Genome-wide Mendelian Randomization Identifies Potential Drug Targets for Dorsopathies

Mendelian randomization, Genetics, Drug target, Dorsopathies, Colocalization 14 analysis, Genome-wide association studies, Expression quantitative trait locus. 15

16

Dorsopathies are a group of musculoskeletal disorders affecting the spinal column and 17 related structures, contributing significantly to global disability rates and healthcare costs. Despite 18 their prevalence, the genetic and biological mechanisms underlying dorsopathies are not fully 19 understood. 20

Summary-data-based Mendelian Randomization (SMR) and colocalization analysis were 21 employed, using data from genome-wide association studies (GWAS) and cis-expression quantitative 22 trait loci (cis-eQTLs) databases. Genes with a colocalization posterior probability (PP.H4) above 0.7 23 in SMR results were selected for additional analysis. These selected genes underwent MR analysis to 24 examine possible causal connections with dorsopathies, and sensitivity analyses were carried out to 25 ensure robustness. Additionally, two transcriptome-wide association studies (TWAS) were utilized to 26 confirm and screen for potential drug targets. 27

We identified four essential genes linked to dorsopathies: NLRC4, CGREF1, KHK, and 28 RNF212. Mendelian randomization (MR) analysis revealed a potential causal link between these 29 genes and dorsopathies. Elevated transcription levels of NLRC4, CGREF1, and KHK correlated with 30 reduced dorsopathies risk, while increased levels of RNF212 were associated with heightened risk of 31 dorsopathies. Regarding methylation sites, an increase in cg04686953 fully mediated the decreased 32 risk of dorsopathies by RNF212. Similarly, the risk effect of cg26638505 and cg18948125 was 33 entirely mediated by NLRC4, while CGREF1 predominantly mediated the risk-increasing effect of 34 cg06112415 and the decrease effect of cg22740783. 35 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 2

Dorsopathies were associated with four pivotal genes: NLRC4, CGREF1, KHK, and 36 RNF212. Methylation analysis identified cg04686953 and cg22740783 as protective against 37 dorsopathies risk, while cg26638505, cg18948125, and cg06112415 exhibited a risk-increasing 38 impact. 39 40 1 Introduction 41 Dorsopathies encompass a broad spectrum of musculoskeletal issues that impact the spinal column 42 and its related structures[1]. These conditions are prevalent among individuals across various age 43 groups and socioeconomic backgrounds, leading affected individuals to seek medical assistance. 44 They significantly contribute to the burden of illness, disability, and distress. Since 1990, there has 45 been a notable rise in the prevalence of these disorders, rendering them one of the primary causes of 46 global disability-adjusted life years [2]. However, the pathogenic genes and biological mechanisms 47 of the disease remain largely unknown. Research conducted on the general population has indicated a 48 potential link between certain dorsopathies and lifestyle choices including smoking [3], a higher body 49 mass index[4,5], and lack of physical activity [6]. Additionally, these dorsopathies have been found 50 to coincide with various other health complications such as anxiety, depression, diabetes, 51 cardiovascular, respiratory, and gastrointestinal ailments[7,8]. Furthermore, specific dorsopathies like 52 osteopenia, osteomalacia, and tuberculosis can be influenced by factors such as diet, living 53 environment, and psychological aspects. These particular disorders often occur alongside systemic 54 comorbidities such as endocrine dysfunction and infection [8]. Relevant limitations in observational 55 epidemiological studies include the potential for confounding, reverse causation, and diverse biases, 56 which hinder comprehensive understanding of disease pathogenesis and identification of treatment 57 targets. 58 However, the utilization of Mendelian randomization (MR) techniques, which rely on the random 59 allocation of genetic variations, allows for the simulation of randomized controlled trials and can 60 effectively mitigate the impact of confounding variables. In relation to research on dorsopathies, this 61 implies a more precise evaluation of the causal association between drug targets and disease while 62 excluding any factors that may potentially disrupt the findings [9]. MR utilizes genetic variants as 63 instrumental variables to assess the causal impact of an exposure on outcomes. This approach has 64 been extensively utilized in other research studies related to diseases and has effectively facilitated 65 the identification of potential therapeutic targets for diverse medical conditions.[10] However, there 66 is limited research on employing MR methods to explore potential drug targets for dorsopathies. 67 Therefore, to gain a deeper understanding of the pathogenesis of dorsopathies and identify more 68 effective treatment approaches, further MR studies are needed to evaluate the drug targets for 69 dorsopathies. This will help eliminate factors that could interfere with the results and offer new 70 perspectives and methods for the treatment of dorsopathies. In theory, single nucleotide 71 polymorphisms (SNPs) are distributed randomly and not impacted by environmental factors, making 72 them an ideal tool for establishing causality. MR is a type of instrumental variable analysis that 73 primarily employs SNPs as genetic instruments to determine the causal impact of an exposure (in this 74 instance, circulatory proteins) on outcomes.[11]. Previous studies have successfully utilized MR to 75 identify biomarkers and treatment targets for various diseases, such as aortic aneurysms [12], 76 multiple sclerosis [13], and breast cancer [14]. 77 H owever, in previous studies, the use of MR To analyze drug targets for dorsopathies is very rare. To 78 address this research gap, our study focuses on utilizing genes as factors of exposure in drug target 79 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 3 research, investigating the role of genes in dorsopathies development and their potential as viable 80 drug targets. Genes have stable genetic characteristics, and they are unaffected by environmental 81 factors, allowing for a more accurate assessment of the causal relationship between genes and 82 dorsopathies [15]. By exploring dorsopathies' potential drug targets through genes as the exposure 83 factor, we aim to offer new perspectives and methods for the treatment and prevention of 84 dorsopathies. 85 To investigate potential pathogenic genes related to dorsopathies, we employed a comprehensive 86 analytical approach including summary-data-based Mendelian randomization (SMR) analysis, 87 colocalization analysis, genome-wide association study (GWAS), and transcriptome-wide association 88 study (TWAS). We utilized a meta-analysis dataset of cis-expression quantitative trait loci (cis-89 eQTLs) from peripheral blood samples as exposure data, along with results from the extensive 90 FinnGen database. Following preliminary analysis, we identified candidate genes and established 91 causal inference using MR methods. Additionally, we conducted mediation analysis of gene-92 mediated methylation sites to explore disease related methylation sites. 93 2 Method 94 2.1 Datasets 95 Summary-level data for the GWAS on dorsopathies were obtained from the FinnGen consortium, 96 comprising 117,411 cases and 294,770 controls. This dataset represents the most recent GWAS 97 findings and comprises the largest cohort of dorsopathies cases documented to date. The primary 98

of FinnGen is to accumulate and rigorously analyze genomic and national health register 99 data from 500,000 Finnish individuals.[16] In the context of drug development studies, we prioritized 100 cis-eQTLs that were in closer proximity to the target gene. These cis-eQTLs used for SMR analyses 101 were sourced from the eQTLGen Consortium [17] and the eQTL meta-analyses conducted on 102 peripheral blood samples from a cohort of 31,684 individuals. For the TWAS analyses of eQTL data, 103 we utilized the GTEx v8 European whole blood dataset. Given our specific emphasis on dorsopathies, 104 we meticulously extracted comprehensive eQTL results exclusively from the whole blood samples 105 within the GTEx dataset [18]. 106 107 2.2 SMR analyses 108 We conducted SMR and heterogeneity in dependent instruments (HEIDI) tests analyses on cis 109 regions using the SMR software (version 1.03) [19]. The methodologies for SMR analyses are 110 detailed in the original work. In brief, SMR analyses employs a well-established MR approach. This 111 technique employs a SNP at a prominent cis-eQTL as an instrumental variable (IV). The summary-112 level eQTL data serve as the exposure variable, and the GWAS data for a specific trait serve as the 113 outcome variable. The primary objective is to explore a potential causal or pleiotropic association, 114 wherein the same causal variant influences both gene expression and the trait. It is crucial to 115 acknowledge that the SMR method lacks the ability to distinguish between a causal association, in 116 which gene expression causally influences the trait, and a pleiotropic association, in which the same 117 SNP affects both gene expression and the trait. This limitation arises because of the single 118 instrumental variable (IV) in the MR method, which cannot differentiate between causality and 119 pleiotropy. Nevertheless, the HEIDI test can make this distinction by discerning causality and 120 pleiotropy from linkage. Linkage refers to cases in which two different SNPs in linkage 121 disequilibrium (LD) independently influence gene expression and the trait. Although less biologically 122 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 4 intriguing than causality and pleiotropy, the HEIDI test provides clarity in such scenarios. For the 123 HEIDI test, a p-value below 0.05 was considered significant, suggesting that the observed association 124 was attributable to linkage. 125 126 2.3 GWAS analyses 127 Multimarker analyses of genomic annotation (MAGMA) employs a multiple regression model to 128 assess the cumulative effect of multiple SNPs within a specific gene region (± 10 kb)[20]. The 129

panel for calculating LD was derived from Phase 3 of the 1000 Genomes European 130 population. Significance thresholds for GWAS analyses using both SMR and MAGMA were set at a 131 false discovery rate (FDR) below 0.05, corrected using the Benjamini-Hochberg method.[21] 132 133 2.4 TWAS analyses 134 We conducted validations to integrate dorsopathies GWAS and eQTL data of whole blood from 135 GTEx using the FUSION and UTMOST, widely utilized tools in prior TWAS investigations [22,23]. 136 FUSION constructs predictive models using various penalized linear models, such as GBLUP, 137 LASSO, Elastic Net, for the significant cis-heritability genes estimated from SNPs within 500 kb on 138 either side of the gene boundary. Subsequently, it selects the optimal model based on the coefficient 139 of determination (R2) calculated through a fivefold cross-validation. For UTMOST, we performed 140 repeated 49 single-tissue association tests for each tissue. TWAS significance for both single-tissue 141 analyses was determined with a Benjamini-Hochberg corrected FDR value below 0.05. 142 143 2.5 MR analyses 144 In conducting the two-sample MR analyses [24], we utilized the TwoSampleMR R package. The 145 utilization of the two-sample MR framework requires employing two distinct datasets. In this study, 146 genetic instruments, specifically cis-eQTL, served as exposures, while GWAS were utilized to 147 determine outcome traits. The MR methodology investigates the relationship between gene 148 expression and diseases or traits by employing genetic variants associated with gene expression as 149 instrumental variables (exposure) and GWAS for the outcome measures. Mendelian Randomization 150 facilitates exploration into whether alterations in gene expression causally impact diseases or traits. 151 For instruments represented by a single SNP, we employed the Wald ratio. In cases where 152 instruments consisted of multiple SNPs, we implemented the inverse-variance-weighted MR 153 approach. When selecting SNPs, the significance thresholds were defined as P < 5 × 10^(-8) for 154 genome-wide significance, with a linkage disequilibrium parameter (r^2) set to 0.1, and a genetic 155 distance set to 10 MB. 156 157 2.6 Colocalization analyses 158 Conducted colocalization analyses using the coloc package in the R software environment (version 159 4.0.3). Colocalization analyses aims to assess the potential shared causality between SNPs associated 160 with both gene expression and phenotype at a specific locus, thereby indicating the "colocalization" 161 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 5 of these genetic signals. The analyses calculates posterior probabilities (PPs) for five hypotheses: H0 162 denotes no association with either gene expression or phenotype; H1 signifies an association solely 163 with gene expression; H2 indicates an association exclusively with the phenotype; H3 suggests an 164 association with both gene expression and phenotype through independent SNPs; and H4 implies an 165 association with both gene expression and phenotype through shared causal SNPs. A substantial PP 166 for H4 (PP.H4 above 0.70) strongly suggests the presence of shared causal variants influencing both 167 gene expression and phenotype [25]. 168 169 2.7 Methylation & Mediation analysis 170 We hypothesize that methylation sites exert influence on the pathogenic risk of dorsopathies through 171 three primary pathways: a) they indirectly modulate the pathogenic risk by influencing gene 172 expression; b) the methylation sites themselves have a direct impact on the pathogenic risk of 173 dorsopathies; and c) methylation sites may affect the pathogenic risk of dorsopathies by influencing 174 other confounders. It is crucial to underscore that the overall impact of methylation on the pathogenic 175 risk of dorsopathies can be conceptualized as the combined action of these three pathways, expressed 176 as T=a+b+c, where a denotes the indirect modulation of pathogenic risk by influencing gene 177 expression, b denotes the direct effect of methylation sites on pathogenic risk, and c represents their 178 influence on pathogenic risk through impacting confounding factors. Significance tests are conducted 179 separately for T, x, y, and a. If all these parameters show significance, it can be inferred that the gene 180 plays a role in the intermediate pathway, thereby substantiating the existence of the a) pathway. 181 182 3 Result 183 3.1 SMR analyses and colocalization for Preliminary Identifying Potential Genes 184 In our initial analyses phase, we utilized the eQTLGen dataset for SMR analyses to pinpoint genes 185 significantly linked to dorsopathies. Employing the FDR method set our p-value significance 186 threshold. The eQTLGen dataset provided extensive data, yielding 15,633 candidate genes. 187 We identified 269 genes significantly associated with dorsopathies (FDR_P < 0.05). Subsequent 188 scrutiny uncovered four potential drug targets (NLRC4, CGREF1, KHK, RNF212) affecting 189 dorsopathies through various intersecting methods. For clarity, Manhattan plots illustrated SMR 190 results. Following gene identification, the HEIDI heterogeneity test (p_HEIDI > 0.05) sifted out 191 genes lacking horizontal pleiotropy. SMR locus plots further elucidated gene outcomes. 192 We proceeded with colocalization analyses to merge GWAS and blood eQTL data for genes passing 193 the SMR test. This aimed to determine if these genes colocalized with the dorsopathies trait. The 194 colocalization test results robustly supported colocalization between the trait and all four genes 195 (NLRC4: PP.H4 = 0.993, CGREF1: PP.H4 = 0.905, KHK: PP.H4 = 0.773, RNF212: PP.H4 = 0.923) 196 meeting both SMR and HEIDI test criteria. Consequently, we identify these genes as top-priority 197 candidates for subsequent functional studies. 198 199 3.2 MR analyses Validates Potential Gene Causal Relationships 200 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 6 Using cis-eQTL data from the eQTLGen Consortium, we conducted two-sample MR analyses on 201 European summary statistics of individuals with dorsopathies. The discovery cohort, consisting of 202 117,411 cases and 294,770 controls from the FinnGen cohort, underwent inverse variance weighted 203 (IVW) MR analyses to combine effect estimates from each genetic instrument. The analysis revealed 204 associations between the genetically predicted expression of 956 genes and dorsopathies risk 205 following multiple testing adjustments (FDR correction). Notably, NLRC4 (OR = 0.886, 95%CI = 206 0.855-0.917, FDR_P = 1.09e-11), CGREF1 (OR = 0.682, 95%CI = 0.585-0.794, FDR_P = 8.74e-07), 207 KHK (OR = 0.936, 95%CI = 0.921-0.951, FDR_P = 3.67e-16), and RNF212 (OR = 1.145, 95%CI = 208 1.056-1.241, FDR_P = 1.01e-03) were among the genes identified. To ensure the reliability of our 209 findings, we conducted tests for horizontal pleiotropy, which did not reveal any evidence of its 210 presence in the dataset. These additional analyses confirmed the absence of horizontal pleiotropy, 211 bolstering the robustness and validity of our MR genetics findings. 212 213 3.3 Validation 214 3.3.1 TWAS & UTMOST Validates Transcriptome-level Causal Relationships 215 In our quest to enhance causal inference and gain deeper insights into genetic associations with 216 dorsopathies, we conducted a TWAS analyses on the four genes identified in previous analyses, 217 utilizing Fusion and UTMOST software. This comprehensive approach aimed to elucidate the 218 transcriptional associations of these genes with dorsopathies and bolster the evidence supporting their 219 potential causal role. 220 The results revealed significant transcriptional associations for the four genes with dorsopathies, with 221 all colocalization probabilities (PP.H4) exceeding 0.7. Specifically, elevated transcription levels of 222 NLRC4 (Z score = -5.33068, P = 9.78E-08), CGREF1 (Z score = -4.7644, P = 1.89E-06), and KHK 223 (Z score = -4.72939, P = 2.25E-06) were significantly correlated with a decreased risk of 224 dorsopathies, whereas an increased transcription level of RNF212 (Z score = 3.751506, P = 225 0.000176) was significantly associated with an increased risk of dorsopathies. Notably, these TWAS 226 findings were consistent with those from the SMR analyses, providing robust support for the 227 transcriptional associations of these genes with dorsopathies. 228 3.3.2 MAGMA Validates Genome-level Causal Relationships 229 We also conducted a GWAS analyses using MAGMA software on the quartet of genes highlighted in 230 preceding studies. This approach aimed to illuminate the associations of these genes with 231 dorsopathies at the genome level, thereby reinforcing the evidence underpinning their potential causal 232 involvement. 233 Based on the MAGMA analyses, we identified 855 significant genes that passed the FDR test. The 234 outcomes unveiled noteworthy genome-level associations for the aforementioned genes with 235 dorsopathies. Specifically, heightened transcription levels of NLRC4 (FDR_P = 0.01851868), 236 CGREF1 (P = 0.002068503), and KHK (P = 0.009814817) exhibited significant correlations with 237 reduced risk of dorsopathies, whereas increased transcription levels of RNF212 (P = 0.00202954) 238 w ere significantly linked with elevated risk of dorsopathies. Importantly, these GWAS findings 239 corroborated those from the SMR analyses, thereby fortifying the robustness of the transcriptional 240 associations of these genes with dorsopathies. 241 242 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 7 3.4 Methylation analysis & Mediation analysis 243 We evaluated the influence of cg23387401 on RNF212 expression (β = -0.52, P = 2.08e-14) and its 244 link to dorsopathies risk (β = 0.19, P = 1.20E-07). From these findings, we identified how gene-245 mediated methylation impacts dorsopathies risk (β = -0.10, P = 1.35E-05), with an observed overall 246 effect (β = -0.10, P = 2.68E-06) indicating a 93.79% intermediate effect proportion. This suggests 247 that the rise in cg23387401 is wholly mediated by RNF212, leading to reduced dorsopathies risk. 248 Likewise, we assessed cg26638505's impact on NLRC4 expression (β = -0.81, P = 1.25E-09) and its 249 association with dorsopathies risk (β = -0.13, P = 7.87E-09). These calculations revealed the 250 influence of gene-mediated methylation sites on dorsopathies risk (β = 0.11, P = 2.87E-05), with an 251 overall effect (β = 0.09, P = 1.43E-03) and an intermediate effect proportion of approximately 252 120.00%. This suggests that cg26638505 elevation is entirely mediated by NLRC4, reducing 253 dorsopathies risk. 254 We also analyzed cg18948125's effect on NLRC4 expression (β = -0.93, P = 1.72E-08) and its 255 association with dorsopathies risk (β = -0.13, P = 7.87E-09). From this, we derived the impact of 256 gene-mediated methylation sites on dorsopathies risk (β = 0.12, P = 5.50E-05), with an overall 257 effect (β = 0.12, P = 5.59E-04) and an intermediate effect proportion of 104.41%. This suggests that 258 cg18948125 elevation primarily contributes to increased dorsopathies risk, mediated by NLRC4. 259 Similarly, cg22740783's influence on CGREF1 expression (β = 0.32, P = 7.71E-07) and its 260 association with dorsopathies risk (β = -0.38, P = 2.12E-05) was calculated. This allowed us to 261 identify the impact of gene-mediated methylation sites on dorsopathies risk (β = -0.12, P = 1.27E-262 03), with an overall effect (β = -0.13, P = 1.63E-04) and an intermediate effect proportion of 263 92.24%. Hence, we propose that the elevation in cg22740783 is fully mediated by CGREF1, leading 264 to decreased dorsopathies risk. 265 Lastly, we evaluated cg06112415's impact on CGREF1 expression (β = -0.34, P = 1.75E-06) and its 266 association with dorsopathies risk (β = -0.38, P = 2.12E-05). Through these assessments, we 267 identified the effect of gene-mediated methylation sites on dorsopathies risk (β = 0.13, P = 1.49E-268 03), with an overall effect (β = 0.12, P = 5.59E-04) and an intermediate effect proportion of 269 110.63%. This suggests that cg06112415 elevation primarily contributes to increased dorsopathies 270 risk, mediated by CGREF1. 271 272 4 Discussion 273 This study aimed to offer new perspectives and methods for the treatment and prevention of 274 dorsopathies. In our research, the identification of candidate genes and the establishment of causal 275 inference were conducted via MR methods, while the mediation analysis was used to explore disease 276 related gene-mediated methylation sites. As a result, our study revealed four genes as potential 277 therapeutic targets for dorsopathies, including NLRC4, CGREF1, KHK and RNF212. Among these 278 genes, elevated transcription levels of the NLRC4, CGREF1 and KHK were significantly associated 279 with a decreased risk of dorsopathies, whereas a raised transcription level of RNF212 was 280 significantly related a higher risk of dorsopathies. 281 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 8 In order to find the novel drug targets for dorsopathies, we employed an integrative analysis that 282 combined colocalization with MR to assess causal genes for dorsopathies. Later, we conducted tests 283 for horizontal pleiotropy to ensure the reliability of our MR genetics findings. In addition, TWAS 284 analyses on NLRC4, CGREF1, KHK and RNF212 was conducted through Fusion and UTMOST 285 software, which further illuminated the transcriptional associations of these genes with dorsopathies 286 and bolstered the evidence supporting their potential causal role. Moreover, we conducted a GWAS 287 analyses whose results were consistent with those from the SMR analyses, which reinforcing the 288 evidence underpinning the potential causal association between these four genes with dorsopathies. 289 Finally, we evaluated the influence of methylation sites on those genes expression and their links to 290 dorsopathies. The results exhibited that the rise in cg23387401 was entirely mediated by RNF212, the 291 elevations of cg26638505 and cg18948125 were primarily mediated by NLRC4, and the up-292 regulation of cg06112415 was fully mediated by CGREF1, leading to increased dorsopathies risk; 293 while the rise of cg22740783 was totally mediated by CGREF1, resulting in decreased dorsopathies 294 risk. 295 NOD-like receptors (NLRs) family caspase activation and recruitment domain-containing protein 4 296 (NLRC4) gene encodes NLRC4, which is mainly expressed in macrophages, neutrophils, dendritic 297 cells and glial cells[26]. As a member of the caspase recruitment domain-containing NLR family, 298 NLRC4 partners with NLR family of apoptosis inhibitory proteins (NAIP) to assemble 299 inflammasome complexes, which termed NAIP/NLRC4 inflammasomes[27]. Inflammasomes 300 function as intracellular multi-protein platforms that are activated by pathogen-associated molecular 301 patterns or damage-associated molecular patterns, triggering innate immune reactions and 302 inflammatory caspase-activation to defense pathogens and danger signals and maintain the 303 homeostasis[27–29]. The activation of caspases results in the proteolytic activation of the pro-304 inflammatory cytokines interleukin-1β (IL-1β ) and/or interleukin-18 (IL-18)[27]. Meanwhile, 305 activated caspase-1 also cleaves gasdermin D, whose N-terminal fragments become inserted into cell 306 membranes, thereby facilitating the release of more IL-1β and IL-18[28,29]. In particular, 307 inflammasomes were classified into canonical inflammasomes, which activating caspase-1, and non-308 canonical inflammasomes, which primarily activate caspase-4 and/or caspase-5 in human cells[29]. 309 Specifically, the NAIP/NLRC4 inflammasome serves as a as part of the innate immune response and 310 senses a range of intracellular bacteria and bacterial components in the host cell cytosol, such as S. 311 typhimurium, Legionella pneumophila, flagellin and components of the virulence-associated type III 312 secretion apparatus, thereby to mediate host defense against bacterial pathogens[26,27,30]. NAIP 313 proteins function as specific cytosolic receptors for a variety of bacterial protein ligands[31], while 314 NLRC4 interacts directly with caspase-1 to induce cell death, such as the pyroptosis of macrophage, 315 and to cause rapidly initiate inflammation and vascular fluid loss[26]. The localized effects of 316 NAIP/NLRC4 inflammasome could defend against bacterial pathogens. However, the aberrant 317 activity of this cluster also leads to multiple clinical manifestations, including macrophage activation 318 syndrome, neonatal enterocolitis and autoimmune disorders[28,31], and promotes the process of 319 some diseases, such as gliomas[29], premature rupture of membranes[26], ulcerative colitis[28], 320 rheumatoid arthritis[30] and so on. Sim et al. found that non-canonical pathway molecules of 321 NAIP/NLRC4 inflammasomes, including caspase-4, caspase-5, and N-cleaved GSDMD, were 322 significantly increased with each glioma grade[29]. Additionally, Zhu et al. reported that NLRC4 was 323 upregulated in the membranes of patients with premature rupture of fetal membranes and recruited 324 more caspase-1, which promotes the process of rupture of fetal membranes by inducing apoptosis 325 and degrading extracellular matrix[26]. The study of An et al. presented that the persistent activation 326 of NAIP/NLRC4 inflammasome induces macrophage pyroptosis mediated by caspase1-dependent 327 cleavage of GSDMD and releases proinflammatory cytokines including IL-1b and IL-18, facilitating 328 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 9 the occurrence and progression of UC[28]. In addition, research conducted by Delgado-Arévalo et al. 329 indicated that NLRC4 as an inflammasome sensor differentially upregulated in CD1c+ cDC from 330 patients with rheumatoid arthritis, and this sensor seems to be nonredundantly involved in the 331 detection of intracellular dsDNA[30]. Nevertheless, our research observed the significantly inhibitory 332 effect of up-regulated NLRC4 expression on dorsopathies. Thus, further research on the roles and 333 mechanisms of NLRC4 and inflammatsome in the progress of dorsopathies is needed. 334 Cell growth regulator with EF-hand domain 1 (CGREF1) gene encodes a novel secretory protein 335 with 2 Ca2+-binding EF-hand domains which plays an important role in eukaryotic cellular 336 signaling[32]. CGREF1 mRNA has high expressions in HCT116, H1299 and HepG2 cells while 337 expressing at low levels in other cell lines including Raji, Jurkat, BT325, PC12[32]. Mechanistically, 338 CGREF1 is regulated by p53[33], and the overexpression of CGREF1 significantly inhibits the 339 transcriptional activity of AP-1, reduces the phosphorylation of ERK (extracellular signal-regulated 340 kinases) and p38 MAPK (mitogen-activated protein kinases), and suppresses the proliferation of 341 HEK293T and HCT116 cells[32]. Furthermore, CGREF1 can decrease the percent of G2/M and S 342 phase and repress cell proliferation while overexpressing[32]. We found that the increased expression 343 of CGREF1 is significantly related to the decreased risk of dorsopathies. In the research of Xiang et 344 al., the colorectal cancer patients with higher expression of CGREF1 were found to have significantly 345 better overall survival than patients with lower expression, and the role of CGREF1 in the prognosis 346 of Early-onset colorectal cancer was reported[33]. Furthermore, Xie et al. observed that CGREF1 347 expression were high in the tissue of osteosarcoma patients while it were lowly expressed in normal 348 tissue, and speculated that CGREF1 can predict drug resistance to osteosarcoma[34]. However, the 349 biological function of CGREF1 is poorly explored, and further study for its mechanisms and roles in 350 these diseases is warranted. 351 Ketohexokinase (KHK) is an enzyme that phosphorylates fructose to produce fructose-1-phosphate 352 (F-1-P) at the first rate-limiting step in fructose metabolism[35,36]. The gene KHK encodes two 353 isoforms, KHK-C and KHK-A. Considered as the primary enzyme in fructose metabolism, KHK-C is 354 exclusively expressed in a few tissues, especially in the liver, and drives the aforementioned reaction 355 rapidly, resulting in the accumulation of uric acid and transient depletion of intracellular phosphate 356 and ATP[35–38]. In the contrary, KHK-A has a much weaker affinity and a higher Michaelis 357 constant for fructose[38]. However, KHK-A can directly phosphorylate phosphoribosyl 358 pyrophosphate synthetase 1 (PRPS1) by prevention of inhibitory nucleotide binding and facilitation 359 of ATP binding[37]. As a result, KHK-A improves nucleic acid synthesis and promotes the G1/S 360 phase transition in the cell cycle, accelerating cell proliferation[36,38]. 361 Recent studies have highlighted the importance of KHK as the key mechanism stimulating the 362 various adverse metabolic effects of fructose, such as impaired insulin sensitivity, 363 hypertriglyceridemia, and oxidative stress[35]. In addition, fructose metabolism seems to provide 364 c ancer cells with the supplementary fuel required for proliferation and metastasis in colon cancer, and 365 glioma[36,38]. Moreover, alternated expression of KHK is also related to several diseases. For 366 instance, knocked down the expression of KHK dramatically reduced fructose-induced production of 367 reactive oxidative species (ROS) in proximal tubular cells[35], indicating the association between 368 KHK and ROS. Furthermore, KHK-A activates and upregulates PRPS1, which could lead to 369 enhanced nucleic acid synthesis for tumourigenesis[38]. Kim et al. reported that most cancer cell 370 lines predominantly expressed KHK-A rather than KHK-C, and KHK-A overexpression could 371 augment cell invasion[36]. Moreover, in their study, upon fructose stimulation, KHK-A acted as a 372 nuclear protein kinase and triggered Epithelial-mesenchymal transition in breast cancer, promoting 373 breast cancer metastasis[36]. Besides, loss-of-function variants in KHK can cause essential 374 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 10 fructosuria, an autosomal recessive disease characterized by intermittent appearance of fructose in the 375 urine[39]. Yang et al. found the the overexpression of KHK-A in cell lines of oesophageal squamous 376 cell carcinoma, which may finally promote the proliferation, tumourigenicity and motility of ESCC 377 cells[38]. Similarly, KHK-A was considered to play an instrumental role in promoting de novo 378 nucleic acid synthesis and hepatocellular carcinoma development[37]. In our study, the raised 379 expression of KHK is found to be positively related to decreased risk of dorsopathies. In oder to 380 understand the mechanism of changed expression level of KHK in dorsopathies, further research are 381 needed. 382 Regulatory factor X2 (RFX2) gene is essential for maintaining normal spermatogenesis and involved 383 in spermatogenesis impairment and male infertility in mice[40]. In particular, Ring finger protein 212 384 (RNF212) gene is important for crossing over and chiasma formation during meiosis[41]. Mouse 385 Rnf212 has a central role in designating crossover sites and coupling chromosome synapsis to the 386 formation of crossover-specific recombination complexes. In humans, RNF212 has been associated 387 with variation in the genome-wide recombination rate[41]. The protein encoded by RNF212 has 388 homology to two meiotic procrossover factors: Zip3 and ZHP-3[42]. It functions to couple 389 chromosome synapsis to the formation of crossover-specific recombination complexes[43]. 390 Moreover, with the symbolic RING-finger domains, RNF212 protein is a RING-family E3-ligase for 391 Small ubiquitin-like protein (SUMO), and the latter plays an important role in assembly and 392 disassembly of synaptonemal complex by regulating protein–protein inter action during 393 meiosis[42,43,43]. In addition, RNF212 stabilizes association of a subset of MutSγ complexes with 394 recombination sites[42,44]. MutSγ complex is a kind of meiosis-specific recombination factors, 395 working as an attractive target for non-crossover/crossover differentiation[42]. It binds and stabilizes 396 DNA strand-exchange intermediates to promote both homolog synapsis and crossingover[44]. In 397 mammals, every pair of chromosomes obtains at least one crossover, while the majority of 398 recombination sites yield non-crossovers. Non-crossovers are inferred to arise from the disassembly 399 of D-loops and annealing of DNA double-strand breaks ends in a process termed synthesis dependent 400 strand annealing[42]. Designation of crossovers involves the formation of metastable joint molecules 401 and selective localization of SUMO-ligase RNF212 to a minority of recombination sites where it 402 stabilizes pertinent factors, such as MutSγ [42,44]. Furthermore, this differential RNF212-dependent 403 stabilization of key recombination proteins at precrossover sites is thought to be the basic feature of 404 crossover/non-crossover differentiation[42]. It is suggested that RNF212-mediated SUMOylation 405 may stabilizes the association of MutSγ with nascent crossover CO intermediates in a number of 406 ways, such as promoting protein-protein interactions, altering ATP binding and hydrolysis (which 407 modulate the binding and dissociation of MutSγ complexes) or antagonizing ubiquitin-dependent 408 protein turnover[41,42]. 409 Insufficient RNF212 accumulating at recombination sites can lead to crossing-over stochastically 410 fails[43]. Fujiwara et al. reported that Rnf212 knock out (KO) in mice leads to infertility of male and 411 female due to the loss of SPCs(spermatocytes) at post-anaphase stage. Moreover, crossing over is 412 diminished by ≥ 90% in Rnf212− /− mice[43]. In particular, the Rnf212 KO spermatocytes lack 413 chiasmata and exhibit depletion of spermatids and mature spermatozoa[41]. A nonsense mutation in 414 the Rnf212 gene was discovered in repro57 mutant mice, and this mutant mice exhibited male 415 infertility, arrest of spermatogenesis in meiosis, and defects in cytological markers of recombination 416 and chiasma formation, which is similar to the Rnf212 KO phenotype[41]. In humans, the rate of 417 crossing-over varies significantly between individuals, and higher maternal crossover rates have been 418 associated with greater fecundity[42]. However, in the absence of RNF212, designation of crossover 419 sites fails because no MutSγ complexes are stabilized beyond early pachynema[44]. Yu et al. 420 detected that the frequencies of allele C and the genotype CC at the rs4045481 locus in RNF212 gene 421 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 11 were significantly higher in patients with azoospermia in comparison with controls. Furthermore, 422 they reported that homozygous of allele C (genotype CC) may decrease the activity of pre-mRNA 423 due to the disappearance of the binding motifs of SRSF5, leading to the reduced expression of 424 RNF212 and influencing normal spermatogenesis, consequently increasing ther risk of 425 azoospermia[40]. Nevertheless, we found that the increased expression of RNF212 exhibited a 426 positive correlation with dorsopathies. The influence of raised expression of RNF212 is poor 427 classified, and further researches are required to illustrate the mechanism of increased RNF212 level 428 and dorsopathies. 429 To summary, our study revealed the causal associations between the genetically predicted expression 430 of four genes (NLRC4, CGREF1, KHK and RNF212) and dorsopathies risk, which offers new 431 perspectives and strategies for the improved diagnosis, treatment and prevention of dorsopathies. 432 Based on our results, we hypothesized that increased expression level of NLRC4, CGREF1 and KHK 433 are associated with decreased risk of dorsopathies, while the increased expression level of RNF212 434 has the opposite effect. Moreover, further study focusing on the mechanism of these expression 435 changes in dorsopathies are needed, and future clinical studies should be conducted. 436 437 5 Acknowledgements 438 The authors appreciate the publicly available data of the FinnGen consortium, the eQTLGen 439 consortium, the GTEx project, and the MRC IEU OpenGWAS database. 440 441 6 Conflict of interest 442 The authors declare that the research was conducted in the absence of any commercial or financial 443 relationships that could be construed as a potential conflict of interest. 444 445 7 Funding 446 The study and publishing of this article were not supported by any funding. 447 448 8 Reference 449 [1] L.B. Connelly, A. Woolf, P. Brooks, Cost-Effectiveness of Interventions for Musculoskeletal 450 Conditions, in: D.T. Jamison, J.G. Breman, A.R. Measham, G. Alleyne, M. Claeson, D.B. 451 Evans, P. Jha, A. Mills, P. Musgrove (Eds.), Disease Control Priorities in Developing Countries, 452 2nd ed., The International Bank for Reconstruction and Development / The World Bank, 453 Washington (DC), 2006. http://www.ncbi.nlm.nih.gov/books/NBK11713/ (accessed April 30, 454 2024). 455 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 12 [2] L. March, E.U.R. Smith, D.G. Hoy, M.J. Cross, L. Sanchez-Riera, F. Blyth, R. Buchbinder, T. 456 Vos, A.D. Woolf, Burden of disability due to musculoskeletal (MSK) disorders, Best Pract Res 457 Clin Rheumatol 28 (2014) 353–366. https://doi.org/10.1016/j.berh.2014.08.002. 458 [3] R. Shiri, J. Karppinen, P. Leino-Arjas, S. Solovieva, E. Viikari-Juntura, The association 459 between smoking and low back pain: a meta-analysis, Am J Med 123 (2010) 87.e7–35. 460 https://doi.org/10.1016/j.amjmed.2009.05.028. 461 [4] R. Shiri, J. Karppinen, P. Leino-Arjas, S. Solovieva, E. Viikari-Juntura, The association 462 between obesity and low back pain: a meta-analysis, Am J Epidemiol 171 (2010) 135–154. 463 https://doi.org/10.1093/aje/kwp356. 464 [5] A. Koyanagi, A. Stickley, N. Garin, M. Miret, J.L. Ayuso-Mateos, M. Leonardi, S. Koskinen, 465 A. Galas, J.M. Haro, The association between obesity and back pain in nine countries: a cross-466 sectional study, BMC Public Health 15 (2015) 123. https://doi.org/10.1186/s12889-015-1362-9. 467 [6] M. Smuck, M.-C.J. Kao, N. Brar, A. Martinez-Ith, J. Choi, C.C. Tomkins-Lane, Does physical 468 activity influence the relationship between low back pain and obesity?, Spine J 14 (2014) 209–469 216. https://doi.org/10.1016/j.spinee.2013.11.010. 470 [7] T.W. Strine, J.M. Hootman, US national prevalence and correlates of low back and neck pain 471 among adults, Arthritis Rheum 57 (2007) 656–665. https://doi.org/10.1002/art.22684. 472 [8] M. Scherer, H. Hansen, J. Gensichen, K. Mergenthal, S. Riedel-Heller, S. Weyerer, W. Maier, 473 A. Fuchs, H. Bickel, G. Schön, B. Wiese, H.-H. König, H. van den Bussche, I. Schäfer, 474 Association between multimorbidity patterns and chronic pain in elderly primary care patients: 475 a cross-sectional observational study, BMC Fam Pract 17 (2016) 68. 476 https://doi.org/10.1186/s12875-016-0468-1. 477 [9] G. Davey Smith, G. Hemani, Mendelian randomization: genetic anchors for causal inference in 478 epidemiological studies, Human Molecular Genetics 23 (2014) R89-98. 479 https://doi.org/10.1093/hmg/ddu328. 480 [10] M.K. Georgakis, D. Gill, Mendelian Randomization Studies in Stroke: Exploration of Risk 481 Factors and Drug Targets With Human Genetic Data, Stroke 52 (2021) 2992–3003. 482 https://doi.org/10.1161/STROKEAHA.120.032617. 483 [11] M.V. Holmes, M. Ala-Korpela, G.D. Smith, Mendelian randomization in cardiometabolic 484 disease: challenges in evaluating causality, Nature Reviews. Cardiology 14 (2017) 577–590. 485 https://doi.org/10.1038/nrcardio.2017.78. 486 [12] Y. Chen, X. Xu, L. Wang, K. Li, Y. Sun, L. Xiao, J. Dai, M. Huang, Y. Wang, D.W. Wang, 487 Genetic insights into therapeutic targets for aortic aneurysms: A Mendelian randomization 488 study, eBioMedicine 83 (2022) 104199. https://doi.org/10.1016/j.ebiom.2022.104199. 489 [13] J. Lin, J. Zhou, Y. Xu, Potential drug targets for multiple sclerosis identified through Mendelian 490 randomization analysis, Brain 146 (2023) 3364–3372. https://doi.org/10.1093/brain/awad070. 491 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 13 [14] Y. Wang, F. Liu, L. Sun, Y. Jia, P. Yang, D. Guo, M. Shi, A. Wang, G.-C. Chen, Y. Zhang, Z. 492 Zhu, Association between human blood metabolome and the risk of breast cancer, Breast 493 Cancer Res 25 (2023) 9. https://doi.org/10.1186/s13058-023-01609-4. 494 [15] S. Shah, A. Henry, C. Roselli, H. Lin, G. Sveinbjörnsson, G. Fatemifar, Å.K. Hedman, J.B. 495 Wilk, M.P. Morley, M.D. Chaffin, A. Helgadottir, N. Verweij, A. Dehghan, P. Almgren, C. 496 Andersson, K.G. Aragam, J. Ärnlöv, J.D. Backman, M.L. Biggs, H.L. Bloom, J. Brandimarto, 497 M.R. Brown, L. Buckbinder, D.J. Carey, D.I. Chasman, X. Chen, X. Chen, J. Chung, W. 498 Chutkow, J.P. Cook, G.E. Delgado, S. Denaxas, A.S. Doney, M. Dörr, S.C. Dudley, M.E. Dunn, 499 G. Engström, T. Esko, S.B. Felix, C. Finan, I. Ford, M. Ghanbari, S. Ghasemi, V. Giedraitis, F. 500 Giulianini, J.S. Gottdiener, S. Gross, D.F. Guðbjartsson, R. Gutmann, C.M. Haggerty, P. van 501 der Harst, C.L. Hyde, E. Ingelsson, J.W. Jukema, M. Kavousi, K.-T. Khaw, M.E. Kleber, L. 502 Køber, A. Koekemoer, C. Langenberg, L. Lind, C.M. Lindgren, B. London, L.A. Lotta, R.C. 503 Lovering, J. Luan, P. Magnusson, A. Mahajan, K.B. Margulies, W. März, O. Melander, I.R. 504 Mordi, T. Morgan, A.D. Morris, A.P. Morris, A.C. Morrison, M.W. Nagle, C.P. Nelson, A. 505 Niessner, T. Niiranen, M.L. O’Donoghue, A.T. Owens, C.N.A. Palmer, H.M. Parry, M. Perola, 506 E. Portilla-Fernandez, B.M. Psaty, Regeneron Genetics Center, K.M. Rice, P.M. Ridker, S.P.R. 507 Romaine, J.I. Rotter, P. Salo, V. Salomaa, J. van Setten, A.A. Shalaby, D.T. Smelser, N.L. 508 Smith, S. Stender, D.J. Stott, P. Svensson, M.-L. Tammesoo, K.D. Taylor, M. Teder-Laving, A. 509 Teumer, G. Thorgeirsson, U. Thorsteinsdottir, C. Torp-Pedersen, S. Trompet, B. Tyl, A.G. 510 Uitterlinden, A. Veluchamy, U. Völker, A.A. Voors, X. Wang, N.J. Wareham, D. Waterworth, 511 P.E. Weeke, R. Weiss, K.L. Wiggins, H. Xing, L.M. Yerges-Armstrong, B. Yu, F. Zannad, J.H. 512 Zhao, H. Hemingway, N.J. Samani, J.J.V. McMurray, J. Yang, P.M. Visscher, C. Newton-Cheh, 513 A. Malarstig, H. Holm, S.A. Lubitz, N. Sattar, M.V. Holmes, T.P. Cappola, F.W. Asselbergs, 514 A.D. Hingorani, K. Kuchenbaecker, P.T. Ellinor, C.C. Lang, K. Stefansson, J.G. Smith, R.S. 515 Vasan, D.I. Swerdlow, R.T. Lumbers, Genome-wide association and Mendelian randomisation 516 analysis provide insights into the pathogenesis of heart failure, Nature Communications 11 517 (2020) 163. https://doi.org/10.1038/s41467-019-13690-5. 518 [16] M.I. Kurki, J. Karjalainen, P. Palta, T.P. Sipilä, K. Kristiansson, K.M. Donner, M.P. Reeve, H. 519 Laivuori, M. Aavikko, M.A. Kaunisto, A. Loukola, E. Lahtela, H. Mattsson, P. Laiho, P. Della 520 Briotta Parolo, A.A. Lehisto, M. Kanai, N. Mars, J. Rämö, T. Kiiskinen, H.O. Heyne, K. 521 Veerapen, S. Rüeger, S. Lemmelä, W. Zhou, S. Ruotsalainen, K. Pärn, T. Hiekkalinna, S. 522 Koskelainen, T. Paajanen, V. Llorens, J. Gracia-Tabuenca, H. Siirtola, K. Reis, A.G. Elnahas, 523 B. Sun, C.N. Foley, K. Aalto-Setälä, K. Alasoo, M. Arvas, K. Auro, S. Biswas, A. Bizaki-524 Vallaskangas, O. Carpen, C.-Y. Chen, O.A. Dada, Z. Ding, M.G. Ehm, K. Eklund, M. Färkkilä, 525 H. Finucane, A. Ganna, A. Ghazal, R.R. Graham, E.M. Green, A. Hakanen, M. Hautalahti, Å.K. 526 Hedman, M. Hiltunen, R. Hinttala, I. Hovatta, X. Hu, A. Huertas-Vazquez, L. Huilaja, J. 527 Hunkapiller, H. Jacob, J.-N. Jensen, H. Joensuu, S. John, V. Julkunen, M. Jung, J. Junttila, K. 528 Kaarniranta, M. Kähönen, R. Kajanne, L. Kallio, R. Kälviäinen, J. Kaprio, FinnGen, N. 529 Kerimov, J. Kettunen, E. Kilpeläinen, T. Kilpi, K. Klinger, V.-M. Kosma, T. Kuopio, V. Kurra, 530 T. Laisk, J. Laukkanen, N. Lawless, A. Liu, S. Longerich, R. Mägi, J. Mäkelä, A. Mäkitie, A. 531 Malarstig, A. Mannermaa, J. Maranville, A. Matakidou, T. Meretoja, S.V. Mozaffari, M.E.K. 532 Niemi, M. Niemi, T. Niiranen, C.J. O Donnell, M.E. Obeidat, G. Okafo, H.M. Ollila, A. 533 Palomäki, T. Palotie, J. Partanen, D.S. Paul, M. Pelkonen, R.K. Pendergrass, S. Petrovski, A. 534 Pitkäranta, A. Platt, D. Pulford, E. Punkka, P. Pussinen, N. Raghavan, F. Rahimov, D. Rajpal, 535 N.A. Renaud, B. Riley-Gillis, R. Rodosthenous, E. Saarentaus, A. Salminen, E. Salminen, V. 536 Salomaa, J. Schleutker, R. Serpi, H.-Y. Shen, R. Siegel, K. Silander, S. Siltanen, S. Soini, H. 537 Soininen, J.H. Sul, I. Tachmazidou, K. Tasanen, P. Tienari, S. Toppila-Salmi, T. Tukiainen, T. 538 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 14 Tuomi, J.A. Turunen, J.C. Ulirsch, F. Vaura, P. Virolainen, J. Waring, D. Waterworth, R. Yang, 539 M. Nelis, A. Reigo, A. Metspalu, L. Milani, T. Esko, C. Fox, A.S. Havulinna, M. Perola, S. 540 Ripatti, A. Jalanko, T. Laitinen, T.P. Mäkelä, R. Plenge, M. McCarthy, H. Runz, M.J. Daly, A. 541 Palotie, FinnGen provides genetic insights from a well-phenotyped isolated population, Nature 542 613 (2023) 508–518. https://doi.org/10.1038/s41586-022-05473-8. 543 [17] U. Võsa, A. Claringbould, H.-J. Westra, M.J. Bonder, P. Deelen, B. Zeng, H. Kirsten, A. Saha, 544 R. Kreuzhuber, S. Yazar, H. Brugge, R. Oelen, D.H. de Vries, M.G.P. van der Wijst, S. Kasela, 545 N. Pervjakova, I. Alves, M.-J. Favé, M. Agbessi, M.W. Christiansen, R. Jansen, I. Seppälä, L. 546 Tong, A. Teumer, K. Schramm, G. Hemani, J. Verlouw, H. Yaghootkar, R. Sönmez Flitman, A. 547 Brown, V. Kukushkina, A. Kalnapenkis, S. Rüeger, E. Porcu, J. Kronberg, J. Kettunen, B. Lee, 548 F. Zhang, T. Qi, J.A. Hernandez, W. Arindrarto, F. Beutner, BIOS Consortium, i2QTL 549 Consortium, J. Dmitrieva, M. Elansary, B.P. Fairfax, M. Georges, B.T. Heijmans, A.W. Hewitt, 550 M. Kähönen, Y. Kim, J.C. Knight, P. Kovacs, K. Krohn, S. Li, M. Loeffler, U.M. Marigorta, H. 551 Mei, Y. Momozawa, M. Müller-Nurasyid, M. Nauck, M.G. Nivard, B.W.J.H. Penninx, J.K. 552 Pritchard, O.T. Raitakari, O. Rotzschke, E.P. Slagboom, C.D.A. Stehouwer, M. Stumvoll, P. 553 Sullivan, P.A.C. ’t Hoen, J. Thiery, A. Tönjes, J. van Dongen, M. van Iterson, J.H. Veldink, U. 554 Völker, R. Warmerdam, C. Wijmenga, M. Swertz, A. Andiappan, G.W. Montgomery, S. 555 Ripatti, M. Perola, Z. Kutalik, E. Dermitzakis, S. Bergmann, T. Frayling, J. van Meurs, H. 556 Prokisch, H. Ahsan, B.L. Pierce, T. Lehtimäki, D.I. Boomsma, B.M. Psaty, S.A. Gharib, P. 557 Awadalla, L. Milani, W.H. Ouwehand, K. Downes, O. Stegle, A. Battle, P.M. Visscher, J. 558 Yang, M. Scholz, J. Powell, G. Gibson, T. Esko, L. Franke, Large-scale cis- and trans-eQTL 559 analyses identify thousands of genetic loci and polygenic scores that regulate blood gene 560 expression, Nature Genetics 53 (2021) 1300–1310. https://doi.org/10.1038/s41588-021-00913-561 z. 562 [18] GTEx Consortium, The GTEx Consortium atlas of genetic regulatory effects across human 563 tissues, Science (New York, N.Y.) 369 (2020) 1318–1330. 564 https://doi.org/10.1126/science.aaz1776. 565 [19] Z. Z, Z. F, H. H, B. A, R. Mr, P. Je, M. Gw, G. Me, W. Nr, V. Pm, Y. J, Integration of summary 566 data from GWAS and eQTL studies predicts complex trait gene targets, Nature Genetics 48 567 (2016). https://doi.org/10.1038/ng.3538. 568 [20] C.A. de Leeuw, J.M. Mooij, T. Heskes, D. Posthuma, MAGMA: generalized gene-set analysis 569 of GWAS data, PLoS Comput Biol 11 (2015) e1004219. 570 https://doi.org/10.1371/journal.pcbi.1004219. 571 [21] The 1000 Genomes Project Consortium, Corresponding authors, A. Auton, G.R. Abecasis, 572 Steering committee, D.M. Altshuler, R.M. Durbin, G.R. Abecasis, D.R. Bentley, A. 573 Chakravarti, A.G. Clark, P. Donnelly, E.E. Eichler, P. Flicek, S.B. Gabriel, R.A. Gibbs, E.D. 574 Green, M.E. Hurles, B.M. Knoppers, J.O. Korbel, E.S. Lander, C. Lee, H. Lehrach, E.R. 575 Mardis, G.T. Marth, G.A. McVean, D.A. Nickerson, J.P. Schmidt, S.T. Sherry, J. Wang, R.K. 576 Wilson, Production group, Baylor College of Medicine, R.A. Gibbs, E. Boerwinkle, H. 577 Doddapaneni, Y. Han, V. Korchina, C. Kovar, S. Lee, D. Muzny, J.G. Reid, Y. Zhu, BGI-578 Shenzhen, J. Wang, Y. Chang, Q. Feng, X. Fang, X. Guo, M. Jian, H. Jiang, X. Jin, T. Lan, G. 579 Li, J. Li, Y. Li, S. Liu, X. Liu, Y. Lu, X. Ma, M. Tang, B. Wang, G. Wang, H. Wu, R. Wu, X. 580 Xu, Y. Yin, D. Zhang, W. Zhang, J. Zhao, M. Zhao, X. Zheng, Broad Institute of MIT and 581 Harvard, E.S. Lander, D.M. Altshuler, S.B. Gabriel, N. Gupta, Coriell Institute for Medical 582 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 15 Research, N. Gharani, L.H. Toji, N.P. Gerry, A.M. Resch, European Molecular Biology 583 Laboratory, European Bioinformatics Institute, P. Flicek, J. Barker, L. Clarke, L. Gil, S.E. Hunt, 584 G. Kelman, E. Kulesha, R. Leinonen, W.M. McLaren, R. Radhakrishnan, A. Roa, D. Smirnov, 585 R.E. Smith, I. Streeter, A. Thormann, I. Toneva, B. Vaughan, X. Zheng-Bradley, Illumina, D.R. 586 Bentley, R. Grocock, S. Humphray, T. James, Z. Kingsbury, Max Planck Institute for Molecular 587 Genetics, H. Lehrach, R. Sudbrak, M.W. Albrecht, V.S. Amstislavskiy, T.A. Borodina, M. 588 Lienhard, F. Mertes, M. Sultan, B. Timmermann, M.-L. Yaspo, McDonnell Genome Institute at 589 Washington University, E.R. Mardis, R.K. Wilson, L. Fulton, R. Fulton, US National Institutes 590 of Health, S.T. Sherry, V. Ananiev, Z. Belaia, D. Beloslyudtsev, N. Bouk, C. Chen, D. Church, 591 R. Cohen, C. Cook, J. Garner, T. Hefferon, M. Kimelman, C. Liu, J. Lopez, P. Meric, C. 592 O’Sullivan, Y. Ostapchuk, L. Phan, S. Ponomarov, V. Schneider, E. Shekhtman, K. Sirotkin, D. 593 Slotta, H. Zhang, University of Oxford, G.A. McVean, Wellcome Trust Sanger Institute, R.M. 594 Durbin, S. Balasubramaniam, J. Burton, P. Danecek, T.M. Keane, A. Kolb-Kokocinski, S. 595 McCarthy, J. Stalker, M. Quail, Analysis group, Affymetrix, J.P. Schmidt, C.J. Davies, J. 596 Gollub, T. Webster, B. Wong, Y. Zhan, Albert Einstein College of Medicine, A. Auton, C.L. 597 Campbell, Y. Kong, A. Marcketta, Baylor College of Medicine, R.A. Gibbs, F. Yu, L. Antunes, 598 M. Bainbridge, D. Muzny, A. Sabo, Z. Huang, BGI-Shenzhen, J. Wang, L.J.M. Coin, L. Fang, 599 X. Guo, X. Jin, G. Li, Q. Li, Y. Li, Z. Li, H. Lin, B. Liu, R. Luo, H. Shao, Y. Xie, C. Ye, C. Yu, 600 F. Zhang, H. Zheng, H. Zhu, Bilkent University, C. Alkan, E. Dal, F. Kahveci, Boston College, 601 G.T. Marth, E.P. Garrison, D. Kural, W.-P. Lee, W. Fung Leong, M. Stromberg, A.N. Ward, J. 602 Wu, M. Zhang, Broad Institute of MIT and Harvard, M.J. Daly, M.A. DePristo, R.E. Handsaker, 603 D.M. Altshuler, E. Banks, G. Bhatia, G. Del Angel, S.B. Gabriel, G. Genovese, N. Gupta, H. Li, 604 S. Kashin, E.S. Lander, S.A. McCarroll, J.C. Nemesh, R.E. Poplin, Cold Spring Harbor 605 Laboratory, S.C. Yoon, J. Lihm, V. Makarov, Cornell University, A.G. Clark, S. Gottipati, A. 606 Keinan, J.L. Rodriguez-Flores, European Molecular Biology Laboratory, J.O. Korbel, T. 607 Rausch, M.H. Fritz, A.M. Stütz, European Molecular Biology Laboratory, European 608 Bioinformatics Institute, P. Flicek, K. Beal, L. Clarke, A. Datta, J. Herrero, W.M. McLaren, 609 G.R.S. Ritchie, R.E. Smith, D. Zerbino, X. Zheng-Bradley, Harvard University, P.C. Sabeti, I. 610 Shlyakhter, S.F. Schaffner, J. Vitti, Human Gene Mutation Database, D.N. Cooper, E.V. Ball, 611 P.D. Stenson, Illumina, D.R. Bentley, B. Barnes, M. Bauer, R. Keira Cheetham, A. Cox, M. 612 Eberle, S. Humphray, S. Kahn, L. Murray, J. Peden, R. Shaw, Icahn School of Medicine at 613 Mount Sinai, E.E. Kenny, Louisiana State University, M.A. Batzer, M.K. Konkel, J.A. Walker, 614 Massachusetts General Hospital, D.G. MacArthur, M. Lek, Max Planck Institute for Molecular 615 Genetics, R. Sudbrak, V.S. Amstislavskiy, R. Herwig, McDonnell Genome Institute at 616 Washington University, E.R. Mardis, L. Ding, D.C. Koboldt, D. Larson, K. Ye, McGill 617 University, S. Gravel, National Eye Institute, NIH, A. Swaroop, E. Chew, New York Genome 618 Center, T. Lappalainen, Y. Erlich, M. Gymrek, T. Frederick Willems, Ontario Institute for 619 Cancer Research, J.T. Simpson, Pennsylvania State University, M.D. Shriver, Rutgers Cancer 620 Institute of New Jersey, J.A. Rosenfeld, Stanford University, C.D. Bustamante, S.B. 621 Montgomery, F.M. De La Vega, J.K. Byrnes, A.W. Carroll, M.K. DeGorter, P. Lacroute, B.K. 622 Maples, A.R. Martin, A. Moreno-Estrada, S.S. Shringarpure, F. Zakharia, Tel-Aviv University, 623 E. Halperin, Y. Baran, The Jackson Laboratory for Genomic Medicine, C. Lee, E. Cerveira, J. 624 Hwang, A. Malhotra, D. Plewczynski, K. Radew, M. Romanovitch, C. Zhang, Thermo Fisher 625 Scientific, F.C.L. Hyland, Translational Genomics Research Institute, D.W. Craig, A. 626 Christoforides, N. Homer, T. Izatt, A.A. Kurdoglu, S.A. Sinari, K. Squire, US National 627 Institutes of Health, S.T. Sherry, C. Xiao, University of California, San Diego, J. Sebat, D. 628 Antaki, M. Gujral, A. Noor, K. Ye, University of California, San Francisco, E.G. Burchard, 629 R.D. Hernandez, C.R. Gignoux, University of California, Santa Cruz, D. Haussler, S.J. 630 Katzman, W. James Kent, University of Chicago, B. Howie, University College London, A. 631 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 16 Ruiz-Linares, University of Geneva, E.T. Dermitzakis, University of Maryland School of 632 Medicine, S.E. Devine, University of Michigan, G.R. Abecasis, H. Min Kang, J.M. Kidd, T. 633 Blackwell, S. Caron, W. Chen, S. Emery, L. Fritsche, C. Fuchsberger, G. Jun, B. Li, R. Lyons, 634 C. Scheller, C. Sidore, S. Song, E. Sliwerska, D. Taliun, A. Tan, R. Welch, M. Kate Wing, X. 635 Zhan, University of Montréal, P. Awadalla, A. Hodgkinson, University of North Carolina at 636 Chapel Hill, Y. Li, University of North Carolina at Charlotte, X. Shi, A. Quitadamo, University 637 of Oxford, G. Lunter, G.A. McVean, J.L. Marchini, S. Myers, C. Churchhouse, O. Delaneau, A. 638 Gupta-Hinch, W. Kretzschmar, Z. Iqbal, I. Mathieson, A. Menelaou, A. Rimmer, D.K. Xifara, 639 University of Puerto Rico, T.K. Oleksyk, University of Texas Health Sciences Center at 640 Houston, Y. Fu, X. Liu, M. Xiong, University of Utah, L. Jorde, D. Witherspoon, J. Xing, 641 University of Washington, E.E. Eichler, B.L. Browning, S.R. Browning, F. Hormozdiari, P.H. 642 Sudmant, Weill Cornell Medical College, E. Khurana, Wellcome Trust Sanger Institute, R.M. 643 Durbin, M.E. Hurles, C. Tyler-Smith, C.A. Albers, Q. Ayub, S. Balasubramaniam, Y. Chen, V. 644 Colonna, P. Danecek, L. Jostins, T.M. Keane, S. McCarthy, K. Walter, Y. Xue, Yale University, 645 M.B. Gerstein, A. Abyzov, S. Balasubramanian, J. Chen, D. Clarke, Y. Fu, A.O. Harmanci, M. 646 Jin, D. Lee, J. Liu, X. Jasmine Mu, J. Zhang, Y. Zhang, Structural variation group, BGI-647 Shenzhen, Y. Li, R. Luo, H. Zhu, Bilkent University, C. Alkan, E. Dal, F. Kahveci, Boston 648 College, G.T. Marth, E.P. Garrison, D. Kural, W.-P. Lee, A.N. Ward, J. Wu, M. Zhang, Broad 649 Institute of MIT and Harvard, S.A. McCarroll, R.E. Handsaker, D.M. Altshuler, E. Banks, G. 650 Del Angel, G. Genovese, C. Hartl, H. Li, S. Kashin, J.C. Nemesh, K. Shakir, Cold Spring 651 Harbor Laboratory, S.C. Yoon, J. Lihm, V. Makarov, Cornell University, J. Degenhardt, 652 European Molecular Biology Laboratory, J.O. Korbel, M.H. Fritz, S. Meiers, B. Raeder, T. 653 Rausch, A.M. Stütz, European Molecular Biology Laboratory, European Bioinformatics 654 Institute, P. Flicek, F. Paolo Casale, L. Clarke, R.E. Smith, O. Stegle, X. Zheng-Bradley, 655 Illumina, D.R. Bentley, B. Barnes, R. Keira Cheetham, M. Eberle, S. Humphray, S. Kahn, L. 656 Murray, R. Shaw, Leiden University Medical Center, E.-W. Lameijer, Louisiana State 657 University, M.A. Batzer, M.K. Konkel, J.A. Walker, McDonnell Genome Institute at 658 Washington University, L. Ding, I. Hall, K. Ye, Stanford University, P. Lacroute, The Jackson 659 Laboratory for Genomic Medicine, C. Lee, E. Cerveira, A. Malhotra, J. Hwang, D. 660 Plewczynski, K. Radew, M. Romanovitch, C. Zhang, Translational Genomics Research 661 Institute, D.W. Craig, N. Homer, US National Institutes of Health, D. Church, C. Xiao, 662 University of California, San Diego, J. Sebat, D. Antaki, V. Bafna, J. Michaelson, K. Ye, 663 University of Maryland School of Medicine, S.E. Devine, E.J. Gardner, University of Michigan, 664 G.R. Abecasis, J.M. Kidd, R.E. Mills, G. Dayama, S. Emery, G. Jun, University of North 665 Carolina at Charlotte, X. Shi, A. Quitadamo, University of Oxford, G. Lunter, G.A. McVean, 666 University of Texas MD Anderson Cancer Center, K. Chen, X. Fan, Z. Chong, T. Chen, 667 University of Utah, D. Witherspoon, J. Xing, University of Washington, E.E. Eichler, M.J. 668 Chaisson, F. Hormozdiari, J. Huddleston, M. Malig, B.J. Nelson, P.H. Sudmant, Vanderbilt 669 University School of Medicine, N.F. Parrish, Weill Cornell Medical College, E. Khurana, 670 Wellcome Trust Sanger Institute, M.E. Hurles, B. Blackburne, S.J. Lindsay, Z. Ning, K. Walter, 671 Y. Zhang, Yale University, M.B. Gerstein, A. Abyzov, J. Chen, D. Clarke, H. Lam, X. Jasmine 672 Mu, C. Sisu, J. Zhang, Y. Zhang, Exome group, Baylor College of Medicine, R.A. Gibbs, F. 673 Yu, M. Bainbridge, D. Challis, U.S. Evani, C. Kovar, J. Lu, D. Muzny, U. Nagaswamy, J.G. 674 Reid, A. Sabo, J. Yu, BGI-Shenzhen, X. Guo, W. Li, Y. Li, R. Wu, Boston College, G.T. Marth, 675 E.P. Garrison, W. Fung Leong, A.N. Ward, Broad Institute of MIT and Harvard, G. Del Angel, 676 M.A. DePristo, S.B. Gabriel, N. Gupta, C. Hartl, R.E. Poplin, Cornell University, A.G. Clark, 677 J.L. Rodriguez-Flores, European Molecular Biology Laboratory, European Bioinformatics 678 Institute, P. Flicek, L. Clarke, R.E. Smith, X. Zheng-Bradley, Massachusetts General Hospital, 679 D.G. MacArthur, McDonnell Genome Institute at Washington University, E.R. Mardis, R. 680 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 17 Fulton, D.C. Koboldt, McGill University, S. Gravel, Stanford University, C.D. Bustamante, 681 Translational Genomics Research Institute, D.W. Craig, A. Christoforides, N. Homer, T. Izatt, 682 US National Institutes of Health, S.T. Sherry, C. Xiao, University of Geneva, E.T. Dermitzakis, 683 University of Michigan, G.R. Abecasis, H. Min Kang, University of Oxford, G.A. McVean, 684 Yale University, M.B. Gerstein, S. Balasubramanian, L. Habegger, Functional interpretation 685 group, Cornell University, H. Yu, European Molecular Biology Laboratory, European 686 Bioinformatics Institute, P. Flicek, L. Clarke, F. Cunningham, I. Dunham, D. Zerbino, X. 687 Zheng-Bradley, Harvard University, K. Lage, J. Berg Jespersen, H. Horn, Stanford University, 688 S.B. Montgomery, M.K. DeGorter, Weill Cornell Medical College, E. Khurana, Wellcome 689 Trust Sanger Institute, C. Tyler-Smith, Y. Chen, V. Colonna, Y. Xue, Yale University, M.B. 690 Gerstein, S. Balasubramanian, Y. Fu, D. Kim, Chromosome Y group, Albert Einstein College of 691 Medicine, A. Auton, A. Marcketta, American Museum of Natural History, R. Desalle, A. 692 Narechania, Arizona State University, M.A. Wilson Sayres, Boston College, E.P. Garrison, 693 Broad Institute of MIT and Harvard, R.E. Handsaker, S. Kashin, S.A. McCarroll, Cornell 694 University, J.L. Rodriguez-Flores, European Molecular Biology Laboratory, European 695 Bioinformatics Institute, P. Flicek, L. Clarke, X. Zheng-Bradley, New York Genome Center, Y. 696 Erlich, M. Gymrek, T. Frederick Willems, Stanford University, C.D. Bustamante, F.L. Mendez, 697 G. David Poznik, P.A. Underhill, The Jackson Laboratory for Genomic Medicine, C. Lee, E. 698 Cerveira, A. Malhotra, M. Romanovitch, C. Zhang, University of Michigan, G.R. Abecasis, 699 University of Queensland, L. Coin, H. Shao, Virginia Bioinformatics Institute, D. Mittelman, 700 Wellcome Trust Sanger Institute, C. Tyler-Smith, Q. Ayub, R. Banerjee, M. Cerezo, Y. Chen, 701 T.W. Fitzgerald, S. Louzada, A. Massaia, S. McCarthy, G.R. Ritchie, Y. Xue, F. Yang, Data 702 coordination center group, Baylor College of Medicine, R.A. Gibbs, C. Kovar, D. Kalra, W. 703 Hale, D. Muzny, J.G. Reid, BGI-Shenzhen, J. Wang, X. Dan, X. Guo, G. Li, Y. Li, C. Ye, X. 704 Zheng, Broad Institute of MIT and Harvard, D.M. Altshuler, European Molecular Biology 705 Laboratory, European Bioinformatics Institute, P. Flicek, L. Clarke, X. Zheng-Bradley, 706 Illumina, D.R. Bentley, A. Cox, S. Humphray, S. Kahn, Max Planck Institute for Molecular 707 Genetics, R. Sudbrak, M.W. Albrecht, M. Lienhard, McDonnell Genome Institute at 708 Washington University, D. Larson, Translational Genomics Research Institute, D.W. Craig, T. 709 Izatt, A.A. Kurdoglu, US National Institutes of Health, S.T. Sherry, C. Xiao, University of 710 California, Santa Cruz, D. Haussler, University of Michigan, G.R. Abecasis, University of 711 Oxford, G.A. McVean, Wellcome Trust Sanger Institute, R.M. Durbin, S. Balasubramaniam, 712 T.M. Keane, S. McCarthy, J. Stalker, Samples and ELSI group, A. Chakravarti, B.M. Knoppers, 713 G.R. Abecasis, K.C. Barnes, C. Beiswanger, E.G. Burchard, C.D. Bustamante, H. Cai, H. Cao, 714 R.M. Durbin, N.P. Gerry, N. Gharani, R.A. Gibbs, C.R. Gignoux, S. Gravel, B. Henn, D. Jones, 715 L. Jorde, J.S. Kaye, A. Keinan, A. Kent, A. Kerasidou, Y. Li, R. Mathias, G.A. McVean, A. 716 Moreno-Estrada, P.N. Ossorio, M. Parker, A.M. Resch, C.N. Rotimi, C.D. Royal, K. Sandoval, 717 Y. Su, R. Sudbrak, Z. Tian, S. Tishkoff, L.H. Toji, C. Tyler-Smith, M. Via, Y. Wang, H. Yang, 718 L. Yang, J. Zhu, Sample collection, British from England and Scotland (GBR), W. Bodmer, 719 Colombians in Medellín, Colombia (CLM), G. Bedoya, A. Ruiz-Linares, Han Chinese South 720 (CHS), Z. Cai, Y. Gao, J. Chu, Finnish in Finland (FIN), L. Peltonen, Iberian Populations in 721 Spain (IBS), A. Garcia-Montero, A. Orfao, Puerto Ricans in Puerto Rico (PUR), J. Dutil, J.C. 722 Martinez-Cruzado, T.K. Oleksyk, African Caribbean in Barbados (ACB), K.C. Barnes, R.A. 723 Mathias, A. Hennis, H. Watson, C. McKenzie, Bengali in Bangladesh (BEB), F. Qadri, R. 724 LaRocque, P.C. Sabeti, Chinese Dai in Xishuangbanna, China (CDX), J. Zhu, X. Deng, Esan in 725 Nigeria (ESN), P.C. Sabeti, D. Asogun, O. Folarin, C. Happi, O. Omoniwa, M. Stremlau, R. 726 Tariyal, Gambian in Western Division – Mandinka (GWD), M. Jallow, F. Sisay Joof, T. Corrah, 727 K. Rockett, D. Kwiatkowski, Indian Telugu in the UK (ITU) and Sri Lankan Tamil in the UK 728 (STU), J. Kooner, Kinh in Ho Chi Minh City, Vietnam (KHV), T. Tị nh Hiê`n, S.J. Dunstan, N. 729 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 18 Thuy Hang, Mende in Sierra Leone (MSL), R. Fonnie, R. Garry, L. Kanneh, L. Moses, P.C. 730 Sabeti, J. Schieffelin, D.S. Grant, Peruvian in Lima, Peru (PEL), C. Gallo, G. Poletti, Punjabi in 731 Lahore, Pakistan (PJL), D. Saleheen, A. Rasheed, Scientific management, L.D. Brooks, A.L. 732 Felsenfeld, J.E. McEwen, Y. Vaydylevich, E.D. Green, A. Duncanson, M. Dunn, J.A. Schloss, 733 J. Wang, H. Yang, Writing group, A. Auton, L.D. Brooks, R.M. Durbin, E.P. Garrison, H. Min 734 Kang, J.O. Korbel, J.L. Marchini, S. McCarthy, G.A. McVean, G.R. Abecasis, A global 735

for human genetic variation, Nature 526 (2015) 68–74. 736 https://doi.org/10.1038/nature15393. 737 [22] A. Gusev, A. Ko, H. Shi, G. Bhatia, W. Chung, B.W.J.H. Penninx, R. Jansen, E.J.C. de Geus, 738 D.I. Boomsma, F.A. Wright, P.F. Sullivan, E. Nikkola, M. Alvarez, M. Civelek, A.J. Lusis, T. 739 Lehtimäki, E. Raitoharju, M. Kähönen, I. Seppälä, O.T. Raitakari, J. Kuusisto, M. Laakso, A.L. 740 Price, P. Pajukanta, B. Pasaniuc, Integrative approaches for large-scale transcriptome-wide 741 association studies, Nat Genet 48 (2016) 245–252. https://doi.org/10.1038/ng.3506. 742 [23] Y. Hu, M. Li, Q. Lu, H. Weng, J. Wang, S.M. Zekavat, Z. Yu, B. Li, J. Gu, S. Muchnik, Y. Shi, 743 B.W. Kunkle, S. Mukherjee, P. Natarajan, A. Naj, A. Kuzma, Y. Zhao, P.K. Crane, H. Lu, H. 744 Zhao, A statistical framework for cross-tissue transcriptome-wide association analysis, Nat 745 Genet 51 (2019) 568–576. https://doi.org/10.1038/s41588-019-0345-7. 746 [24] G. Hemani, J. Zheng, B. Elsworth, K.H. Wade, V. Haberland, D. Baird, C. Laurin, S. Burgess, 747 J. Bowden, R. Langdon, V.Y. Tan, J. Yarmolinsky, H.A. Shihab, N.J. Timpson, D.M. Evans, C. 748 Relton, R.M. Martin, G. Davey Smith, T.R. Gaunt, P.C. Haycock, The MR-Base platform 749 supports systematic causal inference across the human phenome, eLife 7 (2018) e34408. 750 https://doi.org/10.7554/eLife.34408. 751 [25] C. Giambartolomei, D. Vukcevic, E.E. Schadt, L. Franke, A.D. Hingorani, C. Wallace, V. 752 Plagnol, Bayesian test for colocalisation between pairs of genetic association studies using 753 summary statistics, PLoS Genetics 10 (2014) e1004383. 754 https://doi.org/10.1371/journal.pgen.1004383. 755 [26] J. Zhu, C. Ma, L. Zhu, J. Li, F. Peng, L. Huang, X. Luan, A role for the NLRC4 inflammasome 756 in premature rupture of membrane, PLoS ONE 15 (2020) e0237847. 757 https://doi.org/10.1371/journal.pone.0237847. 758 [27] D. Zheng, T. Liwinski, E. Elinav, Inflammasome activation and regulation: toward a better 759 understanding of complex mechanisms, Cell Discov 6 (2020) 36. 760 https://doi.org/10.1038/s41421-020-0167-x. 761 [28] Y. An, Z. Zhai, X. Wang, Y. Ding, L. He, L. Li, Q. Mo, C. Mu, R. Xie, T. Liu, W. Zhong, B. 762 Wang, H. Cao, Targeting Desulfovibrio vulgaris flagellin-induced NAIP/NLRC4 763 inflammasome activation in macrophages attenuates ulcerative colitis, Journal of Advanced 764 Research 52 (2023) 219–232. https://doi.org/10.1016/j.jare.2023.08.008. 765 [29] J. Sim, J.W. Ahn, J. Park, Y.J. Kim, J.-Y. Jeong, J.M. Lee, K. Cho, H.J. Ahn, K.S. Sung, J.-S. 766 Moon, J.H. Moon, J. Lim, Non-canonical NLRC4 inflammasomes in astrocytes contribute to 767 glioma malignancy, Inflamm. Res. 72 (2023) 813–827. https://doi.org/10.1007/s00011-023-768 01710-6. 769 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 19 [30] C. Delgado-Arévalo, M. Calvet-Mirabent, A. Triguero-Martínez, E. Vázquez De Luis, A. 770 Benguría-Filippini, R. Largo, D. Calzada-Fraile, O. Popova, I. Sánchez-Cerrillo, I. Tsukalov, R. 771 Moreno-Vellisca, H. De La Fuente, G. Herrero-Beaumont, A. Ramiro, F. Sánchez-Madrid, S. 772 Castañeda, A. Dopazo, I. González Álvaro, E. Martin-Gayo, NLRC4-mediated activation of 773 CD1c+ DC contributes to perpetuation of synovitis in rheumatoid arthritis, JCI Insight 7 (2022) 774 e152886. https://doi.org/10.1172/jci.insight.152886. 775 [31] R.E. Vance, The NAIP/NLRC4 inflammasomes, Current Opinion in Immunology 32 (2015) 776 84–89. https://doi.org/10.1016/j.coi.2015.01.010. 777 [32] W. Deng, L. Wang, Y. Xiong, J. Li, Y. Wang, T. Shi, D. Ma, The novel secretory protein 778 CGREF1 inhibits the activation of AP-1 transcriptional activity and cell proliferation, The 779 International Journal of Biochemistry & Cell Biology 65 (2015) 32–39. 780 https://doi.org/10.1016/j.biocel.2015.05.019. 781 [33] M. Xiang, Y. Gao, Y. Zhou, M. Wang, X. Yao, A novel nomogram based on cell cycle-related 782 genes for predicting overall survival in early-onset colorectal cancer, BMC Cancer 23 (2023) 783 595. https://doi.org/10.1186/s12885-023-11075-y. 784 [34] M. Xie, H. Dai, Q. Gu, C. Xiao, H. Wang, Y. Lei, C. Wu, X. Li, B. Lin, S. Li, Identification of 785 genes contributing to cisplatin resistance in osteosarcoma cells, FEBS Open Bio 13 (2023) 164–786 173. https://doi.org/10.1002/2211-5463.13524. 787 [35] M.T. Le, M.A. Lanaspa, C.M. Cicerchi, J. Rana, J.D. Scholten, B.L. Hunter, C.J. Rivard, R.K. 788 Randolph, R.J. Johnson, Bioactivity-Guided Identification of Botanical Inhibitors of 789 Ketohexokinase, PLoS ONE 11 (2016) e0157458. 790 https://doi.org/10.1371/journal.pone.0157458. 791 [36] J. Kim, J. Kang, Y.-L. Kang, J. Woo, Y. Kim, J. Huh, J.-W. Park, Ketohexokinase-A acts as a 792 nuclear protein kinase that mediates fructose-induced metastasis in breast cancer, Nat Commun 793 11 (2020) 5436. https://doi.org/10.1038/s41467-020-19263-1. 794 [37] X. Li, X. Qian, L.-X. Peng, Y. Jiang, D.H. Hawke, Y. Zheng, Y. Xia, J.-H. Lee, G. Cote, H. 795 Wang, L. Wang, C.-N. Qian, Z. Lu, A splicing switch from ketohexokinase-C to 796 ketohexokinase-A drives hepatocellular carcinoma formation, Nat Cell Biol 18 (2016) 561–571. 797 https://doi.org/10.1038/ncb3338. 798 [38] J. Yang, S. Yang, Q. Wang, J. Pang, Y. Wang, H. Wang, X. Fu, KHK-A promotes the 799 proliferation of oesophageal squamous cell carcinoma through the up-regulation of PRPS1, 800 Arab Journal of Gastroenterology 22 (2021) 40–46. https://doi.org/10.1016/j.ajg.2020.08.007. 801 [39] J.A. Johnston, D.R. Nelson, P. Bhatnagar, S.E. Curtis, Y. Chen, J.G. MacKrell, Prevalence and 802 cardiometabolic correlates of ketohexokinase gene variants among UK Biobank participants, 803 PLoS ONE 16 (2021) e0247683. https://doi.org/10.1371/journal.pone.0247683. 804 [40] C.-H. Yu, T. Xie, R.-P. Zhang, Z.-C. A, Association of the common SNPs in RNF212, STAG3 805 and RFX2 gene with male infertility with azoospermia in Chinese population, European Journal 806 of Obstetrics & Gynecology and Reproductive Biology 221 (2018) 109–112. 807 https://doi.org/10.1016/j.ejogrb.2017.12.030. 808 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 20 [41] A. Riera-Escamilla, A. Enguita-Marruedo, D. Moreno-Mendoza, C. Chianese, E. Sleddens-809 Linkels, E. Contini, M. Benelli, A. Natali, G.M. Colpi, E. Ruiz-Castañé, M. Maggi, W.M. 810 Baarends, C. Krausz, Sequencing of a ‘mouse azoospermia’ gene panel in azoospermic men: 811 identification of RNF212 and STAG3 mutations as novel genetic causes of meiotic arrest, 812 Human Reproduction 34 (2019) 978–988. https://doi.org/10.1093/humrep/dez042. 813 [42] A. Reynolds, H. Qiao, Y. Yang, J.K. Chen, N. Jackson, K. Biswas, J.K. Holloway, F. Baudat, B. 814 De Massy, J. Wang, C. Höög, P.E. Cohen, N. Hunter, RNF212 is a dosage-sensitive regulator of 815 crossing-over during mammalian meiosis, Nat Genet 45 (2013) 269–278. 816 https://doi.org/10.1038/ng.2541. 817 [43] Y. Fujiwara, H. Matsumoto, K. Akiyama, A. Srivastava, M. Chikushi, M. Ann Handel, T. 818 Kunieda, An ENU-induced mutation in the mouse Rnf212 gene is associated with male meiotic 819 failure and infertility, REPRODUCTION 149 (2015) 67–74. https://doi.org/10.1530/REP-14-820 0122. 821 [44] H. Qiao, H.B.D. Prasada Rao, Y. Yang, J.H. Fong, J.M. Cloutier, D.C. Deacon, K.E. Nagel, 822 R.K. Swartz, E. Strong, J.K. Holloway, P.E. Cohen, J. Schimenti, J. Ward, N. Hunter, 823 Antagonistic roles of ubiquitin ligase HEI10 and SUMO ligase RNF212 regulate meiotic 824 recombination, Nat Genet 46 (2014) 194–199. https://doi.org/10.1038/ng.2858. 825 826 9 Tables 827 9.1 Table 1 828 Gene C h Positio n Z_TWA S P_TW AS P_TWA S (FDR) Tissue topSN P B_SM R SE_S MR P_SM R P_SMR (FDR) P_HEI DI N_HEI DI NLRC 4 2 322658 53 -5.33 9.8E- 08 5.1E-05 GTExv8.Whole_Bl ood rs38507 6 -0.13 0.02 7.9E- 09 9.5E-06 3.6E- 01 20 CGRE F1 2 271191 14 -4.76 1.9E- 06 5.0E-04 GTExv8.Whole_Bl ood rs23046 81 -0.38 0.09 2.1E- 05 4.6E-03 4.4E- 01 20 KHK 2 270867 46 -4.73 2.3E- 06 5.7E-04 GTExv8.Whole_Bl ood rs21190 26 -0.06 0.01 2.3E- 06 1.0E-03 2.5E- 01 20 RNF2 12 4 111356 1 3.75 1.8E- 04 0.01 GTExv8.Whole_Bl ood rs49745 91 0.19 0.04 1.2E- 07 9.4E-05 5.5E- 02 20 Table 1 TWAS / SMR/HEIDI results of the GWAS data on Dorsopathies, blood eQTL 829 data.TWAS/SMR/HEIDI results of the GWAS data on Dorsopathies, blood eQTL data. Ch 830 represents chromosome;position indicates the gene's position, Z_TWAS, P_TWAS are the Z-score 831 and p-value of the TWAS test; Tissue represents the tissue source used in the TWAS analyses. 832 topSNP represents the top SNP in the SMR analyses. P_SMR is the p-value for the SMR test; 833 B_SMR is the effect size from the SMR test; SE_SMR is the standard error of B_SMR; P_HEIDI is 834 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 21 the p-value for the HEIDI test; N_HEIDI is the number of SNPs used in the HEIDI test; FDR 835 represents false discovery rate. 836 837 9.2 Table 2 838 839 Marker Gene Methyl ation on Gene Methyl ation on Dorsop athies Gene on Dorsop athies Intermediary analyses prop mediat e d prop mediat ed se b se p b se p b se p indirec t indirec t_se indirec t_p cg0468695 3 RNF21 2 - 0.5 2 0.0 7 2.08E -14 -0.1 0.0 2 2.68E -06 0.1 9 0.0 4 1.20E -07 -0.1 0.02 1.35E-05 93.79 % 0.29 cg2663850 5 NLRC4 - 0.8 1 0.1 3 1.25E -09 0.0 9 0.0 3 1.43E -03 - 0.1 3 0.0 2 7.87E -09 0.11 0.03 2.87E-05 120.0 0 % 0.47 cg1894812 5 NLRC4 - 0.9 3 0.1 7 1.72E -08 0.1 2 0.0 3 5.59E -04 - 0.1 3 0.0 2 7.87E -09 0.12 0.03 5.50E-05 104.4 1 % 0.4 cg2274078 3 CGRE F1 0.3 2 0.0 6 7.71E -07 - 0.1 3 0.0 4 1.63E -04 - 0.3 8 0.0 9 2.12E -05 -0.12 0.04 1.27E-03 92.24 % 0.38 cg0611241 5 CGRE F1 - 0.3 4 0.0 7 1.75E -06 0.1 2 0.0 3 5.59E -04 - 0.3 8 0.0 9 2.12E -05 0.13 0.04 1.49E-03 110.6 3 % 0.47 Table 2 Main Results of Intermediary analyses. b is the effect size, se is the standard error of effect 840 size, and p is the P-value. prop_mediated represents the proportion of indirect effect in the total 841 effect. 842 843 10 Figures 844 10.1 Figure 1 845 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 22 846 Fig.1 Venn Plot of the main results of multi analysis 847 848 10.2 Figure 2 849 850 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 23 Fig.2 Methylation sites affect disease pathways. 851 10.3 Figure 3 852 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 24 853 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 25 Fig.3 Manhattan plot of the MR/SMR/FUSION analysis results using QTLs and Dorsopathies 854 GWAS summary statistics. The red dashed line represents the Bonferroni-corrected significance 855 threshold. 856 857 10.4 Figure 4 858 859 Fig.4 MR results for genes expression significantly associated with dorsopathies after FDR 860 correction. 861 862 10.5 Figure 5 863 864 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 26 Fig. 5 SMR locus plot illustrating dorsopathies at the gene locus utilizing blood eQTL data. In the 865 upper plot, each gray dot represents a SNP identified through GWAS on dorsopathies. A red 866 diamond indicates passage of the SMR test for the probe, while a solid diamond signifies successful 867 completion of both the SMR and HEIDI tests. In the lower plot, each red cross represents an SNP 868 identified in the eQTL study corresponding to each gene. The x-axis displays the genomic positions 869 (Mb, GRCh37) of SNPs, probes, and genes on the chromosome. The y-axis displays the negative 870 logarithm (base 10) of p-values for SNPs identified in the GWAS on dorsopathies, SMR test, and 871 eQTL study corresponding to each gene. 872 873 10.6 Figure 6 874 875 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint 27 Fig. 6 Locus comparison plot displays the results of colocalization analysis for SNPs associated with 876 gene expression in blood and datasets related to dorsopathies. Each dot represents a specific SNP, 877 with the color indicating its linkage disequilibrium (LD) value (r2) with the lead GWAS variant, 878 marked by a purple diamond. In the right panel, the x-axis represents genomic positions in 879 megabases (GRCh37) along the chromosome, while the y-axis shows the -log10 p-values for SNPs 880 from the dorsopathies GWAS (top) and the gene expression eQTL study (bottom). The left panel 881 compares the p-values from the dorsopathies GWAS with those from the gene expression eQTL 882 study. 883 884 . CC-BY 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 20, 2024. ; https://doi.org/10.1101/2024.05.01.24306675doi: medRxiv preprint