would be suitable targets for inhibitors, whereas genes with a high positive GPS-DOE would be more suitable for drug activation. 40 References (METHODS) 72. Psaty, B. M. et al. Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium: Design of prospective meta-analyses of genome-wide association studies from 5 cohorts. Circ Cardiovasc Genet 2 , 73–80 (2009). 73. Bearden, C. E. & Thompson, P. M. Emerging Global Initiatives in Neurogenetics: The Enhancing Neuroimaging Genetics through Meta-analysis (ENIGMA) Consortium. Neuron 94 , 232–236 (2017). 74. Winkler, T. W. et al. Quality control and conduct of genome-wide association meta-analyses. Nat Protoc 9 , 1192–1212 (2014). 75. Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26 , 2190–2191 (2010). 76. Yang, J., Lee, S. H., Goddard, M. E. & Visscher, P. M. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet 88 , 76–82 (2011). 77. Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. Nat Commun 8 , 1826 (2017). 78. Krasheninina, O. et al. Open-source mapping and variant calling for large-scale NGS data from original base-quality scores. Preprint at https://doi.org/10.1101/2020.12.15.356360 (2020). 79. Mbatchou, J. et al. Computationally efficient whole-genome regression for quantitative and binary traits. Nat Genet 53 , 1097–1103 (2021). 80. Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w 1118 ; iso-2; iso-3. Fly 6 , 80–92 (2012). 81. Chun, S. & Fay, J. C. Identification of deleterious mutations within three human genomes. Genome Res 19 , 1553–1561 (2009). 82. Adzhubei, I., Jordan, D. M. & Sunyaev, S. R. Predicting functional effect of human missense mutations using PolyPhen-2. Curr Protoc Hum Genet Chapter 7 , Unit7.20 (2013). 83. Vaser, R., Adusumalli, S., Leng, S. N., Sikic, M. & Ng, P. C. SIFT missense predictions for genomes. Nat Protoc 11 , 1–9 (2016). 84. Schwarz, J. M., Rödelsperger, C., Schuelke, M. & Seelow, D. MutationTaster evaluates disease-causing potential of sequence alterations. Nat Methods 7 , 575–576 (2010). 85. Ionita-Laza, I., Lee, S., Makarov, V., Buxbaum, J. D. & Lin, X. Sequence kernel association tests for the combined effect of rare and common variants. Am J Hum Genet 92 , 841–853 (2013). 86. Liu, Y. et al. ACAT: A Fast and Powerful p Value Combination Method for Rare-Variant Analysis in Sequencing Studies. Am J Hum Genet 104 , 410–421 (2019). 87. Nakata, I. et al. Prevalence and characteristics of age-related macular degeneration in the Japanese population: the Nagahama study. Am J Ophthalmol 156 , 1002-1009.e2 (2013). 88. Evangelou, E. et al. Genetic analysis of over 1 million people identifies 535 new loci associated with blood pressure traits. Nat Genet 50 , 1412–1425 (2018). 89. Yengo, L. et al. Meta-analysis of genome-wide association studies for height and body mass index in ∼700000 individuals of European ancestry. Hum Mol Genet 27 , 3641–3649 (2018). 90. Willer, C. J. et al. Discovery and refinement of loci associated with lipid levels. Nat Genet 45 , 1274–1283 (2013). 91. Xue, A. et al. Genome-wide association analyses identify 143 risk variants and putative regulatory mechanisms for type 2 diabetes. Nat Commun 9 , 2941 (2018). 92. Mishra, A. et al. Stroke genetics informs drug discovery and risk prediction across ancestries. Nature 611 , 115–123 (2022). 93. Woo, D. et al. Meta-analysis of genome-wide association studies identifies 1q22 as a susceptibility locus for intracerebral hemorrhage. Am J Hum Genet 94 , 511–521 (2014). 94. Alzheimer Disease Genetics Consortium (ADGC), et al. Genetic meta-analysis of diagnosed Alzheimer’s disease identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. Nat Genet 51 , 414–430 (2019). 95. Schwartzentruber, J. et al. Genome-wide meta-analysis, fine-mapping and integrative prioritization implicate new Alzheimer’s disease risk genes. Nat Genet 53 , 392–402 (2021). 96. Buniello, A. et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res 47 , D1005–D1012 (2019). 97. Swerdlow, D. I. et al. Selecting instruments for Mendelian randomization in the wake of genome-wide association studies. Int J Epidemiol 45 , 1600–1616 (2016). 98. Bowden, J., Davey Smith, G. & Burgess, S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 44 , 512–525 (2015). 99. Carithers, L. J. & Moore, H. M. The Genotype-Tissue Expression (GTEx) Project. Biopreserv Biobank 13 , 307–308 (2015). 100. Giambartolomei, C. et al. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet 10 , e1004383 (2014). 101. Finucane, H. K. et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat Genet 47 , 1228–1235 (2015). 102. Sey, N. Y. A. et al. A computational tool (H-MAGMA) for improved prediction of brain-disorder risk genes by incorporating brain chromatin interaction profiles. Nat Neurosci 23 , 583–593 (2020). 103. Hao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat Biotechnol 42 , 293–304 (2024). 104. Li, M. et al. Integrative functional genomic analysis of human brain development and neuropsychiatric risks. Science 362 , eaat7615 (2018). Tables Table 1 is available in the Supplementary Files section. Additional Declarations There is NO Competing Interest. Supplementary Files Table1.docx PSMDGenomicsSupplementaryAppendixSubmit.docx Supplementary Appendix PSMDGenomicsSupplementaryTablesSubmit.xlsx Supplementary Tables EXTENDEDDATA.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5926137","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":414419426,"identity":"6b2eaac1-54be-4be5-958d-58fa86e2cd9f","order_by":0,"name":"Stephanie Debette","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0001-8675-7968","institution":"University of Bordeaux","correspondingAuthor":true,"prefix":"","firstName":"Stephanie","middleName":"","lastName":"Debette","suffix":""},{"id":414419427,"identity":"bb00b623-6523-477c-9be4-54a19cf2c6e4","order_by":1,"name":"Yasaman Saba","email":"","orcid":"","institution":"University of Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Yasaman","middleName":"","lastName":"Saba","suffix":""},{"id":414419428,"identity":"68693333-13b7-4bb0-aec7-e227f9f035e5","order_by":2,"name":"Hideaki Suzuki","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Hideaki","middleName":"","lastName":"Suzuki","suffix":""},{"id":414419429,"identity":"a62f3a58-9eaa-4c87-b9ae-27325764ebd4","order_by":3,"name":"Pauline Maillard","email":"","orcid":"","institution":"University of California Davis","correspondingAuthor":false,"prefix":"","firstName":"Pauline","middleName":"","lastName":"Maillard","suffix":""},{"id":414419430,"identity":"ba0afc05-fd4f-4a04-b2ba-b2aa3f536e4e","order_by":4,"name":"Rick Tankard","email":"","orcid":"","institution":"Murdoch University","correspondingAuthor":false,"prefix":"","firstName":"Rick","middleName":"","lastName":"Tankard","suffix":""},{"id":414419431,"identity":"d3724bb7-86a4-4576-a412-003c348aeaf9","order_by":5,"name":"Michelle Luciano","email":"","orcid":"https://orcid.org/0000-0002-7306-3008","institution":"The University of Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Michelle","middleName":"","lastName":"Luciano","suffix":""},{"id":414419432,"identity":"ac197be9-3c90-4484-bb4f-43ec12d50693","order_by":6,"name":"Ilana Caro","email":"","orcid":"https://orcid.org/0000-0003-3075-4682","institution":"University of Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Ilana","middleName":"","lastName":"Caro","suffix":""},{"id":414419433,"identity":"4b7505cb-b6e0-4256-aad0-94221dbaae24","order_by":7,"name":"Neda Jahanshad","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Neda","middleName":"","lastName":"Jahanshad","suffix":""},{"id":414419434,"identity":"7b3ef9ae-be50-4a6a-861a-23f6511430bb","order_by":8,"name":"Marco Duering","email":"","orcid":"","institution":"Medical Image Analysis Center (MIAC) and Department of Biomedical Engineering, University of Basel","correspondingAuthor":false,"prefix":"","firstName":"Marco","middleName":"","lastName":"Duering","suffix":""},{"id":414419435,"identity":"7b9c4978-6f0f-4258-aa18-8916441d3abd","order_by":9,"name":"Frauke Beyer","email":"","orcid":"https://orcid.org/0000-0001-5401-852X","institution":"Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig","correspondingAuthor":false,"prefix":"","firstName":"Frauke","middleName":"","lastName":"Beyer","suffix":""},{"id":414419436,"identity":"8e6830d9-7161-4fd0-82e4-179e36da058a","order_by":10,"name":"Wenjia Bai","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Wenjia","middleName":"","lastName":"Bai","suffix":""},{"id":414419437,"identity":"acaa8a42-8cc9-4ef6-81fd-bfd7b538b481","order_by":11,"name":"Martina Glavan","email":"","orcid":"","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"Martina","middleName":"","lastName":"Glavan","suffix":""},{"id":414419438,"identity":"a9b7a5b5-385a-4a62-89f5-163ba82b80b5","order_by":12,"name":"Quentin Le Grand","email":"","orcid":"https://orcid.org/0000-0002-9299-0747","institution":"University of Bordeaux, Inserm, Bordeaux Population Health Research Center, UMR 1219","correspondingAuthor":false,"prefix":"","firstName":"Quentin","middleName":"Le","lastName":"Grand","suffix":""},{"id":414419439,"identity":"363966a3-b79d-4c4b-8b25-893cba15057b","order_by":13,"name":"Alexander Teumer","email":"","orcid":"","institution":"University Medicine Greifswald","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"","lastName":"Teumer","suffix":""},{"id":414419440,"identity":"adc9b640-56fc-4c73-a7c3-68f8a3ef27ab","order_by":14,"name":"Hervé Lemaitre","email":"","orcid":"","institution":"University of Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Hervé","middleName":"","lastName":"Lemaitre","suffix":""},{"id":414419441,"identity":"94ebe600-5141-4d57-8278-ce3fa5941c1e","order_by":15,"name":"Muralidharan Sargurupremraj","email":"","orcid":"","institution":"Glenn Biggs Institute for Alzheimer’s and Neurodegenerative Diseases, University of Texas Health Sciences Center","correspondingAuthor":false,"prefix":"","firstName":"Muralidharan","middleName":"","lastName":"Sargurupremraj","suffix":""},{"id":414419442,"identity":"67ada72b-3c2d-4590-b46c-34234cb637e5","order_by":16,"name":"Marie-Gabrielle Duperron","email":"","orcid":"","institution":"University of Bordeaux, Bordeaux population health research center; Bordeaux Hospital","correspondingAuthor":false,"prefix":"","firstName":"Marie-Gabrielle","middleName":"","lastName":"Duperron","suffix":""},{"id":414419443,"identity":"ce7c6d45-ec50-4a73-9309-4a7e65ab3caa","order_by":17,"name":"Hieab Adams","email":"","orcid":"","institution":"Radboud University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hieab","middleName":"","lastName":"Adams","suffix":""},{"id":414419444,"identity":"0ffe850c-ef73-4215-a38c-f8b6841abe1a","order_by":18,"name":"Hugo Aparicio","email":"","orcid":"","institution":"Boston University","correspondingAuthor":false,"prefix":"","firstName":"Hugo","middleName":"","lastName":"Aparicio","suffix":""},{"id":414419445,"identity":"055dbb13-987b-458a-a57d-58f55015ccdd","order_by":19,"name":"Konstantinos Arfanakis","email":"","orcid":"","institution":"Rush University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Konstantinos","middleName":"","lastName":"Arfanakis","suffix":""},{"id":414419446,"identity":"83678625-35d5-4980-b4ce-7c3f97e40925","order_by":20,"name":"Nicola J. Armstrong","email":"","orcid":"https://orcid.org/0000-0002-4477-293X","institution":"Curtin University","correspondingAuthor":false,"prefix":"","firstName":"Nicola","middleName":"J.","lastName":"Armstrong","suffix":""},{"id":414419447,"identity":"99889ad9-07eb-41f8-8f53-3da542aa5ff6","order_by":21,"name":"Nick Bryan","email":"","orcid":"","institution":"University of Texas at Austin","correspondingAuthor":false,"prefix":"","firstName":"Nick","middleName":"","lastName":"Bryan","suffix":""},{"id":414419448,"identity":"d4656f0e-1a3e-41d9-8a53-ed4156647d04","order_by":22,"name":"Mark Bastin","email":"","orcid":"","institution":"Centre for Clinical Brain Sciences, University of Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Mark","middleName":"","lastName":"Bastin","suffix":""},{"id":414419449,"identity":"55b1f562-41cf-4bb7-be4b-f154a7889b9a","order_by":23,"name":"Alexa Beiser","email":"","orcid":"","institution":"Department of Biostatistics, Boston University School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Alexa","middleName":"","lastName":"Beiser","suffix":""},{"id":414419450,"identity":"e2c7569d-6acc-4323-871c-469362483cf3","order_by":24,"name":"David Bennett","email":"","orcid":"","institution":"Rush University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"Bennett","suffix":""},{"id":414419451,"identity":"e986ad25-a56d-46c8-a0d8-46bb0bee5cd0","order_by":25,"name":"Ravi Bhatt","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Ravi","middleName":"","lastName":"Bhatt","suffix":""},{"id":414419452,"identity":"b361d872-e399-46ca-9a5b-d60c23201d06","order_by":26,"name":"Guillaume Bourque","email":"","orcid":"","institution":"McGill University","correspondingAuthor":false,"prefix":"","firstName":"Guillaume","middleName":"","lastName":"Bourque","suffix":""},{"id":414419453,"identity":"446ed441-61f5-49ad-9b08-2b91436a1107","order_by":27,"name":"Henry Brodaty","email":"","orcid":"https://orcid.org/0000-0001-9487-6617","institution":"Centre for Healthy Brain Ageing","correspondingAuthor":false,"prefix":"","firstName":"Henry","middleName":"","lastName":"Brodaty","suffix":""},{"id":414419454,"identity":"38a53272-1394-4ab1-a481-7157b279fd4f","order_by":28,"name":"Sven Cichon","email":"","orcid":"","institution":"University of Basel","correspondingAuthor":false,"prefix":"","firstName":"Sven","middleName":"","lastName":"Cichon","suffix":""},{"id":414419455,"identity":"33dda286-fabe-478d-a818-866ffc229164","order_by":29,"name":"Amelie Colling","email":"","orcid":"","institution":"University of Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Amelie","middleName":"","lastName":"Colling","suffix":""},{"id":414419456,"identity":"6c16e1ba-7163-4e50-8c84-8be9fb556e09","order_by":30,"name":"Ian Deary","email":"","orcid":"","institution":"university of Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Ian","middleName":"","lastName":"Deary","suffix":""},{"id":414419457,"identity":"949bdb4d-f33e-4b68-8b04-6f9dd1ac9105","order_by":31,"name":"Christian Enzinger","email":"","orcid":"","institution":"Medical University of Graz","correspondingAuthor":false,"prefix":"","firstName":"Christian","middleName":"","lastName":"Enzinger","suffix":""},{"id":414419458,"identity":"eb4ee5cb-4ca8-4ed7-a9dd-28f9d38e9f3b","order_by":32,"name":"Lars Forsberg","email":"","orcid":"","institution":"Karolinska Institutet","correspondingAuthor":false,"prefix":"","firstName":"Lars","middleName":"","lastName":"Forsberg","suffix":""},{"id":414419459,"identity":"513c0436-36b5-461f-beed-a12d8a403cf1","order_by":33,"name":"Piyush Gampawar","email":"","orcid":"https://orcid.org/0000-0003-4759-3686","institution":"Research Unit-Genetic Epidemiology, Institute of Molecular Biology and Biochemistry, Medical University of Graz","correspondingAuthor":false,"prefix":"","firstName":"Piyush","middleName":"","lastName":"Gampawar","suffix":""},{"id":414419460,"identity":"f6451e11-43df-49be-8073-be385ee0377d","order_by":34,"name":"Rebecca Gottesman","email":"","orcid":"","institution":"National Institute of Neurological Disorders and Stroke","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Gottesman","suffix":""},{"id":414419461,"identity":"1923bdf4-e3ac-4cbd-8097-28848ea96011","order_by":35,"name":"Val borg Gudmundsdottir","email":"","orcid":"","institution":"University of Iceland","correspondingAuthor":false,"prefix":"","firstName":"Val","middleName":"borg","lastName":"Gudmundsdottir","suffix":""},{"id":414419462,"identity":"01b5baf6-ca44-4137-b64e-284273a3858a","order_by":36,"name":"Mohamad Habes","email":"","orcid":"https://orcid.org/0000-0001-9447-5805","institution":"Biggs Alzheimer's Institute","correspondingAuthor":false,"prefix":"","firstName":"Mohamad","middleName":"","lastName":"Habes","suffix":""},{"id":414419463,"identity":"8cd6199d-a4d9-4861-aa3c-5f895507a791","order_by":37,"name":"Tsuyoshi Hachiya","email":"","orcid":"","institution":"Iwate Medical University","correspondingAuthor":false,"prefix":"","firstName":"Tsuyoshi","middleName":"","lastName":"Hachiya","suffix":""},{"id":414419464,"identity":"fc355906-19ec-4fa0-a9e1-af5e1d633dbc","order_by":38,"name":"Edith Hofer","email":"","orcid":"","institution":"Medical University of Graz","correspondingAuthor":false,"prefix":"","firstName":"Edith","middleName":"","lastName":"Hofer","suffix":""},{"id":414419465,"identity":"fb914e10-9e2d-451e-babf-360992be1041","order_by":39,"name":"Justine Hsu","email":"","orcid":"","institution":"Weill Cornell Medicine","correspondingAuthor":false,"prefix":"","firstName":"Justine","middleName":"","lastName":"Hsu","suffix":""},{"id":414419466,"identity":"6ca429e9-aad1-4d49-a7af-08445fa103c5","order_by":40,"name":"Mohammad Ikram","email":"","orcid":"https://orcid.org/0000-0003-0372-8585","institution":"Erasmus University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Mohammad","middleName":"","lastName":"Ikram","suffix":""},{"id":414419467,"identity":"67e02143-870b-4522-a8e9-07128f946c99","order_by":41,"name":"M. Imtiaz","email":"","orcid":"","institution":"German Center for Neurodegenerative Diseases","correspondingAuthor":false,"prefix":"","firstName":"M.","middleName":"","lastName":"Imtiaz","suffix":""},{"id":414419468,"identity":"7ab4b426-47c7-4b3f-bd0f-a53f779a8691","order_by":42,"name":"Tasfiya Islam","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Tasfiya","middleName":"","lastName":"Islam","suffix":""},{"id":414419469,"identity":"632cd300-cf20-4466-82fc-268da7e8be95","order_by":43,"name":"Jiyang Jiang","email":"","orcid":"https://orcid.org/0000-0002-2147-6302","institution":"Centre for Healthy Brain Ageing","correspondingAuthor":false,"prefix":"","firstName":"Jiyang","middleName":"","lastName":"Jiang","suffix":""},{"id":414419470,"identity":"b68f656e-8b2d-403a-90b9-883331b9643f","order_by":44,"name":"Takahisa Kawaguchi","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Takahisa","middleName":"","lastName":"Kawaguchi","suffix":""},{"id":414419471,"identity":"f8b72656-b9b3-4190-87bf-8674aab8e0c9","order_by":45,"name":"Kengo Kinoshita","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Kengo","middleName":"","lastName":"Kinoshita","suffix":""},{"id":414419472,"identity":"e0d16cdc-8537-4327-85a8-d5b4ddf4597e","order_by":46,"name":"Maria Knol","email":"","orcid":"https://orcid.org/0000-0002-3597-1531","institution":"Erasmus MC University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"","lastName":"Knol","suffix":""},{"id":414419473,"identity":"a3411944-8ba8-43c7-8c88-7c43d95bc2a3","order_by":47,"name":"Alexandra Koch","email":"","orcid":"","institution":"German Center for Neurodegenerative Diseases","correspondingAuthor":false,"prefix":"","firstName":"Alexandra","middleName":"","lastName":"Koch","suffix":""},{"id":414419474,"identity":"bc3793f3-72a4-4b3d-a09f-62e6f283fd61","order_by":48,"name":"You Kojima","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"You","middleName":"","lastName":"Kojima","suffix":""},{"id":414419475,"identity":"955d95f8-849d-4181-b1b4-f73eadf5a53f","order_by":49,"name":"Mark Lathrop","email":"","orcid":"","institution":"University of McGill Genome Center","correspondingAuthor":false,"prefix":"","firstName":"Mark","middleName":"","lastName":"Lathrop","suffix":""},{"id":414419476,"identity":"f4f607bd-6ea5-4e5d-af28-cce6117e84fc","order_by":50,"name":"Sue Leurgans","email":"","orcid":"https://orcid.org/0000-0003-2536-1665","institution":"Rush Alzheimer Disease Center, Rush University Medical Center, Chicago, United States","correspondingAuthor":false,"prefix":"","firstName":"Sue","middleName":"","lastName":"Leurgans","suffix":""},{"id":414419477,"identity":"68168cb7-fca9-42d5-b185-0b7dda130ee6","order_by":51,"name":"Fumihiko Matsuda","email":"","orcid":"","institution":"Kyoto University","correspondingAuthor":false,"prefix":"","firstName":"Fumihiko","middleName":"","lastName":"Matsuda","suffix":""},{"id":414419478,"identity":"6cb0e4a5-4152-4a9d-89fc-57074bce52f8","order_by":52,"name":"Naoko Mori","email":"","orcid":"","institution":"Akita University Graduate School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Naoko","middleName":"","lastName":"Mori","suffix":""},{"id":414419479,"identity":"16c78ad4-2ea9-4de5-a50c-1f0a9e25d0c7","order_by":53,"name":"Thomas Mosley","email":"","orcid":"","institution":"University of Mississippi Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Mosley","suffix":""},{"id":414419480,"identity":"1420a956-6254-419c-8c7a-e45f8f15c473","order_by":54,"name":"Ikuko Motoike","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Ikuko","middleName":"","lastName":"Motoike","suffix":""},{"id":414419481,"identity":"0ed77866-b51e-48d6-b5ee-6ca91d94265e","order_by":55,"name":"Shunji Mugikura","email":"","orcid":"","institution":"Tohoku University","correspondingAuthor":false,"prefix":"","firstName":"Shunji","middleName":"","lastName":"Mugikura","suffix":""},{"id":414419482,"identity":"78525e39-39d6-4f9f-af70-ebb8c92738be","order_by":56,"name":"Ilya Nasrallah","email":"","orcid":"https://orcid.org/0000-0003-2346-7562","institution":"University of Pennsylvania","correspondingAuthor":false,"prefix":"","firstName":"Ilya","middleName":"","lastName":"Nasrallah","suffix":""},{"id":414419483,"identity":"b0bfc167-9c3a-4099-bf23-225e5dfd2ca7","order_by":57,"name":"Talia Nir","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Talia","middleName":"","lastName":"Nir","suffix":""},{"id":414419484,"identity":"823a1d01-d1c6-4d26-b3b4-738461bcab85","order_by":58,"name":"Yash Patel","email":"","orcid":"","institution":"University of Toronto","correspondingAuthor":false,"prefix":"","firstName":"Yash","middleName":"","lastName":"Patel","suffix":""},{"id":414419485,"identity":"d8467dc1-1d71-46f0-834c-1a70328f811e","order_by":59,"name":"James Prendergast","email":"","orcid":"","institution":"University of Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"","lastName":"Prendergast","suffix":""},{"id":414419486,"identity":"d7e5bc12-4e08-4880-8d85-6a76b7e30372","order_by":60,"name":"Jose Romero","email":"","orcid":"https://orcid.org/0000-0002-1101-2950","institution":"Neurology Department, Boston University School of Medicine and The Framingham Heart Study, Framingham, MA","correspondingAuthor":false,"prefix":"","firstName":"Jose","middleName":"","lastName":"Romero","suffix":""},{"id":414419487,"identity":"670d91bb-70cd-4dd2-93e6-9c91acd54bad","order_by":61,"name":"Perminder Sachdev","email":"","orcid":"https://orcid.org/0000-0002-9595-3220","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Perminder","middleName":"","lastName":"Sachdev","suffix":""},{"id":414419488,"identity":"0586ac29-2e77-45e6-94a6-38994f57b3f5","order_by":62,"name":"Claudia Satizabal","email":"","orcid":"","institution":"University of Texas Health Sciences Center","correspondingAuthor":false,"prefix":"","firstName":"Claudia","middleName":"","lastName":"Satizabal","suffix":""},{"id":414419489,"identity":"64a14709-8c4b-4f29-9776-40e53e1c2c8f","order_by":63,"name":"Markus Scholz","email":"","orcid":"https://orcid.org/0000-0002-4059-1779","institution":"University Leipzig","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"","lastName":"Scholz","suffix":""},{"id":414419490,"identity":"8814535e-b498-4c14-8c60-204cad0f73c6","order_by":64,"name":"Kevin Sheth","email":"","orcid":"https://orcid.org/0000-0003-2003-5473","institution":"Yale University School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Kevin","middleName":"","lastName":"Sheth","suffix":""},{"id":414419491,"identity":"d3f29be8-df2a-44d8-a579-569ef768f4f1","order_by":65,"name":"Jean Shin","email":"","orcid":"","institution":"Centre Hospitalier Universitaire Sainte-Justine","correspondingAuthor":false,"prefix":"","firstName":"Jean","middleName":"","lastName":"Shin","suffix":""},{"id":414419492,"identity":"b51fbb74-bb6a-4489-af05-964b6fe9e2c3","order_by":66,"name":"Sigurdur Sigurdsson","email":"","orcid":"","institution":"University of Iceland","correspondingAuthor":false,"prefix":"","firstName":"Sigurdur","middleName":"","lastName":"Sigurdsson","suffix":""},{"id":414419493,"identity":"497f0179-6ab0-488d-96d3-d22f91856c34","order_by":67,"name":"Anbupalam Thalamuthu","email":"","orcid":"https://orcid.org/0000-0002-7114-1260","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Anbupalam","middleName":"","lastName":"Thalamuthu","suffix":""},{"id":414419494,"identity":"257c61e4-e59c-4477-8a37-fb5cf40044df","order_by":68,"name":"Julian Trollor","email":"","orcid":"https://orcid.org/0000-0002-7685-2977","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Julian","middleName":"","lastName":"Trollor","suffix":""},{"id":414419495,"identity":"c7662974-2bb3-4417-846d-2439ab4f9827","order_by":69,"name":"Ami Tsuchida","email":"","orcid":"","institution":"University of Bordeaux, Bordeaux population health research center","correspondingAuthor":false,"prefix":"","firstName":"Ami","middleName":"","lastName":"Tsuchida","suffix":""},{"id":414419496,"identity":"1360b1c5-ed0f-4db5-9211-cc4c38997eed","order_by":70,"name":"Arno Villringer","email":"","orcid":"https://orcid.org/0000-0003-2604-2404","institution":"Max Planck Institute for Human Cognitive and Brain Sciences","correspondingAuthor":false,"prefix":"","firstName":"Arno","middleName":"","lastName":"Villringer","suffix":""},{"id":414419497,"identity":"1b8c3d82-5e80-4ef4-b509-35286eead898","order_by":71,"name":"Joanna Wardlaw","email":"","orcid":"https://orcid.org/0000-0002-9812-6642","institution":"University of Edinburgh","correspondingAuthor":false,"prefix":"","firstName":"Joanna","middleName":"","lastName":"Wardlaw","suffix":""},{"id":414419498,"identity":"c3fb91cb-7b1d-4efb-b13f-70ca3917b71e","order_by":72,"name":"Wei Wen","email":"","orcid":"https://orcid.org/0000-0003-2753-3870","institution":"University of New South Wales","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Wen","suffix":""},{"id":414419499,"identity":"696d1c51-6973-4751-802e-3a521cad33b6","order_by":73,"name":"Jingyun Yang","email":"","orcid":"","institution":"Rush University Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Jingyun","middleName":"","lastName":"Yang","suffix":""},{"id":414419500,"identity":"f2dd1edd-494a-40f8-aebe-04b8443efe44","order_by":74,"name":"Qiong Yang","email":"","orcid":"https://orcid.org/0000-0002-3658-1375","institution":"Boston University School of Public Health","correspondingAuthor":false,"prefix":"","firstName":"Qiong","middleName":"","lastName":"Yang","suffix":""},{"id":414419501,"identity":"f18292c5-d6e3-4d0f-ac3c-a08640268ed3","order_by":75,"name":"Menglei Zhang","email":"","orcid":"","institution":"Yale School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Menglei","middleName":"","lastName":"Zhang","suffix":""},{"id":414419502,"identity":"1105a02a-7e36-4784-b5f2-5a65a51fda00","order_by":76,"name":"Alyssa Zhu","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Alyssa","middleName":"","lastName":"Zhu","suffix":""},{"id":414419503,"identity":"b6f7aabb-a0a6-416b-b97d-926e5b7e012a","order_by":77,"name":"Thomas Mühleisen","email":"","orcid":"","institution":"Institute of Neurosciences and Medicine","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Mühleisen","suffix":""},{"id":414419504,"identity":"ee8dde56-087b-469e-a4d6-799998247c19","order_by":78,"name":"Karen Mather","email":"","orcid":"https://orcid.org/0000-0003-4143-8941","institution":"Centre for Healthy Brain Ageing, Psychiatry, University of New South Wales (UNSW)","correspondingAuthor":false,"prefix":"","firstName":"Karen","middleName":"","lastName":"Mather","suffix":""},{"id":414419505,"identity":"a6e7769f-fc2e-481e-a7d0-1f319c193c0e","order_by":79,"name":"Ron Do","email":"","orcid":"https://orcid.org/0000-0002-3144-3627","institution":"Icahn School of Medicine at Mount Sinai","correspondingAuthor":false,"prefix":"","firstName":"Ron","middleName":"","lastName":"Do","suffix":""},{"id":414419506,"identity":"36de77da-181d-4cc4-b8b8-abfcc8c02d12","order_by":80,"name":"Zdenka Pausova","email":"","orcid":"","institution":"University of Montreal","correspondingAuthor":false,"prefix":"","firstName":"Zdenka","middleName":"","lastName":"Pausova","suffix":""},{"id":414419507,"identity":"edc0ff76-7cb0-4ba9-8e8c-132e08660316","order_by":81,"name":"Philip De Jager","email":"","orcid":"https://orcid.org/0000-0002-8057-2505","institution":"Columbia University Irving Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Philip","middleName":"","lastName":"De Jager","suffix":""},{"id":414419508,"identity":"814aad98-e364-42da-b1e8-4e8880801c8d","order_by":82,"name":"Fabrice Crivello","email":"","orcid":"","institution":"University of Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Fabrice","middleName":"","lastName":"Crivello","suffix":""},{"id":414419509,"identity":"9f009b5b-1fce-4334-9ac4-074069d1cdb4","order_by":83,"name":"Svenja Caspers","email":"","orcid":"","institution":"Institute of Neuroscience and Medicine (INM-1), Research Centre Jülich","correspondingAuthor":false,"prefix":"","firstName":"Svenja","middleName":"","lastName":"Caspers","suffix":""},{"id":414419510,"identity":"ab8c7565-5fd7-43b9-93cc-3a09a2ac1c2a","order_by":84,"name":"Veronica Witte","email":"","orcid":"","institution":"Max Planck Institute for Human Cognitive and Brain Sciences","correspondingAuthor":false,"prefix":"","firstName":"Veronica","middleName":"","lastName":"Witte","suffix":""},{"id":414419511,"identity":"7597b1e1-c340-49a4-a740-3ad80bd61b8c","order_by":85,"name":"Christophe Tzourio","email":"","orcid":"https://orcid.org/0000-0002-6517-2984","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Christophe","middleName":"","lastName":"Tzourio","suffix":""},{"id":414419512,"identity":"9ad58881-9fab-4908-aec6-b55bb4573f9a","order_by":86,"name":"Hagen Tilgner","email":"","orcid":"","institution":"Weill Cornell Medicine","correspondingAuthor":false,"prefix":"","firstName":"Hagen","middleName":"","lastName":"Tilgner","suffix":""},{"id":414419513,"identity":"e9087bcb-c8a0-4a67-a2c3-22d08ef4a2fe","order_by":87,"name":"Nenad Sestan","email":"","orcid":"","institution":"Yale School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Nenad","middleName":"","lastName":"Sestan","suffix":""},{"id":414419514,"identity":"e625ea3c-4271-4d6b-8fce-1fd418b4c461","order_by":88,"name":"Gennady Roshchupkin","email":"","orcid":"","institution":"Department of Epidemiology, Erasmus MC University Medical Center Rotterdam","correspondingAuthor":false,"prefix":"","firstName":"Gennady","middleName":"","lastName":"Roshchupkin","suffix":""},{"id":414419515,"identity":"751aeec3-c738-4938-9014-2c535699c67c","order_by":89,"name":"Tomasi Paus","email":"","orcid":"","institution":"University of Montreal","correspondingAuthor":false,"prefix":"","firstName":"Tomasi","middleName":"","lastName":"Paus","suffix":""},{"id":414419516,"identity":"323745c4-46f3-443b-a451-4476eb2e169f","order_by":90,"name":"Sudha Seshadri","email":"","orcid":"https://orcid.org/0000-0001-6135-2622","institution":"Glenn Biggs Institute for Alzheimer's \u0026 Neurodegenerative Diseases, University of Texas Health Sciences Centre","correspondingAuthor":false,"prefix":"","firstName":"Sudha","middleName":"","lastName":"Seshadri","suffix":""},{"id":414419517,"identity":"1d90bba9-d4c5-44e7-9885-a80a7413efd4","order_by":91,"name":"Myriam Fornage","email":"","orcid":"https://orcid.org/0000-0003-0677-8158","institution":"The University of Texas Health Science Center at Houston","correspondingAuthor":false,"prefix":"","firstName":"Myriam","middleName":"","lastName":"Fornage","suffix":""},{"id":414419518,"identity":"75707960-5638-4b3c-b691-081118d6d6c3","order_by":92,"name":"Charles DeCarli","email":"","orcid":"https://orcid.org/0000-0003-1914-2693","institution":"University of California, Davis","correspondingAuthor":false,"prefix":"","firstName":"Charles","middleName":"","lastName":"DeCarli","suffix":""},{"id":414419519,"identity":"c4d47912-ee22-4d2a-9a7f-d455c5558f9c","order_by":93,"name":"Lenore Launer","email":"","orcid":"https://orcid.org/0000-0002-3238-7612","institution":"National Institute on Aging, National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Lenore","middleName":"","lastName":"Launer","suffix":""},{"id":414419520,"identity":"23712a2e-062f-47b6-a928-173fb3af2c92","order_by":94,"name":"Helena Schmidt","email":"","orcid":"","institution":"Medical University of Graz","correspondingAuthor":false,"prefix":"","firstName":"Helena","middleName":"","lastName":"Schmidt","suffix":""},{"id":414419521,"identity":"756d0680-0917-4689-9353-bb1385608826","order_by":95,"name":"Vilmundur Gudnason","email":"","orcid":"https://orcid.org/0000-0001-5696-0084","institution":"Icelandic Heart Association","correspondingAuthor":false,"prefix":"","firstName":"Vilmundur","middleName":"","lastName":"Gudnason","suffix":""},{"id":414419522,"identity":"5edc71d6-b045-42c6-9f7c-077f16497444","order_by":96,"name":"Paul Matthews","email":"","orcid":"","institution":"Imperial College London","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Matthews","suffix":""},{"id":414419523,"identity":"c153545e-c9e0-4a60-aefd-b8326b2a382e","order_by":97,"name":"Monique Breteler","email":"","orcid":"https://orcid.org/0000-0002-0626-9305","institution":"German Center for Neurodegenerative Diseases","correspondingAuthor":false,"prefix":"","firstName":"Monique","middleName":"","lastName":"Breteler","suffix":""},{"id":414419524,"identity":"0b7e2993-354e-49df-95c1-184343ed3569","order_by":98,"name":"Aniket Mishra","email":"","orcid":"","institution":"Bordeaux Population Health research center, Université de Bordeaux","correspondingAuthor":false,"prefix":"","firstName":"Aniket","middleName":"","lastName":"Mishra","suffix":""}],"badges":[],"createdAt":"2025-01-29 18:45:49","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5926137/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5926137/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":76281994,"identity":"66f69f35-e0d1-4486-9311-fa7513926c2a","added_by":"auto","created_at":"2025-02-14 10:40:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":327177,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdeogram of genome-wide association results with PSMD across all GWAS models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIdeogram showing 29 genome-wide significant (GWS) PSMD risk loci. Shape corresponds to PSMD models: circles, full model; triangle, age 36-65; diamonds, age \u0026gt; 65. Colors correspond to PSMD association analyses: red, inverse variance weighted (IVW); blue, gene-based test (VEGAS); black, cross-ancestries meta-analyses (MR-MEGA); grey, conditional joint analyses (GCTA-COJO); dark blue, multi trait analysis (MTAG). Nearest genes to lead variants are displayed.\u003c/p\u003e","description":"","filename":"Fig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/d022d69c6d8af6103229be77.png"},{"id":76281988,"identity":"12e1c801-36b1-4fb7-abdb-1d2616e9841b","added_by":"auto","created_at":"2025-02-14 10:40:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":86801,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffect size correlation across ancestries\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePlots showing Pearson’s correlation coefficient (\u003cem\u003er\u003c/em\u003e) between the effect sizes (\u003cem\u003eβ\u003c/em\u003e) of the 23 PSMD risk alleles in Europeans and East Asians (Full model;\u0026nbsp;\u003cem\u003er\u003c/em\u003e = 0.57,\u0026nbsp;\u003cem\u003eP\u003c/em\u003e = 4 x 10\u003csup\u003e−3\u003c/sup\u003e). N = 23 independent PSMD-risk variants\u0026nbsp;from the IVW meta-analyses\u0026nbsp;were used to compute Pearson’s correlation coefficients (\u003cem\u003er\u003c/em\u003e) of the effect sizes between ancestries. The dots represent the effect-size (\u003cem\u003eβ\u003c/em\u003e)\u0026nbsp;estimates and the bars represent the 95% CI of the estimates. Two-sided\u0026nbsp;\u003cem\u003eP\u003c/em\u003e\u0026nbsp;values of the deviation of Pearson’s correlation coefficient from zero are reported. Color corresponds to genome-wide significant association (\u003cem\u003eP\u003c/em\u003e \u0026lt; 5 × 10\u003csup\u003e−8\u003c/sup\u003e) in individual ancestries: blue, European only; purple, European and cross-ancestry; red, cross-ancestry only.\u0026nbsp;\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/f43abaa3e398c7866253e2fa.png"},{"id":76281985,"identity":"f42cda99-520f-44fd-bd88-cdb3b7de3764","added_by":"auto","created_at":"2025-02-14 10:40:41","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":147523,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTrumpet plot of genetic associations from the whole exome association study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDistribution of genetic variants associated with PSMD based on allele frequency and effect size estimates. Red dots represent genetic variants identified from GWAS summary statistics. They include variants highlighted by single-variant association tests and gene-based burden tests. Blue dots correspond to genetic variants identified from Whole Exome Sequencing (WES) association studies. Dots’ size is proportional to -log10 (p-value). Closest gene to rare variants (allele frequency \u0026lt; 0.005) is indicated. The trend in relation between allele frequency and effect size of PSMD-associated genetic variants is represented by the black curves (linear regression model using a polynomial function). These curves illustrate the genetic architecture of PSMD according to the current results.\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/0c83e242b827e0244d720f12.png"},{"id":76281984,"identity":"a5bf9771-c507-48e1-b07f-9c563b86574a","added_by":"auto","created_at":"2025-02-14 10:40:41","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":290050,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation of genetically determined WMH with PSMD across the lifespan\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAssociation of the weighted GRS of WMH volume identified in older adults of European ancestry (mean age 65 years) with PSMD in the four age strata: 9-11, 18-35, 36-65, and \u0026gt;65 years. N corresponds to the sample size of each age stratum tested. Effect estimates and P-value correspond to the association of the WMH wGRS with PSMD.\u003c/p\u003e","description":"","filename":"Fig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/ff3bb17e287e2416f3f23739.png"},{"id":76282722,"identity":"85d39e0a-57e8-4bd9-957e-79e09ba83079","added_by":"auto","created_at":"2025-02-14 10:48:45","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":278878,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation of PSMD loci with MRI-markers of cSVD, vascular risk factors, and neurological diseases\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eVenn diagram displaying significant associations of genome-wide significant risk loci for PSMD with A. vascular risk factors B. MRI-markers of cSVD and other diffusion tensor imaging metrics, and C. neurological diseases. Significance threshold after multiple testing: \u003cem\u003eP \u003c/em\u003e\u0026lt; 1.08x10\u003csup\u003e-4\u003c/sup\u003e, corresponding to a correction for 22 independent phenotypes, 3 association models, and 21 independent loci. Two independent signals were observed at the \u003cem\u003eSH3PXD2A\u003c/em\u003e-\u003cem\u003eSTN1\u003c/em\u003e locus, only one of which was associated with other traits\u003c/p\u003e","description":"","filename":"Fig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/0599656d9a7188c463d8cc23.png"},{"id":76282716,"identity":"c9e50cd0-155f-4ce9-af5f-0b798d168609","added_by":"auto","created_at":"2025-02-14 10:48:44","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":568787,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHeatmap of genetic correlation of PSMD with risk factors, neurological diseases, and other MRI-markers of brain aging\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGenetic correlations and Mendelian randomization causal estimates of vascular risk factors, neurological diseases, and MRI-markers of cSVD. Genetic correlations or causal estimates that are significant after multiple testing correction (P \u0026lt; 7.58 × 10\u003csup\u003e−4\u003c/sup\u003e) are marked with an asterisk. Two-sided P-values were calculated using LD score regression (LDSC)\u0026nbsp;for genetic correlations and inverse variance weighted analysis for MR. The colors represent the direction of the genetic correlation (positive in red, negative in blue).\u003c/p\u003e","description":"","filename":"Fig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/a9680c328d1650e708fedb4c.png"},{"id":76282715,"identity":"9d90d976-8da6-42ba-a6c3-0bea4edd30c6","added_by":"auto","created_at":"2025-02-14 10:48:44","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":176573,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTranscriptome-wide association study (TWAS) of PSMD in multiple tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHeatmap of the transcriptome-wide association results of PSMD (full model, age 36-65 years and age \u0026gt; 65 years strata) reaching transcriptome wide significance with colocalization; Colored squares are TWAS significant two-sided p-values after multiple testing correction (p\u0026lt;2.7x10\u003csup\u003e-6\u003c/sup\u003e); * Conditionally significant (p\u0026lt;0.05) and COLOC PP4 ≥ 0.70; Genes are presented on the x-axis, those underlined in blue are in a\u0026nbsp;PSMD GWAS locus, those underlined in purple are not within a genome-wide significant PSMD locus (\u003cstrong\u003eMethods\u003c/strong\u003e); Tissue types are on the y-axis (pink: arterial; orange: heart; green: brain; blue: fibroblasts).\u003c/p\u003e","description":"","filename":"Fig.7.png","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/17c8ac957f21ca124c269ed9.png"},{"id":76284492,"identity":"d675d7c8-f49a-4635-83ee-624b90bf8702","added_by":"auto","created_at":"2025-02-14 10:56:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4921093,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/070ccdbf-c273-4fc1-b013-b9c4846a76c1.pdf"},{"id":76282002,"identity":"6a5a4761-8c30-478a-9512-a9c4b58a26e3","added_by":"auto","created_at":"2025-02-14 10:40:43","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":39057,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"Table1.docx","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/77a018778b6411ab94347225.docx"},{"id":76284462,"identity":"c6cf4b31-0bd7-4164-b887-d82271dc0ce2","added_by":"auto","created_at":"2025-02-14 10:56:43","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":120331,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Appendix\u003c/p\u003e","description":"","filename":"PSMDGenomicsSupplementaryAppendixSubmit.docx","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/210343cd6ad75dcdb11e3d3a.docx"},{"id":76282001,"identity":"fa2722eb-07e5-4a73-8e94-a8993d1544f3","added_by":"auto","created_at":"2025-02-14 10:40:43","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":360396,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Tables\u003c/p\u003e","description":"","filename":"PSMDGenomicsSupplementaryTablesSubmit.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/a727b7807d56f43d3df00916.xlsx"},{"id":76281998,"identity":"e3fe8e2b-bcbd-438f-bc6d-3b11bcaa7545","added_by":"auto","created_at":"2025-02-14 10:40:43","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":2576138,"visible":true,"origin":"","legend":"","description":"","filename":"EXTENDEDDATA.docx","url":"https://assets-eu.researchsquare.com/files/rs-5926137/v1/1cdb3da9ff254d0ba0f0440b.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Genomics of diffusion-imaging integrating GWAS, exome data and single-cell sequencing unravels lifespan determinants of cerebral small vessel disease","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eCerebral small vessel disease (cSVD) is a leading cause of stroke, cognitive decline, and dementia, and also a major source of postural balance, gait, and mood disturbances in older age.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e In addition, large population-based cohort studies have shown that covert cSVD, detectable with brain imaging in the absence of positive neurological history, is extremely common in the general population with increasing age, portending a considerably increased risk of stroke, dementia, and disability. Covert cSVD could therefore represent a major target to prevent stroke and dementia and promote healthier brain aging. However, there is no mechanism-based treatment to date for cSVD. Imaging features most commonly used to define cSVD include volume of white matter hyperintensities (WMH), lacunes, microbleeds, and perivascular space burden.\u003csup\u003e\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e Most of these imaging markers represent advanced stages of the disease, reflecting consequences of alterations of small vessel structure and function on the brain parenchyma. Their quantification is still in great part based on labor-intensive visual reading of brain scans or heterogeneous automated software, subject to variability and bias.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eDiffusion tensor imaging (DTI) is the most commonly used magnetic resonance imaging (MRI) technique to study variations in white-matter microstructure in cSVD,\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e detectable long before the occurrence of the aforementioned macrostructural lesions on conventional MRI. The typical pattern of DTI in cSVD is a reduction in directionality as captured by fractional anisotropy (FA), and a higher magnitude of diffusion as captured by mean diffusivity (MD).\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e Recently, a robust and fully automated DTI measure known as the peak width of skeletonized mean diffusivity (PSMD), which is highly sensitive to change in longitudinal analyses and reproducible across different MRI scanners, was proposed as a novel marker of cSVD.\u003csup\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e PSMD is the width of the distribution of MD in core white matter regions, thereby reflecting variability of MD. PSMD increases steadily with age across the adult lifespan, in contrast with MD and FA, which have a U-shaped pattern.\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e PSMD is associated with WMH volume, lacunes, total brain volume, global cognition, executive function, and processing speed.\u003csup\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e Its association with processing speed was shown to be stronger than for WMH volume, lacunes, and total brain volume.\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e Therefore, PSMD might represent a useful phenotype to identify genetic susceptibility to cSVD across the lifespan.\u003c/p\u003e \u003cp\u003eIn recent years, large collaborative genome-wide association studies (GWAS), mostly conducted in older individuals aged 65 years on average, have identified over 70 genetic risk loci for cSVD.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e Interestingly, genetic risk variants for WMH volume detected in older age already showed association with DTI metrics at age 20, in a direction compatible with variations preceding WMH occurrence.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e Several GWAS of DTI-based indices of white-matter microstructure have also been published, mostly on UK Biobank.\u003csup\u003e\u003cspan additionalcitationids=\"CR14 CR15 CR16\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e The largest of these (N\u0026thinsp;=\u0026thinsp;43,802 participants, mean age 54.2 years), identified 121 loci associated with region-specific FA and/or MD.\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e To our knowledge, no genomic association study of the summary measure PSMD, offering a more stable measure of diffusivity, that better represents microstructural damage across white matter tracts than traditional, region-specific DTI metrics, has been conducted so far.\u003c/p\u003e \u003cp\u003eHere we performed a meta-GWAS of PSMD in 58,403 individuals from 24 population-based cohort studies (89% European [EUR], 10% East-Asian [EAS], and 1% African ancestry [AFR]), and a whole-exome association study of PSMD in 32,957 population-based participants leveraging whole-exome and whole-genome sequencing data. Our main objective was to identify novel and early molecular signatures of cSVD risk. We further examined shared genetic determinants between PSMD and other MRI-markers of cSVD, putative vascular risk factors, and common neurological diseases. Finally, by integrating our genomic findings with tissue and cell-specific transcriptomic and proteomic data, we sought to detect putative causal genes and mechanisms to be prioritized for experimental follow-up and drug discovery.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eGenetic discovery through GWAS\u003c/h2\u003e \u003cp\u003eOur study population for the GWAS meta-analyses of PSMD comprised 58,403 participants from 24 population-based cohorts (mean age 58.46\u0026thinsp;\u0026plusmn;\u0026thinsp;6.19 years, range 18\u0026ndash;100 years, 48.20% women, \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e). We tested association of PSMD with ~\u0026thinsp;7.5\u0026nbsp;million common single nucleotide polymorphisms (SNPs) with minor allele frequency (MAF)\u0026thinsp;\u0026gt;\u0026thinsp;0.01 (see \u003cb\u003eMethods\u003c/b\u003e and \u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e for genotyping and imputation methods). Participants were of European (N\u0026thinsp;=\u0026thinsp;51,921, 89%), East-Asian (N\u0026thinsp;=\u0026thinsp;5,691, 10%), and African-American (N\u0026thinsp;=\u0026thinsp;521, 1%) ancestry. PSMD, which corresponds to the dispersion of MD values across white matter tracts, was calculated using the same fully automated, publicly available script in all cohorts (\u003cb\u003eMethods\u003c/b\u003e and \u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e for MRI protocols in each cohort). GWAS of PSMD were conducted using linear regression with ln(PSMD*10000) as the dependent variable to correct for skewness, adjusting for age, sex, total intracranial volume (TIV), principal components of population stratification, study site, and familial relationships when applicable. Inverse-variance-weighted (IVW) meta-analyses were conducted using METAL, within each ancestry, followed by a meta-analysis across ancestries. To identify age-specific associations we also conducted secondary GWAS stratified on age: \u0026le;35 years (N\u0026thinsp;=\u0026thinsp;3,411), 36\u0026ndash;65 years (N\u0026thinsp;=\u0026thinsp;23,782), and \u0026gt;\u0026thinsp;65 years (N\u0026thinsp;=\u0026thinsp;24,654, \u003cb\u003eMethods, Extended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThere was no evidence for systematic inflation of association statistics at the cohort or meta-analysis level (\u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e). In the European-only GWAS, 16 independent genome-wide significant loci (P\u0026thinsp;\u0026lt;\u0026thinsp;5x10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e) were identified across all PSMD models (full model and age-stratified analyses, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, \u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;5\u003c/b\u003e). In sensitivity analyses without adjustment for TIV, associations were substantially unchanged (\u003cb\u003eSupplementary Table\u0026nbsp;6\u003c/b\u003e). Using GCTA-COJO to detect independent associations within individual loci by conditioning on the lead variant,\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e we identified five additional independent signals, one at chr10q24.33 (\u003cem\u003eSH3PXD2A\u003c/em\u003e), one at chr6p21.33 (\u003cem\u003eMUC21\u003c/em\u003e), and three at chr6p22.1 \u003cem\u003e(HLA-G\u003c/em\u003e, \u003cem\u003eHLA-H\u003c/em\u003e, \u003cem\u003eLOC401242, TRIM27\u003c/em\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;7, Extended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), consistent with LD clumping. In the cross-ancestry IVW GWAS meta-analyses, two additional loci were identified at chr1p31.3 (\u003cem\u003eJAK1\u003c/em\u003e) and chr12q13.13 (\u003cem\u003eATP5G2\u003c/em\u003e) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;8\u003c/b\u003e). In addition, multi-ancestry meta-regression (MR-MEGA)\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e was performed for loci showing heterogeneity in allelic effect across ancestries (\u003cem\u003eP-het\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01, \u003cb\u003eMethods\u003c/b\u003e, \u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), identifying two additional loci at chr9q22.31 (\u003cem\u003eIARS\u003c/em\u003e) and chr6p24.3 (low-frequency variant near \u003cem\u003eTFAP2A\u003c/em\u003e), leading to a total of 25 independent loci associated with PSMD (\u003cb\u003eSupplementary Table\u0026nbsp;9\u003c/b\u003e).\u003c/p\u003e \u003cp\u003ePer-allele effect sizes showed moderate correlation (r\u0026thinsp;=\u0026thinsp;0.57) between European and East-Asian participants, the two largest contributing ancestries (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;10\u003c/b\u003e). Despite the smaller sample size of the Japanese cohort study by Tohoku Medical Megabank Organization (ToMMo, N\u0026thinsp;=\u0026thinsp;5,961 vs. 55,060 in the EUR GWAS meta-analysis), a weighted genetic risk score (wGRS) of European PSMD lead SNPs or best proxies in East-Asians (\u003cb\u003eMethods\u003c/b\u003e) was significantly associated with PSMD in the ToMMo cohort (beta\u0026thinsp;=\u0026thinsp;0.077\u0026thinsp;\u0026plusmn;\u0026thinsp;0.017; P\u0026thinsp;=\u0026thinsp;6.82x10\u003csup\u003e-6\u003c/sup\u003e).\u003c/p\u003e \u003cp\u003eTo identify probable causal variants at genome-wide significant risk loci, we performed multiple-causal-variant fine-mapping using SuSiE\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e in Europeans (\u003cb\u003eMethods\u003c/b\u003e). We identified 23 95% credible set (CS-trait) pairs, overall and across age strata, at 15 PSMD risk loci, each having a 95% posterior probability of containing a causal variant with multiple CS identified. None of these loci had a single credible SNP, but 2 loci had only 2 credible SNPs (rs12521212 and rs7728421 at chr5q23.2, and rs62434144 and rs275350 at chr6q25.1, \u003cb\u003eSupplementary Table\u0026nbsp;11\u003c/b\u003e). Of note, rs275350 is a \u003cem\u003ecis\u003c/em\u003e protein quantitative trait locus (\u003cem\u003ecis\u003c/em\u003e-pQTL, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;8.56x10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e) in plasma for LRP11,\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e involved in lipid metabolism, response to extreme cold,\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e and binge eating\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e (\u003cb\u003eSupplementary Table\u0026nbsp;12\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eFurthermore, to enhance statistical power, and given a significant genetic correlation (r\u003csub\u003eg\u003c/sub\u003e=0.56, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;7.24x10\u003csup\u003e\u0026minus;\u0026thinsp;30\u003c/sup\u003e) between PSMD and WMH volume, we ran multitrait genome-wide association analyses using MTAG,\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e to increase statistical power by including summary statistics from GWAS of both PSMD and WMH volume.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e We identified 6 additional genome-wide significant loci for PSMD (\u003cb\u003eSupplementary Table\u0026nbsp;13, Extended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), at \u003cem\u003eAMZ2P1\u003c/em\u003e, \u003cem\u003eVTA1\u003c/em\u003e, \u003cem\u003eFOXF2\u003c/em\u003e, \u003cem\u003eDEPDC1B\u003c/em\u003e, \u003cem\u003eAPOE\u003c/em\u003e, and \u003cem\u003eRAPGEF4\u003c/em\u003e, which were not associated at genome-wide significance with either PSMD or WMH volume\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e in single trait GWAS. Thus, across all methods (IVW, GCTA-COJO, MR-MEGA, and MTAG), we identified 31 loci associated with PSMD at genome-wide significance (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn gene-based analyses, we tested the combined association of variants within genes with PSMD in European ancestry participants (\u003cb\u003eMethods, Supplementary Table\u0026nbsp;14\u003c/b\u003e). Using MAGMA\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e we identified 33 gene-wide significant associations (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;2.62x10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) of which nine in five loci did not reach genome-wide significance in the GWAS (Full model: \u003cem\u003eCALCRL, ATP5G2\u003c/em\u003e, and \u003cem\u003eNPAS4\u003c/em\u003e; age\u0026thinsp;\u0026gt;\u0026thinsp;65 group: \u003cem\u003eGPSM3/VWA7/PBX2/AGER/RNF5\u003c/em\u003e, \u003cem\u003eAPOC1\u003c/em\u003e); using VEGAS\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e we identified 6 additional gene-wide significant associations at 5 loci (\u003cem\u003eWDR12/ICA1L, TTL, TFPI\u003c/em\u003e (same locus as \u003cem\u003eCALCRL\u003c/em\u003e), \u003cem\u003eMOG\u003c/em\u003e, \u003cem\u003eSLC54A4\u003c/em\u003e) for the full model, and one additional gene, \u003cem\u003eAPOE\u003c/em\u003e, for the age\u0026thinsp;\u0026gt;\u0026thinsp;65 stratum (same locus as \u003cem\u003eAPOC1\u003c/em\u003e), all in suggestive GWAS loci (5x10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e\u0026lt;P\u0026thinsp;\u0026lt;\u0026thinsp;5x10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;15\u003c/b\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eGenetic discovery through whole-exome association study (WEAS)\u003c/h3\u003e\n\u003cp\u003eWe systematically tested associations of rare and common exonic variants with PSMD using the whole-exome sequencing (WES) resource from UK Biobank (N\u0026thinsp;=\u0026thinsp;29,938 with PSMD data), using similar association models as for GWAS; significant associations were followed up in the BRIDGET (BRain Imaging, cognition, Dementia and next generation Genomics) whole genome sequencing resource (N\u0026thinsp;=\u0026thinsp;1,647, \u003cb\u003eMethods\u003c/b\u003e, \u003cb\u003eSupplementary Methods\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eThe single variant association analysis identified 4 loci associated with PSMD in the full model (all ages combined) in UK Biobank at P\u0026thinsp;\u0026lt;\u0026thinsp;5x10\u003csup\u003e\u0026minus;\u0026thinsp;8\u003c/sup\u003e, including common synonymous variants at two PSMD GWAS loci (\u003cem\u003eTRIM47\u003c/em\u003e and \u003cem\u003eEFEMP1\u003c/em\u003e) and two novel rare risk variants, namely a synonymous variant (rs148195895, MAF\u0026thinsp;=\u0026thinsp;1.22\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) in \u003cem\u003eST3GAL5\u003c/em\u003e and a 20kb deletion (10:88910872:TATTGAAAATCCCACTAATCA:T, MAF\u0026thinsp;=\u0026thinsp;2.54\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) in the intron of \u003cem\u003eSTAMBPL1\u003c/em\u003e (\u003cb\u003eSupplementary Table\u0026nbsp;16\u003c/b\u003e). In participants aged\u0026thinsp;\u0026gt;\u0026thinsp;65 years, we additionally identified a significant association of a common intronic variant in \u003cem\u003eNOTCH4\u003c/em\u003e at chr6p21.33, another PSMD GWAS locus (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e). The common variant associations at \u003cem\u003eTRIM47\u003c/em\u003e and \u003cem\u003eEFEMP1\u003c/em\u003e showed the same direction of effect in the much smaller BRIDGET dataset and became more significant in the meta-analysis of UK Biobank and BRIDGET than in UK Biobank alone. Rare variants in \u003cem\u003eST3GAL5\u003c/em\u003e and \u003cem\u003eSTAMBPL1\u003c/em\u003e were monomorphic in BRIDGET and the meta-analysis of all single variant association statistics in UK Biobank and BRIDGET did not yield additional PSMD risk loci (\u003cb\u003eSupplementary Table\u0026nbsp;16)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eNext, we performed a gene-based burden analysis, testing associations of PSMD with 310,705 variant sets representing 18,190 genes in UK Biobank, and reporting associations at gene-wide (p\u0026thinsp;\u0026lt;\u0026thinsp;2.75\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) and set-wide (p\u0026thinsp;\u0026lt;\u0026thinsp;1.31\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e) significance thresholds. Analyses were conducted with five different variant masks and four MAF thresholds (\u003cb\u003eMethods\u003c/b\u003e). We identified gene-wide significant associations with variant burden in five genes in the full model (\u003cb\u003eSupplementary Table\u0026nbsp;17A\u003c/b\u003e), of which two (\u003cem\u003eKIF13B\u003c/em\u003e, \u003cem\u003eCABYR\u003c/em\u003e) also reached set-wide significance. In participants aged\u0026thinsp;\u0026gt;\u0026thinsp;65 years, we additionally identified a gene-wide significant association with the burden of singleton variants in \u003cem\u003eITPKC\u003c/em\u003e (\u003cb\u003eSupplementary Table\u0026nbsp;17A\u003c/b\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Most gene-based burden associations were discovered for sets of singletons, rare (MAF\u0026thinsp;\u0026lt;\u0026thinsp;0.01), and very rare (MAF\u0026thinsp;\u0026lt;\u0026thinsp;0.001) variants. Given the smaller sample size and underrepresentation of rare and very rare alleles, we were unable to replicate these gene-based burden associations in BRIDGET (\u003cb\u003eSupplementary Table\u0026nbsp;17A\u003c/b\u003e). In secondary analyses using effect-agnostic gene-based tests with SKAT and ACATV, we identified significant associations of PSMD with burden of rare variants in \u003cem\u003eST3GAL5\u003c/em\u003e (\u003cb\u003eSupplementary Table\u0026nbsp;17B\u003c/b\u003e), driven by the aforementioned synonymous variant \u0026lsquo;rs148195895\u0026rsquo;.\u003c/p\u003e \u003cp\u003eWe further tested associations of PSMD with burden of rare variants in eight familial cSVD and Alzheimer's disease (AD) genes, representing 160 variant sets. We identified set-wide significant (p\u0026thinsp;\u0026lt;\u0026thinsp;3.13\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) associations of PSMD with burden of loss-of-function (LOF) and missense very rare variants (MAF\u0026thinsp;\u0026lt;\u0026thinsp;0.001) in \u003cem\u003eHTRA1\u003c/em\u003e (p\u0026thinsp;=\u0026thinsp;5.49\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e), gene-wide significant (p\u0026thinsp;\u0026lt;\u0026thinsp;6.25\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) associations with burden of singletons in \u003cem\u003eNOTCH3\u003c/em\u003e (p\u0026thinsp;=\u0026thinsp;5.65\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e) and close to gene-wide significant associations with burden of very rare variants in \u003cem\u003ePSEN2\u003c/em\u003e (p\u0026thinsp;=\u0026thinsp;7.10\u0026times;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;17C\u003c/b\u003e).\u003c/p\u003e\n\u003ch3\u003eLifespan exploration of PSMD and relationship with genetically determined WMH\u003c/h3\u003e\n\u003cp\u003eTo identify age-specific effects we conducted an age-stratified meta-analysis of GWAS and used the p-value for heterogeneity as a proxy of p-value for interaction. No significant evidence for interaction with age was observed for any of the 16 genome-wide significant risk loci in the European ancestry IVW meta-analysis (p\u0026thinsp;\u0026lt;\u0026thinsp;3.1x10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Using LD-score regression (LDSC)\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, we estimated the heritability of PSMD at 15\u0026thinsp;\u0026plusmn;\u0026thinsp;1% overall, 5.4\u0026thinsp;\u0026plusmn;\u0026thinsp;11%, 19\u0026thinsp;\u0026plusmn;\u0026thinsp;2%, and 18\u0026thinsp;\u0026plusmn;\u0026thinsp;3% at age 18\u0026ndash;35, 36\u0026ndash;65, and \u0026gt;\u0026thinsp;65 respectively (\u003cb\u003eSupplementary Table\u0026nbsp;18\u003c/b\u003e). We found strong genetic correlation between PSMD measured in the middle- (36\u0026ndash;65) and older-age (\u0026gt;\u0026thinsp;65) group (r\u003csub\u003eg\u003c/sub\u003e=0.57\u0026thinsp;\u0026plusmn;\u0026thinsp;0.11, P\u0026thinsp;=\u0026thinsp;3.31x10\u003csup\u003e\u0026minus;\u0026thinsp;7\u003c/sup\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;18\u003c/b\u003e), while the youngest age group was too small to establish reliable genetic correlation. A wGRS of PSMD combining independent genome-wide significant variants from the European ancestry PSMD GWAS (full model) was significantly associated with PSMD already in children in an independent sample (ABCD [N\u0026thinsp;=\u0026thinsp;3,769, mean age 9.96\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63 years], p\u0026thinsp;=\u0026thinsp;1.78x10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;19\u003c/b\u003e). Individually, one PSMD risk locus (common variant in \u003cem\u003eSMG6\u003c/em\u003e) was significantly associated with PSMD in all age groups (P\u0026thinsp;=\u0026thinsp;2.21x10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e in children).\u003c/p\u003e \u003cp\u003eNext, to explore whether PSMD is also associated with genetic susceptibility to established cSVD features in different age groups throughout the lifetime, we used a wGRS of WMH volume derived from the largest European-ancestry WMH GWAS and tested its association with PSMD in independent European-ancestry participants in four age strata (\u0026lt;\u0026thinsp;18 years \u0026ndash; N\u0026thinsp;=\u0026thinsp;3,769, 18\u0026ndash;35 years - N\u0026thinsp;=\u0026thinsp;3,265, 36\u0026ndash;65 years - N\u0026thinsp;=\u0026thinsp;4,159; and \u0026gt;\u0026thinsp;65 years - N\u0026thinsp;=\u0026thinsp;3,356, \u003cb\u003eMethods\u003c/b\u003e). The WMH wGRS was significantly associated with PSMD across all four age strata: effect sizes were similar across the adult lifetime, and an order of magnitude smaller in children (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Of the 25 independent lead WMH genetic risk variants, 4 were associated with PSMD at p\u0026thinsp;\u0026lt;\u0026thinsp;0.002 (0.05/25), at \u003cem\u003eVCAN\u003c/em\u003e and \u003cem\u003eCOL4A2\u003c/em\u003e at age 18\u0026ndash;35, and \u003cem\u003eVCAN\u003c/em\u003e, \u003cem\u003eCARF1\u003c/em\u003e, and \u003cem\u003eRASL12\u003c/em\u003e at age\u0026thinsp;\u0026lt;\u0026thinsp;18 (\u003cb\u003eSupplementary Table\u0026nbsp;20\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eImaging and clinical correlates of PSMD\u003c/h3\u003e\n\u003cp\u003eWe explored shared genetic variation of PSMD with 22 independent traits (i) traditional MRI-markers of cSVD or other DTI metrics, (ii) vascular risk factors, and (iii) the most common neurological diseases associated with cSVD, i.e. stroke, its subtypes, and AD, using the largest published GWAS (\u003cb\u003eMethods\u003c/b\u003e, \u003cb\u003eSupplementary Methods, Supplementary Table\u0026nbsp;21A\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eOf the 21 European-ancestry PSMD GWAS loci (lead or proxy variants with r\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.90), after accounting for the number of independent variants and traits tested (p\u0026thinsp;\u0026lt;\u0026thinsp;1.08x10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e), 10 (48%) were associated with at least one traditional MRI-marker of cSVD (WMH [38%], PVS [24%]), and 5 (24%) with other DTI metrics (FA or MD)). Moreover, 9 loci (43%) were associated with at least one vascular risk factor, mostly blood pressure; and 4 (19%) with risk of stroke (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, \u003cb\u003eSupplementary Table\u0026nbsp;21B\u003c/b\u003e). Systematic screening for PSMD risk variants in the GWAS catalog highlighted shared genetic variants with additional complex traits, especially regional volumes and cortical phenotypes on brain MRI (\u003cb\u003eSupplementary Table\u0026nbsp;22\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eUsing LD-score regression,\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e accounting for 22 traits explored (p\u0026thinsp;\u0026lt;\u0026thinsp;7.58x10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e), we found significant genetic correlations of PSMD with WMH, WM-PVS, MD, and FA (|r\u003csub\u003eg\u003c/sub\u003e|\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026ge;\u003c/span\u003e0.6 for WMH, MD and FA), with diastolic and systolic blood pressure (DBP, SBP), any stroke and ischemic stroke, and additionally with intracerebral hemorrhage in participants\u0026thinsp;\u0026gt;\u0026thinsp;65 years (r\u003csub\u003eg\u003c/sub\u003e\u0026gt;0.6, \u003cb\u003eSupplementary Table\u0026nbsp;23\u003c/b\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, \u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFinally, we used two-sample Mendelian randomization to explore the causal association of vascular risk factors with PSMD, and of PSMD with neurological diseases, using generalized summary-data-based Mendelian randomization (GSMR)\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e, and confirming significant associations (p\u0026thinsp;\u0026lt;\u0026thinsp;7.58x10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e) with TwoSampleMR\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e (\u003cb\u003eMethods, Supplementary Table\u0026nbsp;24\u003c/b\u003e). We used large published GWAS for instruments and outcomes after removing overlapping samples (mainly through contributions of UK Biobank, \u003cb\u003eMethods\u003c/b\u003e). After multiple testing correction, genetically determined higher SBP and DBP were significantly associated with higher PSMD in participants aged\u0026thinsp;\u0026gt;\u0026thinsp;65 years, while genetically predicted higher PSMD was significantly associated with increased risk of intracerebral hemorrhage (\u003cb\u003eSupplementary Table\u0026nbsp;24\u003c/b\u003e). There was no evidence for reverse causation (MR-Steiger)\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e or horizontal pleiotropy (weighted mode, MR-Egger intercept).\u003c/p\u003e\n\u003ch3\u003eFunctional exploration of identified PSMD loci\u003c/h3\u003e\n\u003cp\u003eUsing MAGMA\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e we identified significant enrichment of PSMD GWAS loci in the \u0026ldquo;superoxide_generating_nadph_oxidase_activator_activity\u0026rdquo; oxidative stress-related pathway (full model) and the \u0026ldquo;negative regulation of cytosolic calcium ion concentration\u0026rdquo; pathway (\u0026gt;\u0026thinsp;65 years, \u003cb\u003eSupplementary Tables\u0026nbsp;25\u003c/b\u003e). We used VEGAS2Pathway\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e to explore pathway enrichment for genes identified the WEAS (gene-based burden results): the most significant pathway, although not significant after multiple testing correction, was related to ubiquitination (\u0026ldquo;protein_K11-linked_ubiquitination\u0026rdquo;, \u003cb\u003eSupplementary Tables\u0026nbsp;26\u003c/b\u003e), previously described as a central pathway in verbal memory\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eNext, to generate hypotheses of putative causal genes and directions of effect, we conducted transcriptome-wide association studies (TWAS) using TWAS-Fusion\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e leveraging the PSMD GWAS summary statistics (European-only) and GTEx v8 multi-tissue expression quantitative trait loci (eQTL) from brain, vascular, and blood tissue. We identified 62 transcriptome-wide expression-trait associations for PSMD that were significant in colocalization analyses (TWAS-COLOC), suggesting evidence for a shared causal variant between the corresponding gene expression and PSMD (\u003cb\u003eSupplementary Table\u0026nbsp;27\u003c/b\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). Overall, 34 genes showed transcriptome-wide significant associations with PSMD: 16 were in 6 genome-wide significant risk loci, while 18 were in 13 loci that did not reach genome-wide significance in the PSMD GWAS (\u003cem\u003eJAK1\u003c/em\u003e, \u003cem\u003eICA1L\u003c/em\u003e/\u003cem\u003eNBEAL1\u003c/em\u003e/\u003cem\u003eFAM117B\u003c/em\u003e/\u003cem\u003eWDR12\u003c/em\u003e, \u003cem\u003eCALCRL, FAM107B\u003c/em\u003e, \u003cem\u003eDPP3\u003c/em\u003e, \u003cem\u003eATP5MC2\u003c/em\u003e, \u003cem\u003eUPF3A\u003c/em\u003e, \u003cem\u003eFAM47E\u003c/em\u003e, \u003cem\u003eNDUFAF2\u003c/em\u003e, \u003cem\u003eCASC1\u003c/em\u003e, \u003cem\u003eDEPDC1B\u003c/em\u003e, \u003cem\u003eANO1\u003c/em\u003e, and \u003cem\u003eAC007993.3\u003c/em\u003e/\u003cem\u003eMPP2\u003c/em\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFurthermore, in proteome-wide association studies (PWAS) lower \u003cem\u003ecis\u003c/em\u003e-regulated protein abundance of \u003cem\u003eICA1L\u003c/em\u003e and \u003cem\u003eTRIM47\u003c/em\u003e in the dorsolateral prefrontal cortex (DLPFC) was associated with higher PSMD, with evidence for colocalization (posterior probability of hypothesis 4 for one shared SNP [PP4]\u0026thinsp;\u0026gt;\u0026thinsp;0.93, full model and age\u0026thinsp;\u0026gt;\u0026thinsp;65, \u003cb\u003eSupplementary Table\u0026nbsp;28, Extended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eApplying the STEAP pipeline to several publicly available brain single-cell sequencing resources in humans and mice (\u003cb\u003eMethods, Supplementary Table\u0026nbsp;29\u003c/b\u003e), we found significant enrichment in brain vascular endothelial cells for PSMD (full model, age 36\u0026ndash;65 and \u0026gt;\u0026thinsp;65), including markedly in human fetal brain vascular endothelial cells. Enrichment in immune-response-related cell types and pericytes was also observed (\u003cb\u003eSupplementary Table\u0026nbsp;30\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eTwo genome-wide significant loci for PSMD and WMH, at \u003cem\u003eSMG6\u003c/em\u003e and \u003cem\u003eVCAN\u003c/em\u003e,\u003csup\u003e\u003cem\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/em\u003e\u003c/sup\u003e both showed early significant associations with PSMD, starting in childhood (\u003cb\u003eSupplementary Tables\u0026nbsp;19\u0026ndash;20\u003c/b\u003e), with reduced expression of \u003cem\u003eSMG6\u003c/em\u003e and \u003cem\u003eVCAN\u003c/em\u003e in brain tissue and fibroblasts being associated with higher PSMD and WMH values in TWAS (\u003cb\u003eSupplementary Table\u0026nbsp;27\u003c/b\u003e).\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eWe therefore explored expression patterns of these two genes in the human brain across the lifespan (\u003cb\u003eMethods\u003c/b\u003e),\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e and found that both had maximal expression in the prenatal period, suggesting developmental effects (\u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea-b). Next, we utilized an unpublished single-nuclei RNA sequencing dataset across the human lifespan in the DLPFC to decipher cell-type specific expression patterns (\u003cb\u003eMethods\u003c/b\u003e). \u003cem\u003eVCAN\u003c/em\u003e expression showed an increasing trend in oligodendrocyte progenitor cells (OPCs) and a decreasing trend in mature oligodendrocytes across the lifecourse, with sharply decreasing expression in endothelial cells until early childhood, supporting developmental involvement, consistent with trends previously described in cSVD rat models\u003csup\u003e\u003cspan additionalcitationids=\"CR38\" citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003eSMG6\u003c/em\u003e exhibited dominant expression in vascular and leptomeningeal cells (VLMCs) throughout the lifespan and a trend towards decreasing expression in OPCs and neurons during development and increasing expression in adulthood (\u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee-f\u003cb\u003e).\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo further substantiate these findings we used single-cell RNA sequencing data from a mouse developmental time course (\u003cb\u003eMethods\u003c/b\u003e).\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e We observed that \u003cem\u003eVCAN\u003c/em\u003e was highly expressed in oligodendrocytes across brain regions, with highest expression at P14 decreasing slowly, thereafter, consistent with patterns seen in humans. \u003cem\u003eSMG6\u003c/em\u003e was expressed strongly in neurons with increasing expression during postnatal development, as in humans, and weaker expression in progenitor cells, vascular cells, and oligodendrocytes (\u003cb\u003eExtended Data\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec-d).\u003c/p\u003e \u003cp\u003eFinally, among genes showing transcriptome-wide significant association with PSMD with colocalization evidence, we prioritized putative drug targets using the genetic priority score (GPS) browser (\u003cb\u003eMethods\u003c/b\u003e).\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e We found evidence of very high GPS (\u0026gt;\u0026thinsp;2.1) for \u003cem\u003eSMG6\u003c/em\u003e and \u003cem\u003eCARF\u003c/em\u003e with ischemic heart disease, conferring a 9.7-fold increased likelihood of having a drug indication in Open Targets (\u003cb\u003eSupplementary Table\u0026nbsp;31\u003c/b\u003e for all PSMD-associated genes with a GPS\u0026thinsp;\u0026gt;\u0026thinsp;1.5).\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn a first, cross-ancestry GWAS of PSMD, a novel fully automated diffusion imaging metric associated with cSVD, in up to 58,403 participants from 24 population-based cohort studies, we identified 31 independent genome-wide significant risk loci, over half of which had not been identified before in GWAS for other MRI-markers of cSVD. In addition, a whole-exome association study in 29,938 participants identified associations of PSMD with rare variants in \u003cem\u003eST3GAL5\u003c/em\u003e and \u003cem\u003eSTAMBPL1;\u003c/em\u003e and burden of rare LoF and singleton variants in \u003cem\u003eKIF13B\u003c/em\u003e and \u003cem\u003eCABYR\u003c/em\u003e. Genetically determined larger WMH volume, the most studied MRI-marker of cSVD, was strongly associated with PSMD across the lifetime, starting in childhood. Mendelian randomization supported causal associations of high blood pressure with PSMD and of PSMD with stroke, especially intracerebral hemorrhage. Using TWAS and PWAS we provided evidence for causal implication of 34 genes predominantly through genetically regulated gene expression and protein levels in vascular and brain tissue. Integration with single-cell sequencing data in humans and mice displayed significant enrichment of PSMD loci in genes expressed in fetal brain vascular endothelial cells and revealed early developmental changes in cell-type specificity for genes with strong lifespan effects (\u003cem\u003eVCAN\u003c/em\u003e and \u003cem\u003eSMG6\u003c/em\u003e).\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eOur findings provide strong genetic evidence that PSMD is a relevant MRI-marker for cSVD, PSMD showed significant genetic correlation with WMH (r\u003csub\u003eg\u003c/sub\u003e=0.63) and to a lesser extent PVS (r\u003csub\u003eg\u003c/sub\u003e=0.26), more prominently so in participants aged\u0026thinsp;\u0026gt;\u0026thinsp;65 years (r\u003csub\u003eg\u003c/sub\u003e=0.72 and 0.43). Several genome-wide significant PSMD loci were shared with other MRI-markers of cSVD, especially WMH (at \u003cem\u003eKCNK2-CENPF, SH3PXD2A-STN1, TRIM47, NBEAL1-ICA1L, CALCRL\u003c/em\u003e, \u003cem\u003ePLEKHG1\u003c/em\u003e, \u003cem\u003eLOC100505841\u003c/em\u003e)\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. The WEAS also revealed significant associations of PSMD with the burden of very rare variants in genes causing monogenic cSVD, especially \u003cem\u003eHTRA1\u003c/em\u003e.\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e Furthermore, corroborating a recent observational study\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e and in line with other MRI-markers of cSVD\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e, 43% of PSMD loci were associated with blood pressure, with evidence for a putative causal association of high blood pressure with higher PSMD values. The association of higher genetically determined PSMD with increased risk of stroke, especially intracerebral hemorrhage, is consistent with known clinical complications of cSVD.\u003c/p\u003e \u003cp\u003eAn interesting feature of PSMD is that it can be measured in a standardized automated fashion at any time in life, allowing to capture subtle changes in the overall white matter microstructure prior to the occurrence of traditional cSVD markers such as WMH. Strikingly, genetically determined larger WMH volume was significantly associated with higher PSMD values across the full lifespan,\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e with effect sizes an order of magnitude smaller in children. Although the cross-sectional nature of our analysis prompts important caution, we speculate that, rather than reflecting early subtle features of cSVD, this association could perhaps correspond, at least in part, to increased vulnerability, or lesser resilience of the brain white matter to vascular insults occurring later in life. As for PVS,\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e we saw strong enrichment of PSMD risk loci in genes expressed in fetal brain vascular endothelial cells, notably in all age strata, supporting an important role of vascular developmental factors. Particularly marked lifespan associations were observed for two loci, with evidence for a causal involvement of \u003cem\u003eVCAN\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e and \u003cem\u003eSMG6\u003c/em\u003e in TWAS, and demonstration of maximal brain expression of these prenatally. We have previously shown an association of common variants in \u003cem\u003eVCAN\u003c/em\u003e, a robust risk locus for cSVD,\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e with lower neurite density index in young adults, specifically in regions harboring the highest frequency of WMH in older age.\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e \u003cem\u003eVCAN\u003c/em\u003e is an extracellular matrix proteoglycan involved in development, inflammation, and remyelination, and has been suggested as a potential drug target for multiple sclerosis and possibly cSVD.\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e Observed changes in the cell-type specificity of its expression throughout the lifespan provide precious insights for the design of future experiments. \u003cem\u003eSMG6\u003c/em\u003e is involved in non-sense-mediated mRNA decay,\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e which regulates axonal guidance during neurodevelopment, and has been implicated in several neurodevelopmental and neurodegenerative diseases.\u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e Interestingly, \u003cem\u003eSMG6\u003c/em\u003e was also predicted to have a very high likelihood of being a suitable drug target using GREP.\u003c/p\u003e \u003cp\u003ePublished GWAS for MRI-markers of cSVD have been conducted in populations of nearly exclusively European ancestry so far (\u0026gt;\u0026thinsp;95%),\u003csup\u003e11\u003c/sup\u003e although cSVD is even more prevalent in other ancestry groups, especially East-Asians.\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e Our study included 11% of non-European participants (10% East-Asian). While the overall correlation of effect sizes between European and East-Asian ancestry participants was moderate to good, some variants showed large discrepancies in effect size, especially those with important differences in allele frequency (\u0026gt;\u0026thinsp;30%, at \u003cem\u003eJAK1\u003c/em\u003e, \u003cem\u003eLOC100505841\u003c/em\u003e, \u003cem\u003eTRIM47\u003c/em\u003e). Enhancing non-European contributions to cSVD GWAS will be crucial to enhance the discovery of novel risk variants and optimize transportability of genetic risk prediction and genomics-driven drug discovery.\u003c/p\u003e \u003cp\u003eThe most significant association with PSMD by far, consistent across age strata and ancestries, was observed for the chr1q41 locus near \u003cem\u003eKCNK2\u003c/em\u003e and \u003cem\u003eCENPF\u003c/em\u003e, previously shown to be associated with WMH, PVS,\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e cortical thickness and surface.\u003csup\u003e\u003cspan additionalcitationids=\"CR51\" citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e \u003cem\u003eKCNK2\u003c/em\u003e encodes a voltage-gated potassium channel, involved in neuronal migration, blood-brain barrier function,\u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e and mechanosensing.\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e \u003cem\u003eCENPF\u003c/em\u003e encodes centromere protein F, a mitotic protein involved in cellular differentiation, vesicle transport, cortical neurogenesis,\u003csup\u003e\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e and endothelial cell proliferation, particularly within the brain, where its upregulation in endothelial cells has been associated with ischemic stroke in mice.\u003csup\u003e\u003cspan additionalcitationids=\"CR57\" citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e At chr17q25, the leading WMH risk locus\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e was also associated at genome-wide significance with PSMD in the full model and older adults (\u0026gt;\u0026thinsp;65 years). Interestingly, higher PSMD values were associated with lower expression levels of \u003cem\u003eTRIM47\u003c/em\u003e in brain tissues and cultured fibroblasts (TWAS)\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e and with lower TRIM47 protein levels in the DLPFC, supporting a putative causal involvement of \u003cem\u003eTRIM47\u003c/em\u003e with LoF effects, in line with recent experimental evidence.\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eThe present GWAS also highlights novel genome-wide significant loci not previously associated with other MRI-markers of cSVD, and TWAS and PWAS provide evidence for putative causal genes to be prioritized for functional follow-up, e.g. \u003cem\u003eSMG6\u003c/em\u003e, \u003cem\u003eTRIM47\u003c/em\u003e, and \u003cem\u003eICA1L\u003c/em\u003e. Additionally, several genes showing transcriptome-wide significant associations with PSMD and significant colocalization, located outside of genome-wide significant risk loci, warrant further explorations. These include for instance \u003cem\u003eJAK1\u003c/em\u003e (Janus Kinase 1), involved in interferon-mediated signaling, and Mendelian diseases associated with immune dysregulation (OMIM 618999), \u003cem\u003eDEPDC1B\u003c/em\u003e (DEP Domain Containing 1B), contributing to positive regulation of the Wnt signaling pathway, itself central in brain-specific angiogenesis and blood-brain barrier integrity\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e,\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e, or \u003cem\u003eATP5MC2\u003c/em\u003e (ATP Synthase Membrane Subunit C Locus 2), involved in energy metabolism and mitochondrial function. Overall, PSMD risk loci were enriched in genes involved in oxidative stress pathways, converging with recent experimental findings in \u003cem\u003eTRIM47\u003c/em\u003e knock-out mice.\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eOur whole exome association study identified novel associations of PSMD with single rare variants in \u003cem\u003eST3GAL5\u003c/em\u003e and \u003cem\u003eSTAMBPL1\u003c/em\u003e and burden of rare variants in \u003cem\u003eKIF13B\u003c/em\u003e and \u003cem\u003eCABYR\u003c/em\u003e. \u003cem\u003eST3GAL5\u003c/em\u003e encodes the GM3 synthase, deficiency of which causes salt and pepper developmental regression syndrome, an autosomal recessive neurocutaneous disorder characterized by recurrent seizures and impaired brain development\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e,\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e. Mutations in \u003cem\u003eSTAMBPL1\u003c/em\u003e cause microcephaly\u0026ndash;capillary malformation syndrome\u003csup\u003e\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003eKIF13B\u003c/em\u003e encodes a polarized transporter of VEGF-A receptor to the plasma membrane of endothelial cells\u003csup\u003e\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e\u003c/sup\u003e, playing a key role in VEGF-A-induced neovascularization and angiogenesis, especially under pathological conditions\u003csup\u003e\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e,\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e. Specific isoforms of \u003cem\u003eCABYR\u003c/em\u003e have been suggested to play a role in brain development \u003csup\u003e\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe acknowledge limitations. Images were acquired on different scanners and eras with varying diffusion parameters and sensitivity, however, participating cohorts used standardized image acquisition protocols and identical PSMD quantification algorithms. Furthermore, PSMD was previously shown to have very good inter-scanner reproducibility.\u003csup\u003e\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u003c/sup\u003e The non-European contribution to the study sample is still limited, although substantially larger than previous genomic studies for MRI-markers of cSVD.\u003c/p\u003e \u003cp\u003eIn summary, in this first large cross-ancestry GWAS and WEAS of PSMD, an emerging, fully automated imaging marker of cSVD, we describe numerous novel genetic risk loci comprising both common and rare coding variants. Through extensive integration with multi-omics resources, including at single-cell resolution and across the lifespan, our findings offer multiple lines of evidence suggesting a lifetime process with developmental factors contributing to cSVD susceptibility. They further provide precious leads for gene prioritization towards the identification of novel therapeutic targets.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDetailed acknowledgments are included in the\u003cstrong\u003e\u0026nbsp;Supplementary method.\u0026nbsp;\u003c/strong\u003eWe thank all the participating cohorts for contributing to this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eL.L., H.S., V.Gudnason, P.M.M., M.M.B., A.M., and S.D. jointly supervised research. \u0026nbsp;Y.S., H.Suzuki, P.M., R.M.T, M.Luciano, I.C., and N.J. contributed equally.\u0026nbsp;Y.S. and S.D. designed and conceived the study. Y.S., M.D., F.B., W.B., M.G., Q.L.G, A.T, H.L., M.Sargurupremraj, M.G.D, H.H.H.A., H.J.A., K.A., N.J.A., N.R.B., M.E.B., A.S.B., D.A.B. R.R.B., G.B., H.B., S.Cichon, A.C., I.J.D., C.E., L.F., P.G., R.F.G., V.G., M.H., T.H., E.H., J.H., M.A.Ikram, M.A.I., T.I., J.J., T.K., K.K., M.J.K., A.K., Y.K., M.Lathrop, S.E.L., F.M., N.M., T.H.M., I.N.M., S.M., I.M.N., T.M.N., Y.P., J.G.D.P., J.R.R., P.S.S., C.S., M.S., K.N.S., J.S., S.Sigurdsson, A.Thalamuthu, J.N.T., A.Tsuchida, A.V., J.M.W., W.W., J.Y., Q.Y., M.Z., A.H.Z., T.W.M., K.A.M., R.D., Z.P., P.L.D.J., F.C., S.C., V.A.W., C.T., H.T., N.S., G.R., T.P., S.S., M.F, and C.D. generated the PSMD phenotype, genomic data, conducted cohort-wise GWAS analyses, and provided funding. \u0026nbsp;Y.S.,\u0026nbsp;Q.L.G., A.M., A.C., M-G.D., I.C., R.D, M.G., T.H., H.T., N.S., M.Z., A.M., N.S., H.U.T.,\u0026nbsp;and G.R. contributed to statistical, bioinformatics and functional analyses.\u0026nbsp;Y.S. and S.D. wrote the manuscript. All of the authors provided critical revision.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthic declerations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eH.U.T has presented at user meetings of 10xGenomics, Oxford Nanopore Technologies and Pacific Biosciences, which in some cases has involved payment for travel, accommodation and food. Other authors declared no potential conflicts of interest with respect to research, authorship, and/or publication of this article.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSummary statistics for the GWAS meta-analysis of the European and cross-ancestry are deposited in the GWAS catalog. Individual cohort data are restricted for privacy and legal reasons (national and European restrictions, including GDPR), just like other meta-analyses of GWAS or sequencing data. This is applicable to all cohorts involved (the cohorts included in the meta-analyses). Access to UKB data was conducted under Application Numbers 18545, 23509, and 94113. We used publicly available data analyses described in the manuscript, including data from GTEx (https://gtexportal.org/home/), the Gusev laboratory (http://gusevlab.org/projects/fusion/), RNA sequencing datasets: PsychENCODE DER-22 (www.ncbi.nlm.nih.gov/geo/, accession code GSE97942), GSE67835 (www.ncbi.nlm.nih.gov/geo/, accession code GSE67835), GSE101601, Allen Brain Atlas (http://portal.brain-map.org/), Mousebrain (http://mousebrain.org/), Tabula Muris (https://tabula-muris.ds.czbiohub.org/). All other data supporting the findings of this study are available within the article, the supplementary information, or the supplementary data files.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe used publicly available data from METAL (https://github.com/statgen/METAL), GCTA-cojo (https://yanglab.wetlake.edu.cn/software/gcta/#COJO), FUMA (https://fuma.ctglab.nl), MAGMA (https://ctg.cncr.nl/software/magma), LD-Score Regression (https://github.com/ldsc/), MTAG (https://github/JonJala/mtag), coloc (https://chr1swallace.github.io/coloc/) R package, TwoSampleMR (https:/mrcieu.github.io/TwoSampleMR), STEAP (https://github.com/erwinerdem/STEAP), CELLECT (https://github.com/perslab/CELLECT), S-LDSC (https://github.com/bulik/ldsc), H-MAGMA (https://github.com/thewonlab/H-MAGMA), HBT (https://hbatlas.org), and susieR (https://stephenslab.github.io/susieR/index.html) R packages.\u003cstrong\u003e\u003cbr\u003e\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003ePasi M, Cordonnier C (2020) Clinical Relevance of Cerebral Small Vessel Diseases. Stroke 51:47\u0026ndash;53\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGodin O et al (2008) White matter lesions as a predictor of depression in the elderly: the 3C-Dijon study. Biol Psychiatry 63:663\u0026ndash;669\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuering M et al (2023) Neuroimaging standards for research into small vessel disease-advances since 2013. Lancet Neurol 22:602\u0026ndash;618\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWardlaw JM et al (2013) Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol 12:822\u0026ndash;838\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWardlaw JM et al (2020) Perivascular spaces in the brain: anatomy, physiology and pathology. Nat Rev Neurol 16:137\u0026ndash;153\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNucifora PGP, Verma R, Lee S-K, Melhem ER (2007) Diffusion-tensor MR imaging and tractography: exploring brain microstructure and connectivity. Radiology 245:367\u0026ndash;384\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaykara E et al (2016) A Novel Imaging Marker for Small Vessel Disease Based on Skeletonization of White Matter Tracts and Diffusion Histograms. Ann Neurol 80:581\u0026ndash;592\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeaudet G et al (2020) Age-Related Changes of Peak Width Skeletonized Mean Diffusivity (PSMD) Across the Adult Lifespan: A Multi-Cohort Study. Front Psychiatry 11:342\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZanon Zotin MC et al (2023) Peak Width of Skeletonized Mean Diffusivity: A Neuroimaging Marker for White Matter Injury. Radiology 306:e212780\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBordes C, Sargurupremraj M, Mishra A, Debette S (2022) Genetics of common cerebral small vessel disease. Nat Rev Neurol 18:84\u0026ndash;101\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuperron M-G et al (2023) Genomics of perivascular space burden unravels early mechanisms of cerebral small vessel disease. Nat Med 29:950\u0026ndash;962\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSargurupremraj M et al (2020) Cerebral small vessel disease genomics and its implications across the lifespan. Nat Commun 11:6285\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePersyn E et al (2020) Genome-wide association study of MRI markers of cerebral small vessel disease in 42,310 participants. Nat Commun 11:2175\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith SM et al (2021) An expanded set of genome-wide association studies of brain imaging phenotypes in UK Biobank. Nat Neurosci 24:737\u0026ndash;745\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao B et al (2021) Large-scale GWAS reveals genetic architecture of brain white matter microstructure and genetic overlap with cognitive and mental health traits (n\u0026thinsp;=\u0026thinsp;17,706). Mol Psychiatry 26:3943\u0026ndash;3955\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao B et al (2021) Common genetic variation influencing human white matter microstructure. Science 372:eabf3736\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRutten-Jacobs LCA et al (2018) Genetic Study of White Matter Integrity in UK Biobank (N\u0026thinsp;=\u0026thinsp;8448) and the Overlap With Stroke, Depression, and Dementia. Stroke 49:1340\u0026ndash;1347\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J et al (2012) Conditional and joint multiple-SNP analysis of GWAS summary statistics identifies additional variants influencing complex traits. Nat Genet 44:369\u0026ndash;375\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026auml;gi R et al (2017) Trans-ethnic meta-regression of genome-wide association studies accounting for ancestry increases power for discovery and improves fine-mapping resolution. Hum Mol Genet 26:3639\u0026ndash;3650\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang G, Sarkar A, Carbonetto P, Stephens M (2020) A simple new approach to variable selection in regression, with application to genetic fine mapping. J R Stat Soc Ser B Stat Methodol 82:1273\u0026ndash;1300\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerkingstad E et al (2021) Large-scale integration of the plasma proteome with genetics and disease. Nat Genet 53:1712\u0026ndash;1721\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang X et al (2022) Transcriptome analysis provides new insights into cold adaptation of corsac fox (\u003cem\u003eVulpes Corsac\u003c/em\u003e). Ecol Evol 12:e8866\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLyu Y et al (2024) Recent selection and introgression facilitated high-altitude adaptation in cattle. Sci Bull 69:3415\u0026ndash;3424\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurstein D et al (2023) Genome-wide analysis of a model-derived binge eating disorder phenotype identifies risk loci and implicates iron metabolism. Nat Genet 55:1462\u0026ndash;1470\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurley P et al (2018) Multi-trait analysis of genome-wide association summary statistics using MTAG. Nat Genet 50:229\u0026ndash;237\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Leeuw CA, Mooij JM, Heskes T, Posthuma D (2015) MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput Biol 11:e1004219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMishra A, Macgregor S (2015) VEGAS2: Software for More Flexible Gene-Based Testing. Twin Res Hum Genet 18:86\u0026ndash;91\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBulik-Sullivan BK et al (2015) LD Score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat Genet 47:291\u0026ndash;295\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu Z et al (2018) Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun 9:224\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHemani G et al (2018) The MR-Base platform supports systematic causal inference across the human phenome. Elife 7:e34408\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHemani G, Tilling K (2017) Davey Smith, G. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet 13:e1007081\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMishra A, MacGregor SA (2017) Novel Approach for Pathway Analysis of GWAS Data Highlights Role of BMP Signaling and Muscle Cell Differentiation in Colorectal Cancer Susceptibility. Twin Res Hum Genet 20:1\u0026ndash;9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDebette S et al (2015) Genome-wide studies of verbal declarative memory in nondemented older people: the Cohorts for Heart and Aging Research in Genomic Epidemiology consortium. Biol Psychiatry 77:749\u0026ndash;763\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGusev A et al (2016) Integrative approaches for large-scale transcriptome-wide association studies. Nat Genet 48:245\u0026ndash;252\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKang HJ et al (2011) Spatio-temporal transcriptome of the human brain. Nature 478:483\u0026ndash;489\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQuick S et al (2022) Loss of the heterogeneous expression of flippase ATP11B leads to cerebral small vessel disease in a normotensive rat model. Acta Neuropathol 144:283\u0026ndash;303\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRajani RM et al (2018) Reversal of endothelial dysfunction reduces white matter vulnerability in cerebral small vessel disease in rats. Sci Transl Med 10:eaam9507\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLopez LM et al (2015) Genes From a Translational Analysis Support a Multifactorial Nature of White Matter Hyperintensities. Stroke 46:341\u0026ndash;347\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoglekar A et al (2024) Single-cell long-read sequencing-based mapping reveals specialized splicing patterns in developing and adult mouse and human brain. Nat Neurosci 27:1051\u0026ndash;1063\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuffy \u0026Aacute; et al (2024) Development of a human genetics-guided priority score for 19,365 genes and 399 drug indications. Nat Genet 56:51\u0026ndash;59\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTraylor M et al (2019) Genetic variation in PLEKHG1 is associated with white matter hyperintensities (n\u0026thinsp;=\u0026thinsp;11,226). Neurology 92:e749\u0026ndash;e757\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTraylor M et al (2021) Genetic basis of lacunar stroke: a pooled analysis of individual patient data and genome-wide association studies. Lancet Neurol 20:351\u0026ndash;361\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDebette S, Markus HS (2022) Stroke Genetics: Discovery, Insight Into Mechanisms, and Clinical Perspectives. Circ Res 130:1095\u0026ndash;1111\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetrea RE et al (2024) Hypertension Trends and White Matter Brain Injury in the Offspring Framingham Heart Study Cohort. Hypertension 81:87\u0026ndash;95\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLe Grand Q et al (2024) Diffusion imaging genomics provides novel insight into early mechanisms of cerebral small vessel disease. Mol Psychiatry. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41380-024-02604-7\u003c/span\u003e\u003cspan address=\"10.1038/s41380-024-02604-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGhorbani S et al (2022) Versican promotes T helper 17 cytotoxic inflammation and impedes oligodendrocyte precursor cell remyelination. Nat Commun 13:2445\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEberle AB, Lykke-Andersen S, M\u0026uuml;hlemann O, Jensen TH (2009) SMG6 promotes endonucleolytic cleavage of nonsense mRNA in human cells. Nat Struct Mol Biol 16:49\u0026ndash;55\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetrić Howe M, Patani R (2023) Nonsense-mediated mRNA decay in neuronal physiology and neurodegeneration. Trends Neurosci 46:879\u0026ndash;892\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMok V et al (2014) Race-ethnicity and cerebral small vessel disease\u0026ndash;comparison between Chinese and White populations. Int J Stroke 9:36\u0026ndash;42\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Meer D et al (2020) Understanding the genetic determinants of the brain with MOSTest. Nat Commun 11:3512\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShadrin AA et al (2021) Vertex-wise multivariate genome-wide association study identifies 780 unique genetic loci associated with cortical morphology. NeuroImage 244:118603\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofer E et al (2020) Genetic correlations and genome-wide associations of cortical structure in general population samples of 22,824 adults. Nat Commun 11:4796\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBando Y, Hirano T, Tagawa Y (2014) Dysfunction of KCNK potassium channels impairs neuronal migration in the developing mouse cerebral cortex. Cereb Cortex 24:1017\u0026ndash;1029\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLengyel M, Enyedi P, Czirj\u0026aacute;k G (2021) Negative Influence by the Force: Mechanically Induced Hyperpolarization via K2P Background Potassium Channels. Int J Mol Sci 22:9062\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaters AM et al (2015) The kinetochore protein, \u003cem\u003eCENPF\u003c/em\u003e, is mutated in human ciliopathy and microcephaly phenotypes. J Med Genet 52:147\u0026ndash;156\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMahmoud AD et al (2019) The Human-Specific and Smooth Muscle Cell-Enriched LncRNA SMILR Promotes Proliferation by Regulating Mitotic CENPF mRNA and Drives Cell-Cycle Progression Which Can Be Targeted to Limit Vascular Remodeling. Circul Res 125:535\u0026ndash;551\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang F et al (2022) Endothelial cell heterogeneity and microglia regulons revealed by a pig cell landscape at single-cell level. Nat Commun 13:3620\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan Y et al (2023) Identification of brain endothelial cell-specific genes and pathways in ischemic stroke by integrated bioinformatical analysis. Brain Circulation 9:228\u0026ndash;239\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFornage M et al (2011) Genome-wide association studies of cerebral white matter lesion burden: the CHARGE consortium. Ann Neurol 69:928\u0026ndash;939\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMishra A et al (2022) Gene-mapping study of extremes of cerebral small vessel disease reveals TRIM47 as a strong candidate. Brain 145:1992\u0026ndash;2007\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDupl\u0026agrave;a C et al (2024) Cerebral Small Vessel Disease genetic determinant TRIM47 controls brain homeostasis \u003cem\u003evia\u003c/em\u003e the NRF2 antioxidant system. Preprint https://doi. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eorg/10.1101/2024.10.08.616723\u003c/span\u003e\u003cspan address=\"10.1101/2024.10.08.616723\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchevenels G et al (2024) A brain-specific angiogenic mechanism enabled by tip cell specialization. Nature 628:863\u0026ndash;871\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBats M-L et al (2022) Wnt/frizzled Signaling in Endothelium: A Major Player in Blood-Retinal- and Blood-Brain-Barrier Integrity. Cold Spring Harb Perspect Med 12:a041219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFragaki K et al (2013) Refractory epilepsy and mitochondrial dysfunction due to GM3 synthase deficiency. Eur J Hum Genet 21:528\u0026ndash;534\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBowser LE et al (2019) Recessive GM3 synthase deficiency: Natural history, biochemistry, and therapeutic frontier. Mol Genet Metab 126:475\u0026ndash;488\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcDonell LM et al (2013) Mutations in STAMBP, encoding a deubiquitinating enzyme, cause microcephaly-capillary malformation syndrome. Nat Genet 45:556\u0026ndash;562\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamada KH et al (2014) KIF13B regulates angiogenesis through Golgi to plasma membrane trafficking of VEGFR2. J Cell Sci 127:4518\u0026ndash;4530\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaters SB et al (2021) VEGFR2 Trafficking by KIF13B Is a Novel Therapeutic Target for Wet Age-Related Macular Degeneration. Invest Ophthalmol Vis Sci 62:5\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaters SB et al (2022) KIF13B-mediated VEGFR2 trafficking is essential for vascular leakage and metastasis in vivo. Life Sci Alliance 5:e202101170\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHsu H-C et al (2005) Characterization of two non-testis-specific CABYR variants that bind to GSK3beta with a proline-rich extensin-like domain. Biochem Biophys Res Commun 329:1108\u0026ndash;1117\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaillard P et al (2022) Instrumental validation of free water, peak-width of skeletonized mean diffusivity, and white matter hyperintensities: MarkVCID neuroimaging kits. Alzheimers Dement (Amst) 14:e12261\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eEthics statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study complies with all relevant ethical regulations, and all participants gave written, informed consent. Institutional review boards that approved each contributing study are described in \u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStudy design\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn total, 62,172 participants were included in this study, of whom 55,690 of European (EUR), 5,961 of East-Asian (EAS), and 521 of African-American (AFR) ancestry. Of these 4,269 participants aged 18 years and older (3,411 EUR, 858 EAS) were included in the GWAS meta-analysis, while data from 3,769 children was used to explore lifespan associations with PSMD. Individuals included in this project participated in 25 cohorts\u003cstrong\u003e\u0026nbsp;(Extended Data Fig. 1\u003c/strong\u003e) from the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortium,\u003csup\u003e72\u003c/sup\u003e UK Biobank (UKB), the Brain Imaging, cognition, Dementia, and Next generation GEnomics \u0026nbsp;(BRIDGET) consortium, and the Enhancing Neuroimaging Genetics through meta-analysis (ENIGMA) consortium.\u003csup\u003e73\u003c/sup\u003e Characteristics of study participants for each cohort are provided in \u003cstrong\u003eSupplementary Tables 1-3\u003c/strong\u003e. All participants gave written informed consent, and institutional review boards approved individual studies (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e, \u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePeak width of Skeletonized Mean Diffusivity (PSMD)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eVarious metrics were derived from the DTI data using a script developed by Baykara et al (http://www.psmd-marker.com),\u003csup\u003e7\u003c/sup\u003e and adapted by Beaudet et al.\u003csup\u003e8\u003c/sup\u003e The fully automated calculation of PSMD, corresponding to the dispersion of MD values across white matter tracts, \u0026nbsp;was conducted in two steps: 1) skeletonization of the DTI data using the FA map and 2) histogram analysis of the MD within the white matter skeleton after application of a custom mask as described in Baykara et al.\u003csup\u003e7\u003c/sup\u003e PSMD was computed as the difference between the 95\u003csup\u003eth\u003c/sup\u003e and the 5\u003csup\u003eth\u003c/sup\u003e percentiles of the skeletonized and masked MD volume voxel value distribution.\u003csup\u003e7,8\u003c/sup\u003e MRI scans were acquired using scanners with field strengths ranging from 1.5 to 3.0 Tesla. MRI protocols of each cohort can be found in the \u003cstrong\u003eSupplementary Methods\u0026nbsp;\u003c/strong\u003eand \u003cstrong\u003eSupplementary Table 2\u003c/strong\u003e. Due to a skewed distribution of PSMD in several studies, we used the natural logarithmic transformation of the variable with the following formula ln(PSMD*10000).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenotyping and Imputation\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInformation on genotyping platforms, quality control (QC), and imputation methods for each participating cohort are provided in \u003cstrong\u003eSupplementary Table 3\u003c/strong\u003e. Each cohort conducted genotyping using different commercially available genotyping platforms. Genotyped data were mostly imputed to the Haplotype Reference Consortium (HRC) reference panel, while a few studies used the 1000 Genome reference panel (Phase I version 3), in particular African ancestry samples.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGenome-wide association analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEach cohort performed a linear regression model to determine association between ln(PSMD*10000) and allele dosages of SNPs using additive genetic effects adjusted for age, sex, TIV, and principal components of population stratification, study site, and familial relationships, where relevant. Furthermore, each cohort performed secondary association analyses in three age strata: 18-35, 36-65, and \u0026gt;65.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eQC of summary statistics shared by each cohort was conducted following recommendations of Winkler et al.\u003csup\u003e74\u003c/sup\u003e using EasyQC R package (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e). Rare variants with a minor allele frequency (MAF) \u0026lt; 0.01, with low imputation quality (R\u003csup\u003e2\u003c/sup\u003e \u0026lt; 0.5), or a product of MAF, R\u003csup\u003e2\u003c/sup\u003e,\u003csup\u003e\u0026nbsp;\u003c/sup\u003eand sample size less than 10 were excluded.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFirst, a fixed-effects Inverse-Variance Weighted (IVW) meta-analysis was performed using METAL\u003csup\u003e75\u003c/sup\u003e within each ancestry group (EUR, EAS, and AFR) followed by a meta-analysis across ancestries. Genomic control was applied to each study-specific GWAS with a genomic inflation factor greater than 1 (\u0026lambda;\u003csub\u003eGC\u003c/sub\u003e \u0026gt; 1.00) to correct for any residual population stratification. After meta-analysis, only SNPs represented in more than four of the participating studies and/or more than one-third of the sample size and with no evidence of between-study heterogeneity (P\u003csub\u003eHet\u0026nbsp;\u003c/sub\u003e\u0026gt; 1x10\u003csup\u003e-4\u003c/sup\u003e) were considered. Quantile-Quantile (Q-Q) plots, Manhattan plots, and genomic inflation factors for all meta-analyses are presented in \u003cstrong\u003eExtended Data Fig. 2\u0026nbsp;\u003c/strong\u003eand \u003cstrong\u003eSupplementary Table 4\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAcross all meta-analysis models, we considered variants reaching a P-value \u0026lt; 5x10\u003csup\u003e-8\u003c/sup\u003e as genome-wide significant. To identify independent SNPs within genome-wide risk loci, we used the PLINK clumping function, considering linkage disequilibrium (LD) with both an LD threshold of r\u003csup\u003e2\u0026nbsp;\u003c/sup\u003e\u0026gt; 0.1 and a physical distance of \u0026plusmn; 1 Mb from the index SNPs of a given locus.\u003c/p\u003e\n\u003cp\u003eIn order to identify variants that were independently associated with PSMD within 1 MB of lead SNPs, the step-wise conditional regression and joint analysis (GCTA-COJO)\u003csup\u003e18\u003c/sup\u003e was performed using Genome-wide Complex Trait Analysis (GCTA, version 1.26.0).\u003csup\u003e76\u003c/sup\u003e LD patterns were selected based on HRC imputed data of 1,862 participants from the i-Share study (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e). \u0026nbsp;Next, fixed-effects IVW meta-analysis was performed to conduct cross-ancestry GWAS of PSMD using the METAL software. We used the Cochran Q test as implemented in METAL to test for heterogeneity of effects between ancestries. Furthermore, we conducted a multi-ancestry meta-analysis using the MR-MEGA software\u003csup\u003e19\u003c/sup\u003e, which uses meta-regression to model allelic effects including axes of genetic variation as covariates in the model.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExploration of age-specific genetic associations and heritability estimates\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe SNP-based heritability estimates were calculated using LD score regression (LDSC package https://github.com/bulik/ldsc/)\u003csup\u003e28\u003c/sup\u003e and the European LD-score files calculated from the 1000G reference panel provided by the developers, overall and in each age stratum.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGene-based analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed gene-based analyses on European PSMD GWAS meta-analyses, using the Multi-marker Analysis of Genomic Annotation (MAGMA)\u003csup\u003e26\u003c/sup\u003e software implemented in FUMA\u003csup\u003e77\u003c/sup\u003e (19,090 protein coding genes) and VEGAS2\u003csup\u003e27\u003c/sup\u003e software (18,432 protein coding, autosomal genes). Both methods considered variants in the gene or within 10 kb on either side of a gene\u0026rsquo;s transcription site to compute a gene-based P-value. We performed MAGMA gene-based tests using the default parameters, whereas the VEGAS2 analyses were performed using the \u0026ldquo;\u0026mdash;top 10\u0026rdquo; parameter that tests enrichment of the top 10% variants assigned to a gene accounting for the linkage disequilibrium between variants and the total number of variants within a gene. Gene-wide significance was defined at P-value \u0026lt; 2.62x10\u003csup\u003e-6\u003c/sup\u003e for gene-based tests using MAGMA and P-value \u0026lt; 2.71x10\u003csup\u003e-6\u003c/sup\u003e for VEGAS2. Genes were considered in the same locus if they were \u0026lt; 200 kb apart from each other.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-Trait Analysis of GWAS with white matter hyperintensity volume\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe applied Multi-Trait Analysis of GWAS (MTAG),\u003csup\u003e25\u003c/sup\u003e performing a joint analysis of summary statistics from the EUR PSMD GWAS with the largest EUR WMH GWAS (N=48,454),\u003csup\u003e12\u003c/sup\u003e to uncover additional genetic risk loci for PSMD, as PSMD and WMH were strongly correlated. MTAG estimates per SNP effect size for each trait by incorporating information contained in other correlated traits. The effect size re-estimated by MTAG is a generalized estimate of IVW meta-analysis by integrating GWAS summary statistics from different traits, where the P-value is derived from the re-estimated effect size.\u003csup\u003e25\u0026nbsp;\u003c/sup\u003eWe prioritized associations fulfilling the following conditions: (1) MTAG P-value for PSMD \u0026lt; 5x10\u003csup\u003e-8\u003c/sup\u003e, (2) univariate GWAS P-value for PSMD \u0026lt; 0.05; and (3) MTAG P-value for PSMD was lower the univariate GWAS P-value for WMH.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePSMD whole exome association study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed a whole exome association study (WEAS) to identify (rare) exonic variants associated with PSMD. We conducted discovery analyses on processed population level OQFE \u0026nbsp;(an update \u0026nbsp;Functional Equivalence (FE) protocol) UK biobank whole exome sequencing (WES) data\u003csup\u003e78\u003c/sup\u003e with PSMD information (N=29,938) accessible through the UKB-RAP platform. The REGENIE software\u003csup\u003e79\u003c/sup\u003e (v3.1.1) was used to perform single variant association tests, gene-based burden test, gene-based Sequencing Kernel Association Test (SKAT), and Aggregated Cauchy Association Test with Variance component (ACATV) tests. The REGENIE software employs a two-step approach \u0026ndash; step 1 is applied on a small set of directly genotyped variants that captures a good fraction of the phenotype variance, and step 2 is applied on full WES data testing different association models (e.g. single variant, gene-based, etc.). Prior to REGENIE Step 1, we performed quality control (QC) the directly genotyped array data removing variants with minor allele frequency (MAF) \u0026lt;0.01, minor allele count (MAC) \u0026lt;20, Hardy Weinberg equilibrium (HWE) p-value \u0026lt;1\u0026times;10\u003csup\u003e-15\u003c/sup\u003e, genotype missingness \u0026gt;10%, and samples with \u0026gt;10% missingness. Prior to running REGENIE Step 2, we performed QC of WES data removing variants with HWE p-value\u0026lt;1\u0026times;10\u003csup\u003e-15\u003c/sup\u003e and genotype missingness \u0026gt;10%.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor single variant association testing, we also removed variants with minor allele count (MAC) \u0026lt;10. We used the variant annotation provided in the helper files of population level OQFE UKB WES data generated using SnpEff software.\u003csup\u003e80\u003c/sup\u003e Using this annotation we created the following variant masks for gene-based burden tests: M1: Loss of function (LoF) variants only, M2: LoF and missense variants predicted to be deleterious by five predictor software (LRT,\u003csup\u003e81\u003c/sup\u003e PolyPhen-2 HDIV, PolyPhen-2 HVAR, \u003csup\u003e82\u003c/sup\u003e SIFT,\u003csup\u003e83\u003c/sup\u003e and MutationTaster\u003csup\u003e84\u003c/sup\u003e), \u0026nbsp;M3: LoF and missense variants predicted to be deleterious by at least one of the five predictor software, M4: LoF and all missense variants, M5: LoF, all missense, and all synonymous variants. Additionally, we considered four strata based on maximum MAF of variants namely: singletons, MAF\u0026lt;0.001, MAF\u0026lt;0.01, and MAF\u0026lt;0.05. Variant sets with MAC\u0026lt;5 were removed from gene-based burden tests. Overall 310,705 variant sets of 18,190 genes were tested for gene-based burden analyses, leading to a gene-wide and set-wide significance threshold of p\u0026lt;2.75\u0026times;10\u003csup\u003e-6\u003c/sup\u003e and p\u0026lt;1.31\u0026times;10\u003csup\u003e-7\u003c/sup\u003e respectively.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe also performed secondary effect-agnostic gene-based tests using SKAT\u003csup\u003e85\u003c/sup\u003e and ACATV.\u003csup\u003e86\u003c/sup\u003e For this we considered the default strategy in REGENIE considering all variants in each mask category irrespective of their allele frequency. REGENIE collapses ultra-rare variants (MAC\u0026le;10) into a burden mask to include them in these tests. To avoid reporting ACATV/SKAT associations that are driven by the burden masks of only one or two singleton variants (burden set MAC\u0026lt;5), we manually removed significant variant sets containing less than five singleton variants and all non-ultra-rare variant with single variant association with p-values \u0026gt;0.1. Correcting for 62,355 variant sets considering all variants in respective masks without allele frequency stratification the significant threshold was set at p\u0026lt;8.02\u0026times;10\u003csup\u003e-7\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eDiscovered associations were followed-up in independent BRIDGET (BRain Imaging, cognition, Dementia and next generation Genomics) whole genome sequencing data (N=1,647). Details on the BRIDGET WGS data processing and QC is described in the \u003cstrong\u003eSupplementary Methods\u003c/strong\u003e. BRIDGET data was also analyzed using the REGENIE software, employing similar QC and association parameters as described for UK biobank. The BRIDGET data was annotated using the SnpEff software and the same five variant masks (M1-M5) were created as for gene-based association tests in UK Biobank. Due to small sample size, the gene-based burden test was performed only for variant set with maximum MAF\u0026lt;0.05. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssociation of genetically determined WMH with PSMD across the lifespan\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe explored the association of genetically determined weighted genetic risk score (wGRS) WMH with PSMD across the adult age strata and in younger age stratum in a non-overlapping sample between WMH, and PSMD (18-35, N=3,265; 36-65, N=4,159; \u0026gt;65, N=3,356). Associations were tested using linear mixed models adjusted for age, sex, total intracranial volume, and the first four principal components of populations stratifications (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e). For associations with individual SNPS the significant threshold was set at P\u0026lt;2x10\u003csup\u003e-3\u003c/sup\u003e (0.05/25). The aggregate effect of 25 WMH risk variants with DTI metrics was estimated by using the \u0026ldquo;gtx\u0026rdquo; package in R. In secondary analyses, we searched for an association of PSMD genome-wide significant loci identified in the GWAS with PSMD in 3,769 children aged 9.9\u0026plusmn;0.63 years, participating in ABCD study, using the same approach as described above. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssociation of genetically determined PSMD with PSMD across\u003c/strong\u003e \u003cstrong\u003eancestries\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe tested the association of a European PSMD weighted genetic risk score (wGRS) with PSMD in the Japanese cohort using the R-package \u0026ldquo;gtx\u0026rdquo; (http://cran.r-project.org/web/packages/gtx/). \u0026nbsp;We used a LD reference panel including 7,062 unrelated individuals from the Nagahama study.\u003csup\u003e87\u003c/sup\u003e Among the 16 PSMD lead SNPs, 5 were not present in the Nagahama reference panel, so we used as tag SNPs the SNP in LD r\u003csup\u003e2\u003c/sup\u003e\u0026gt;0.7, window size = 1 Mb (European HRC-imputed 3C-Dijon study or 1000Gp3 reference panel) which was present in the Nagahama reference panel and had the most significant p-value in the PSMD European meta-analysis.\u0026nbsp;Two SNPs (chr6:31807540 and chr6:1365883) were rare variants in EAS. The 14 SNPs included in the GRS were independent using Nagahama reference panel (r\u003csup\u003e2\u003c/sup\u003e\u0026lt;0.10).\u0026nbsp;To construct the wGRS, we used the most significant SNP (or in LD r\u003csup\u003e2\u003c/sup\u003e\u0026gt;0.7, window size = 1 Mb, using Nagahama reference panel) in the Tohoku Megabank cohort which was\u0026nbsp;present in the European PSMD meta-analysis. SNPs were weighted by the SNP effect sizes in the European GWAS meta-analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eShared genetic variation of PSMD with related vascular, neurological, and psychiatric traits\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe first assessed (in European ancestry participants) whether PSMD risk loci were associated with: (i) putative risk factors (SBP,\u003csup\u003e88\u003c/sup\u003e DBP,\u003csup\u003e88\u003c/sup\u003e pulse pressure (PP),\u003csup\u003e88\u003c/sup\u003e body mass index (BMI),\u003csup\u003e89\u003c/sup\u003e high density lipoprotein (HDL) cholesterol,\u003csup\u003e90\u003c/sup\u003e low density lipoprotein (LDL) cholesterol,\u003csup\u003e90\u003c/sup\u003e triglycerides,\u003csup\u003e90\u003c/sup\u003e type 2 diabetes,\u003csup\u003e91\u003c/sup\u003e (ii) other MRI-markers of brain aging (WMH burden,\u003csup\u003e12\u003c/sup\u003e FA,\u003csup\u003e13\u003c/sup\u003e MD,\u003csup\u003e13\u003c/sup\u003e PVS in WM,\u003csup\u003e11\u003c/sup\u003e BG,\u003csup\u003e11\u003c/sup\u003e and HIP\u003csup\u003e11\u003c/sup\u003e); (iii) stroke (any stroke,\u003csup\u003e92\u003c/sup\u003e any ischemic stroke,\u003csup\u003e92\u003c/sup\u003e large artery stroke,\u003csup\u003e92\u003c/sup\u003e cardio-embolic stroke,\u003csup\u003e92\u003c/sup\u003e small vessel stroke,\u003csup\u003e92\u003c/sup\u003e intracerebral hemorrhage [ICH]\u003csup\u003e93\u003c/sup\u003e), and AD.\u003csup\u003e94\u003c/sup\u003e\u0026nbsp; P-value \u0026lt; 3.98x10\u003csup\u003e-5\u003c/sup\u003e correcting for 22 independent phenotypes, the 3 PSMD GWAS models, and 19 independent genome-wide significant PSMD risk loci (EUR IVW analysis (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe then used LD-score regression (LDSC package https://github.com/bulik/ldsc)\u003csup\u003e28\u003c/sup\u003e to assess the genetic correlation between PSMD and the aforementioned traits, using summary statistics from the largest publicly available GWASs.\u003csup\u003e11\u0026ndash;13,88\u0026ndash;93,95\u003c/sup\u003e To decrease the potential bias due to poor imputation quality, the summary statistics were filtered to the subset of HapMap3 SNPs for each trait. A P-value \u0026le; 7.58x10\u003csup\u003e-4\u003c/sup\u003e correcting for 22 phenotypes and for the 3 tested models (overall and by age strata) was considered significant.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAdditionally, we used the Functional Mapping and Annotation of Genome-wide Association studies (FUMA)\u003csup\u003e77\u003c/sup\u003e to obtain extensive functional annotation for genome-wide significant SNPs, and to identify SNPs \u0026nbsp;associated with any other trait from the GWAS catalog at genome-wide significant level.\u003csup\u003e\u0026nbsp;96\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMendelian randomization\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMendelian randomization (MR) was used to seek evidence for a causal relation of putative vascular risk factors (SBP, DBP, PP, BMI, LDL- and HDL-cholesterol, triglyceride, type 2 diabetes) with PSMD, and of PSMD with the most common neurological diseases associated cSVD, stroke, and AD.\u003c/p\u003e\n\u003cp\u003eWe used the following two-sample MR:approaches: the Generalized Summary-data-based Mendelian Randomization (MR) method implemented in GCTA (version 1.93.2beta) software (GCTA-GSMR),\u003csup\u003e29\u0026nbsp;\u003c/sup\u003eand the TwoSampleMR software.\u003csup\u003e30\u003c/sup\u003e\u0026nbsp; To build our instruments for MR, we used genetic risk variants for the aforementioned traits (exposures). Only independent SNPs (LD-r\u003csup\u003e2\u003c/sup\u003e \u0026lt;0.05 for GSMR, and LD-r\u003csup\u003e2\u003c/sup\u003e \u0026lt;0.01 \u0026nbsp;reaching genome-wide significance (P-value \u0026lt; 5x10\u003csup\u003e-8\u003c/sup\u003e) were included.\u003csup\u003e97\u003c/sup\u003e To select these SNPs, we clumped the summary statistics of the vascular risk factors (window: 1000kb, r\u003csup\u003e2\u003c/sup\u003e\u0026lt;0.01) after filtering the SNPs, excluding variants with MAF \u0026lt; 0.01, ambiguous alleles, and non-matching alleles between the exposure and PSMD summary statistics and variants with an average imputation score \u0026lt; 0.9 in the PSMD GWAS.\u003c/p\u003e\n\u003cp\u003eIn GSMR, we removed SNPs that have pleiotropic effects on both exposure and outcome by using the heterogeneity in independent instrument (HEIDI)-outlier method (pHEIDI \u0026lt; 0.01) and ran GSMR, based on a two-step least square approach, to estimate the effects of the exposures on the outcomes.\u003csup\u003e29\u003c/sup\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn TwosampleMR,\u003csup\u003e30\u003c/sup\u003e we harmonized data between exposure and outcome using the default parameters. F-statistic, which should be over 10 to avoid weak instrument bias, was calculated to confirm the attainment of the relevance assumption. We used IVW for the primary analyses, and sensitivity analyses including weighted median, and MR-Egger\u003csup\u003e98\u003c/sup\u003e methods to confirm the observed effects. We \u0026nbsp;applied MR-Egger intercept to assess the horizontal pleiotropy (P\u0026ge;0.05), In addition, we confirmed the directionality of observed associations with Steiger test\u003csup\u003e31\u003c/sup\u003e (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e). A P-value \u0026le; 7.58x10\u003csup\u003e-4\u003c/sup\u003e correcting for 22 independent phenotypes and 3 tested models was considered significant.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePathway analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe conducted pathway analyses with MAGMA gene set analyses\u003csup\u003e26\u003c/sup\u003e implemented in FUMA, on EUR PSMD GWAS summary statistics, using the 1000G phase3 reference panel. 10,678 gene sets (curated gene sets: 4,761, Go terms: 5,917) from MsigdB v6.2 were used, and a P-value \u0026lt; 3.8x10\u003csup\u003e-5\u0026nbsp;\u003c/sup\u003ecorrecting for 1,320 independent gene sets was considered significant. We also used the VEGAS2Pathway approach, which aggregates association strengths of individual variants into pre-specified biological pathways using VEGAS-derived gene association P-values for PSMD. The empirical significance threshold for VEGAS2Pathway was 1x10\u003csup\u003e-5\u003c/sup\u003e accounting for 6,213 correlated pathways.\u003csup\u003e32\u0026nbsp;\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTranscriptome-wide association study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed transcriptome-wide association studies (TWAS) of PSMD using TWAS-Fusion\u003csup\u003e34\u0026nbsp;\u003c/sup\u003e leveraging the association statistics from the EUR PSMD GWAS (Full model, Age 36-65, Age \u0026gt; 65) and precomputed gene expression weights from GTEx v8 gene expression reference panels.\u003csup\u003e99\u003c/sup\u003e\u0026nbsp; TWAS Z score (association statistic between predicted expression and PSMD) was derived from the integration of expression reference panels (SNP-expression weights), GWAS summary statistics (SNP-PSMD effect estimates), and linkage disequilibrium reference panels (SNP correlation matrix).\u003csup\u003e34\u003c/sup\u003e Transcriptome-wide significance genes (eGenes) and corresponding eQTLs were determined using Bonferroni correction based on the average number of features (6140.5 genes) tested across tissues considering all three independent models tested (p\u0026lt;2.7 x 10\u003csup\u003e-6\u003c/sup\u003e). Identified genes were then tested in conditional analysis as implemented in Fusion software. Colocalization analysis was then conducted on the conditionally significant genes (p\u0026lt;0.05) using COLOC\u003csup\u003e100\u003c/sup\u003e to estimate the posterior probability of a shared causal variant between gene expression and PSMD (PP4\u0026ge;0.70). eGene regions with eQTL not in LD (r\u003csup\u003e2\u003c/sup\u003e\u0026lt;0.01) with the lead SNP for genome-wide significant PSMD risk loci were considered as novel. (\u003cstrong\u003eSupplementary Methods\u003c/strong\u003e)\u003cstrong\u003e\u003cem\u003e.\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell type enrichment analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe conducted a cell-type enrichment analysis using \u003cstrong\u003eS\u003c/strong\u003eingle cell \u003cstrong\u003eT\u003c/strong\u003eype \u003cstrong\u003eE\u003c/strong\u003enrichment \u003cstrong\u003eA\u003c/strong\u003enalysis for \u003cstrong\u003eP\u003c/strong\u003ehenotypes (STEAP;\u0026nbsp;https://github.com/erwinerdem/STEAP/). This is an extension to\u0026nbsp;CELLECT\u0026nbsp;and uses\u0026nbsp;S-LDSC,\u003csup\u003e101\u003c/sup\u003e\u0026nbsp; MAGMA,\u003csup\u003e26\u003c/sup\u003e,\u0026nbsp;and\u0026nbsp;H-MAGMA\u003csup\u003e102\u003c/sup\u003e for enrichment analysis \u003cstrong\u003e(Supplementary Methods)\u003c/strong\u003e.\u0026nbsp;PSMD GWAS summary statistics were first munged.\u0026nbsp;Then, expression specificity profiles were calculated using human and mouse single cell RNA-seq databases (\u003cstrong\u003eSupplementary Table 27\u003c/strong\u003e). Cell-type enrichment was calculated with three models: MAGMA, H-MAGMA (incorporating chromatin interaction profiles from human brain tissues in MAGMA), and stratified LD score regression. P-values were corrected for the number of independent cell types in each database (Bonferroni correction).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLifetime brain gene expression profile\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo look for developmental processes, we examined the lifetime expression of genes in loci reaching genome-wide significance with PSMD Phenotype in the children (ABCD) study and the younger age stratum 18-35, as well as WMH genome-wide significant genes which significantly associate with PSMD. We used a public database (https://hbatlas.org/)\u003csup\u003e35\u003c/sup\u003e that contained genome-wide exon-level transcriptome data from 1,340 tissue samples from 16 brain regions of 57 postmortem human brains, spanning from embryonic development to late adulthood. Next, we used single-cell RNA sequencing data from a mouse developmental time course (postnatal days 14, 21, 28, and 56) for visual cortex and hippocampus as well as further P56 data from cerebellum, thalamus, and striatum\u003csup\u003e39\u003c/sup\u003e and calculated the average gene expression and percentage of cells that express \u003cem\u003eSMG6\u003c/em\u003e and \u003cem\u003eVCAN\u003c/em\u003e using the Seurat software.\u003csup\u003e103\u003c/sup\u003e Furthermore, we analyzed unpublished single-nuclei RNA sequencing data from the dorsolateral prefrontal cortex (DLPFC) across the human lifespan, covering prenatal and postnatal development as well as aging, to investigate the temporal dynamic gene expression pattern. Raw counts for each cell type were aggregated across 137 libraries (from 114 donors) as pseudo-bulked counts and scaled to log2(CPM+1). The expression dynamics were smoothed by loess function with a span of 0.8, and a 95% confidence interval, implemented in R version 4.3.2.\u003csup\u003e104\u003c/sup\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePrioritizing drug targets for enhanced drug discovery\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo find evidence of drug target genes, we searched for identified TWAS genes associated with PSMD using browser of Genetic Priority Score for Therapeutic Targets (https://rstudio-connect.hpc.mssm.edu/geneticpriorityscore/). Genes with a high negative GPS-DOE would be suitable targets for inhibitors, whereas genes with a high positive GPS-DOE would be more suitable for drug activation.\u003csup\u003e40\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReferences (METHODS)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e72. \u0026nbsp; \u0026nbsp; \u0026nbsp; Psaty, B. M. \u003cem\u003eet al.\u003c/em\u003e Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium: Design of prospective meta-analyses of genome-wide association studies from 5 cohorts. \u003cem\u003eCirc Cardiovasc Genet\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 73\u0026ndash;80 (2009).\u003c/p\u003e\n\u003cp\u003e73. \u0026nbsp; \u0026nbsp; \u0026nbsp; Bearden, C. E. \u0026amp; Thompson, P. M. Emerging Global Initiatives in Neurogenetics: The Enhancing Neuroimaging Genetics through Meta-analysis (ENIGMA) Consortium. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e94\u003c/strong\u003e, 232\u0026ndash;236 (2017).\u003c/p\u003e\n\u003cp\u003e74. \u0026nbsp; \u0026nbsp; \u0026nbsp; Winkler, T. W. \u003cem\u003eet al.\u003c/em\u003e Quality control and conduct of genome-wide association meta-analyses. \u003cem\u003eNat Protoc\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 1192\u0026ndash;1212 (2014).\u003c/p\u003e\n\u003cp\u003e75. \u0026nbsp; \u0026nbsp; \u0026nbsp; Willer, C. J., Li, Y. \u0026amp; Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 2190\u0026ndash;2191 (2010).\u003c/p\u003e\n\u003cp\u003e76. \u0026nbsp; \u0026nbsp; \u0026nbsp; Yang, J., Lee, S. H., Goddard, M. E. \u0026amp; Visscher, P. M. GCTA: a tool for genome-wide complex trait analysis. \u003cem\u003eAm J Hum Genet\u003c/em\u003e \u003cstrong\u003e88\u003c/strong\u003e, 76\u0026ndash;82 (2011).\u003c/p\u003e\n\u003cp\u003e77. \u0026nbsp; \u0026nbsp; \u0026nbsp; Watanabe, K., Taskesen, E., van Bochoven, A. \u0026amp; Posthuma, D. Functional mapping and annotation of genetic associations with FUMA. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 1826 (2017).\u003c/p\u003e\n\u003cp\u003e78. \u0026nbsp; \u0026nbsp; \u0026nbsp; Krasheninina, O. \u003cem\u003eet al.\u003c/em\u003e Open-source mapping and variant calling for large-scale NGS data from original base-quality scores. Preprint at https://doi.org/10.1101/2020.12.15.356360 (2020).\u003c/p\u003e\n\u003cp\u003e79. \u0026nbsp; \u0026nbsp; \u0026nbsp; Mbatchou, J. \u003cem\u003eet al.\u003c/em\u003e Computationally efficient whole-genome regression for quantitative and binary traits. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 1097\u0026ndash;1103 (2021).\u003c/p\u003e\n\u003cp\u003e80. \u0026nbsp; \u0026nbsp; \u0026nbsp; Cingolani, P. \u003cem\u003eet al.\u003c/em\u003e A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w \u003csup\u003e1118\u003c/sup\u003e ; iso-2; iso-3. \u003cem\u003eFly\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 80\u0026ndash;92 (2012).\u003c/p\u003e\n\u003cp\u003e81. \u0026nbsp; \u0026nbsp; \u0026nbsp; Chun, S. \u0026amp; Fay, J. C. Identification of deleterious mutations within three human genomes. \u003cem\u003eGenome Res\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 1553\u0026ndash;1561 (2009).\u003c/p\u003e\n\u003cp\u003e82. \u0026nbsp; \u0026nbsp; \u0026nbsp; Adzhubei, I., Jordan, D. M. \u0026amp; Sunyaev, S. R. Predicting functional effect of human missense mutations using PolyPhen-2. \u003cem\u003eCurr Protoc Hum Genet\u003c/em\u003e \u003cstrong\u003eChapter 7\u003c/strong\u003e, Unit7.20 (2013).\u003c/p\u003e\n\u003cp\u003e83. \u0026nbsp; \u0026nbsp; \u0026nbsp; Vaser, R., Adusumalli, S., Leng, S. N., Sikic, M. \u0026amp; Ng, P. C. SIFT missense predictions for genomes. \u003cem\u003eNat Protoc\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 1\u0026ndash;9 (2016).\u003c/p\u003e\n\u003cp\u003e84. \u0026nbsp; \u0026nbsp; \u0026nbsp; Schwarz, J. M., R\u0026ouml;delsperger, C., Schuelke, M. \u0026amp; Seelow, D. MutationTaster evaluates disease-causing potential of sequence alterations. \u003cem\u003eNat Methods\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 575\u0026ndash;576 (2010).\u003c/p\u003e\n\u003cp\u003e85. \u0026nbsp; \u0026nbsp; \u0026nbsp; Ionita-Laza, I., Lee, S., Makarov, V., Buxbaum, J. D. \u0026amp; Lin, X. Sequence kernel association tests for the combined effect of rare and common variants. \u003cem\u003eAm J Hum Genet\u003c/em\u003e \u003cstrong\u003e92\u003c/strong\u003e, 841\u0026ndash;853 (2013).\u003c/p\u003e\n\u003cp\u003e86. \u0026nbsp; \u0026nbsp; \u0026nbsp; Liu, Y. \u003cem\u003eet al.\u003c/em\u003e ACAT: A Fast and Powerful p Value Combination Method for Rare-Variant Analysis in Sequencing Studies. \u003cem\u003eAm J Hum Genet\u003c/em\u003e \u003cstrong\u003e104\u003c/strong\u003e, 410\u0026ndash;421 (2019).\u003c/p\u003e\n\u003cp\u003e87. \u0026nbsp; \u0026nbsp; \u0026nbsp; Nakata, I. \u003cem\u003eet al.\u003c/em\u003e Prevalence and characteristics of age-related macular degeneration in the Japanese population: the Nagahama study. \u003cem\u003eAm J Ophthalmol\u003c/em\u003e \u003cstrong\u003e156\u003c/strong\u003e, 1002-1009.e2 (2013).\u003c/p\u003e\n\u003cp\u003e88. \u0026nbsp; \u0026nbsp; \u0026nbsp; Evangelou, E. \u003cem\u003eet al.\u003c/em\u003e Genetic analysis of over 1 million people identifies 535 new loci associated with blood pressure traits. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 1412\u0026ndash;1425 (2018).\u003c/p\u003e\n\u003cp\u003e89. Yengo, L. \u003cem\u003eet al.\u003c/em\u003e Meta-analysis of genome-wide association studies for height and body mass index in \u0026sim;700000 individuals of European ancestry. \u003cem\u003eHum Mol Genet\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 3641\u0026ndash;3649 (2018).\u003c/p\u003e\n\u003cp\u003e90. \u0026nbsp; \u0026nbsp; \u0026nbsp; Willer, C. J. \u003cem\u003eet al.\u003c/em\u003e Discovery and refinement of loci associated with lipid levels. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1274\u0026ndash;1283 (2013).\u003c/p\u003e\n\u003cp\u003e91. \u0026nbsp; \u0026nbsp; \u0026nbsp; Xue, A. \u003cem\u003eet al.\u003c/em\u003e Genome-wide association analyses identify 143 risk variants and putative regulatory mechanisms for type 2 diabetes. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 2941 (2018).\u003c/p\u003e\n\u003cp\u003e92. \u0026nbsp; \u0026nbsp; \u0026nbsp; Mishra, A. \u003cem\u003eet al.\u003c/em\u003e Stroke genetics informs drug discovery and risk prediction across ancestries. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e611\u003c/strong\u003e, 115\u0026ndash;123 (2022).\u003c/p\u003e\n\u003cp\u003e93. \u0026nbsp; \u0026nbsp; \u0026nbsp; Woo, D. \u003cem\u003eet al.\u003c/em\u003e Meta-analysis of genome-wide association studies identifies 1q22 as a susceptibility locus for intracerebral hemorrhage. \u003cem\u003eAm J Hum Genet\u003c/em\u003e \u003cstrong\u003e94\u003c/strong\u003e, 511\u0026ndash;521 (2014).\u003c/p\u003e\n\u003cp\u003e94. \u0026nbsp; \u0026nbsp; \u0026nbsp; Alzheimer Disease Genetics Consortium (ADGC), \u003cem\u003eet al.\u003c/em\u003e Genetic meta-analysis of diagnosed Alzheimer\u0026rsquo;s disease identifies new risk loci and implicates A\u0026beta;, tau, immunity and lipid processing. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 414\u0026ndash;430 (2019).\u003c/p\u003e\n\u003cp\u003e95. \u0026nbsp; \u0026nbsp; \u0026nbsp; Schwartzentruber, J. \u003cem\u003eet al.\u003c/em\u003e Genome-wide meta-analysis, fine-mapping and integrative prioritization implicate new Alzheimer\u0026rsquo;s disease risk genes. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 392\u0026ndash;402 (2021).\u003c/p\u003e\n\u003cp\u003e96. \u0026nbsp; \u0026nbsp; \u0026nbsp; Buniello, A. \u003cem\u003eet al.\u003c/em\u003e The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, D1005\u0026ndash;D1012 (2019).\u003c/p\u003e\n\u003cp\u003e97. \u0026nbsp; \u0026nbsp; \u0026nbsp; Swerdlow, D. I. \u003cem\u003eet al.\u003c/em\u003e Selecting instruments for Mendelian randomization in the wake of genome-wide association studies. \u003cem\u003eInt J Epidemiol\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1600\u0026ndash;1616 (2016).\u003c/p\u003e\n\u003cp\u003e98. \u0026nbsp; \u0026nbsp; \u0026nbsp; Bowden, J., Davey Smith, G. \u0026amp; Burgess, S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. \u003cem\u003eInt J Epidemiol\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e, 512\u0026ndash;525 (2015).\u003c/p\u003e\n\u003cp\u003e99. \u0026nbsp; \u0026nbsp; \u0026nbsp; Carithers, L. J. \u0026amp; Moore, H. M. The Genotype-Tissue Expression (GTEx) Project. \u003cem\u003eBiopreserv Biobank\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 307\u0026ndash;308 (2015).\u003c/p\u003e\n\u003cp\u003e100. \u0026nbsp; \u0026nbsp; Giambartolomei, C. \u003cem\u003eet al.\u003c/em\u003e Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. \u003cem\u003ePLoS Genet\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, e1004383 (2014).\u003c/p\u003e\n\u003cp\u003e101. \u0026nbsp; \u0026nbsp; Finucane, H. K. \u003cem\u003eet al.\u003c/em\u003e Partitioning heritability by functional annotation using genome-wide association summary statistics. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 1228\u0026ndash;1235 (2015).\u003c/p\u003e\n\u003cp\u003e102. \u0026nbsp; \u0026nbsp; Sey, N. Y. A. \u003cem\u003eet al.\u003c/em\u003e A computational tool (H-MAGMA) for improved prediction of brain-disorder risk genes by incorporating brain chromatin interaction profiles. \u003cem\u003eNat Neurosci\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 583\u0026ndash;593 (2020).\u003c/p\u003e\n\u003cp\u003e103. \u0026nbsp; \u0026nbsp; Hao, Y. \u003cem\u003eet al.\u003c/em\u003e Dictionary learning for integrative, multimodal and scalable single-cell analysis. \u003cem\u003eNat Biotechnol\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e, 293\u0026ndash;304 (2024).\u003c/p\u003e\n\u003cp\u003e104. \u0026nbsp; \u0026nbsp; Li, M. \u003cem\u003eet al.\u003c/em\u003e Integrative functional genomic analysis of human brain development and neuropsychiatric risks. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e362\u003c/strong\u003e, eaat7615 (2018).\u003c/p\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5926137/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5926137/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePeak width of skeletonized mean diffusivity (PSMD) is an emerging automated diffusion imaging marker showing clinically relevant changes in cerebral small vessel disease (cSVD), a leading cause of stroke and dementia with no mechanism-based treatment. We conducted a genome-wide association study of PSMD in 58,403 participants from 24 population-based cohorts (89% European, 10% East-Asian, 1% African-American), identifying 31 independent common variant associations. Additionally, a whole-exome sequencing analysis in 32,957 participants yielded associations of PSMD with single and burden of rare coding variants in four novel genes. Mendelian randomization supported causal association of higher blood pressure with larger PSMD values, and of larger PSMD with an increased risk of stroke, especially intracerebral hemorrhage. Strikingly, genetic susceptibility to white matter hyperintensities, an established MRI-marker of cSVD, was associated with higher PSMD from early childhood to older age, with prominent lifespan effects for \u003cem\u003eVCAN\u003c/em\u003e and \u003cem\u003eSMG6\u003c/em\u003e. Leveraging unique brain single-cell sequencing resources we showed temporal changes in the cell-type specificity of these genes in the developing brain and overall enrichment of PSMD risk loci in genes expressed in fetal brain endothelial cells. Finally, through extensive integration with multi-omics resources, we provide precious leads for gene prioritization to accelerate drug discovery for cSVD.\u003c/p\u003e","manuscriptTitle":"Genomics of diffusion-imaging integrating GWAS, exome data and single-cell sequencing unravels lifespan determinants of cerebral small vessel disease","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-02-14 10:40:26","doi":"10.21203/rs.3.rs-5926137/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-aging","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nataging","sideBox":"Learn more about [Nature Aging](https://www.nature.com/nataging/)","snPcode":"","submissionUrl":"","title":"Nature Aging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"4eef589f-720c-4d37-88f6-d9018d3786e0","owner":[],"postedDate":"February 14th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":44194076,"name":"Health sciences/Diseases/Neurological disorders/Cerebrovascular disorders"},{"id":44194077,"name":"Biological sciences/Genetics/Genetic association study/Genome-wide association studies"},{"id":44194078,"name":"Health sciences/Diseases/Neurological disorders/Neurovascular disorders"}],"tags":[],"updatedAt":"2025-03-31T23:15:20+00:00","versionOfRecord":[],"versionCreatedAt":"2025-02-14 10:40:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5926137","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5926137","identity":"rs-5926137","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}