Coding and noncoding de novo variation converge on developmental trajectories of cortical layer 5–6 neurons in autism spectrum disorder

Coding and noncoding de novo variation converge on developmental trajectories of cortical layer 5–6 neurons in autism spectrum disorder | 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 Coding and noncoding de novo variation converge on developmental trajectories of cortical layer 5–6 neurons in autism spectrum disorder Ran Elkon, Eleina England, Sapir Shemesh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7841641/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Large-scale sequencing has greatly advanced our understanding of the genetic architecture of autism spectrum disorder (ASD). Whole-exome sequencing (WES) in thousands of trios revealed a major role for de novo protein-truncating variants (PTVs) in loss-of-function (LoF) intolerant genes and identified dozens of high-confidence ASD genes, many central to neuronal development and synaptic signaling. Whole-genome sequencing (WGS) extended these findings by uncovering noncoding contributors, including rare structural and regulatory variants that disrupt gene expression during brain development. Despite these advances, the mechanisms by which diverse mutations converge on shared neurodevelopmental pathways, and the specific cell types and developmental windows most impacted, remain incompletely understood. To address this, we integrated results from large ASD WES/WGS studies with bulk and single-nucleus transcriptomic data. Stratifying LoF-intolerant ASD genes into broadly expressed versus brain-restricted subsets revealed distinct functional roles: broadly expressed genes regulate transcription, chromatin, histone modification, and splicing, whereas brain-restricted genes function mainly in synaptic processes. Clinically, the former tend to be linked to general neurodevelopmental disorders (NDDs), while the latter are more associated with ASD-predominant phenotypes. Intersecting coding and regulatory ASD variants with human prefrontal cortex (PFC) trajectories showed that both converge on the L5-6_TLE4 neuronal lineage, but at different stages: coding de novo variants disrupt postnatal programs of neuronal maturation, while regulatory promoter variants act earlier, on fetal developmental programs. These findings highlight a framework in which distinct variant classes act within different subsets of LoF-intolerant genes, shaping ASD risk through cell type– and stage-specific mechanisms. Health sciences/Diseases/Psychiatric disorders/Autism spectrum disorders Biological sciences/Genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition characterized by difficulties in social communication and interaction, accompanied by restricted and repetitive behaviors and interests. Symptoms typically arise in early childhood and persist across the lifespan, although their severity and impact on daily functioning vary widely among affected individuals. This clinical diversity reflects the complex biology underlying ASD [ 1 ]. ASD is relatively common, with recent epidemiological studies estimating a prevalence of approximately 1–2% worldwide. Reported rates have risen over the past few decades, partly due to broadened diagnostic criteria and increased awareness, and may also reflect the true global burden [ 2 – 4 ]. Genetic studies have provided some of the strongest insights into ASD biology. Family and twin studies estimate ASD heritability at ~ 70–90% [ 5 , 6 ]. Large-scale sequencing efforts—particularly whole-exome sequencing (WES) and whole-genome sequencing (WGS)—have transformed our understanding of ASD’s genetic architecture. WES studies in thousands of parent–child trios demonstrate a key contribution of de novo protein-truncating variants (PTVs) in genes intolerant to loss-of-function (LoF) mutations to ASD risk. Landmark studies identified dozens of high-confidence ASD risk genes, now recognized as critical nodes in neuronal development and synaptic signaling [ 7 – 9 ]. Beyond de novo variants, WES has also revealed an excess burden of rare inherited damaging variants in ASD families, particularly in genes associated with neuronal communication [ 10 , 11 ]. This suggests that both de novo and inherited, rare variants act alongside common polygenic risk to shape ASD liability. More recently, WGS has further expanded these insights by uncovering noncoding contributors to ASD, including rare structural and regulatory variants in promoters and untranslated regions that disrupt gene expression during brain development [ 12 , 13 ]. A key advance from these large-scale sequencing efforts is network-based analysis showing that ASD genes, although individually diverse, converge on interconnected biological modules. Co-expression networks from human brain transcriptomes reveal that de novo ASD risk genes preferentially cluster in modules active during fetal cortical development and are enriched for neuronal differentiation and synaptic function [ 14 , 15 ]. In addition, protein–protein interaction (PPI) network analyses demonstrate that ASD genes form highly connected functional modules tied to chromatin remodeling, transcriptional regulation, synaptic signaling, and ubiquitin-mediated protein degradation [ 9 , 16 ]. Collectively, WES, WGS, and network-based analyses indicate that ASD arises from a broad spectrum of genetic perturbations—ranging from common polygenic influences to rare high-risk mutations—that converge on neuronal and developmental pathways. Elucidating this convergence has provided a framework for understanding the biological underpinnings of ASD and offers a foundation for developing mechanism-based diagnostic and therapeutic strategies. Nevertheless, our understanding of the precise mechanisms by which mutations in highly heterogeneous genes converge on shared neurodevelopmental pathways, and of the specific brain cell types and developmental windows most affected by these genetic aberrations, remains limited. In this study, we integrated results from large-scale ASD WES/WGS studies and bulk and single-nucleus transcriptomic datasets to shed further light on these outstanding challenges. Results To establish a foundation for our analyses, we first examined coding DNVs in ASD. Using data from Satterstrom et al. [ 9 ] we confirmed prior reports that such variants are significantly enriched in LoF-intolerant genes among ASD probands compared to their unaffected siblings ( Fig. S1 ). We next characterized the relationship between genes’ LoF intolerance and their breadth of expression across human tissues. Using the GTEx expression atlas, we counted the number of tissues in which each gene is expressed. We recapitulated the results of Karczewski et al. and found a clear monotonic inverse relationship between tolerance to LoF mutations and expression breadth, with highly tolerant genes expressed in fewer tissues (Fig. 1 A) [ 17 ]. Although genes in the lowest LOEUF decile generally showed the broadest expression patterns, a subset within this group displayed more restricted expression. Notably, these highly LoF-intolerant genes with restricted expression (expressed in fewer than 35 tissues) were significantly enriched across all brain regions (Fig. 1 B). For example, considering genes with restricted expression, more than 70% of the most LoF-intolerant genes (lowest LOEUF decile) are expressed in the hypothalamus, whereas fewer than 20% of the most tolerant genes (highest decile) show hypothalamic expression (Fig. 1 B; p < 9.31 × 10⁻ 48 , chi-square test). Since coding DNVs from ASD cases are enriched in LoF-intolerant genes, and as these genes can be further divided into broadly versus narrowly expressed (with tendency toward brain expression) subsets, we next asked which of these two subsets primarily drives the observed enrichment. To address this, we stratified genes in the lowest LOEUF decile (most intolerant) by expression breadth, defining broadly expressed genes as those expressed in at least 35 tissues. Repeating the enrichment analysis separately for each of the two subsets revealed that the signal was markedly stronger for the LoF-intolerant genes with narrow, brain-preferential expression (Fig. 2 A). To further dissect the functional roles of these two subsets, we examined the molecular pathways they contribute to. Previous studies have shown that ASD-associated genes are enriched for processes related to chromatin remodeling as well as synaptic development and function [ 9 , 16 ]. Here, we applied DOMINO—an algorithm we recently developed for identifying “active modules” within large protein–protein interaction networks [ 18 ]—to the set of highly LoF-intolerant genes carrying at least one DNV in the ASD cohort we analyzed, stratifying them into broadly expressed versus tissue-restricted subsets. This analysis refined the insight that the biological processes disrupted by coding ASD DNVs are driven by two key distinct components: (1) fundamental cellular functions—such as transcriptional regulation, chromatin organization, histone modification, and splicing—mediated by broadly expressed genes, and (2) synaptic processes mediated by genes with narrow, brain-specific expression ( Fig. S2 ). The functional divergence between broadly and narrowly expressed LoF-intolerant ASD genes raised the question of whether these differences also manifest at the phenotypic level. Broadly expressed genes largely govern fundamental cellular processes, whereas narrowly expressed genes are more involved in brain-specific functions such as synaptic signaling. This suggests that mutations in these two sets may yield distinct clinical outcomes. Although many genes are implicated in both ASD and neurodevelopmental disorders (NDDs) more broadly, emerging evidence suggests that functional specialization may underlie some degree of phenotypic specificity [ 19 , 20 ]. To test this idea, we examined whether broadly and narrowly expressed LoF-intolerant genes show distinct associations with ASD- versus DD-predominant gene categories. Fu et al. recently analyzed large ASD and DD cohorts and, based on enrichment for damaging DNVs, classified 466 genes as either DD- or ASD-predominant [ 21 ]. Intersecting the LoF-intolerant genes containing LoF DNVs in the ASD cohort we analyzed with this classification indicated that broadly expressed genes show a tendency toward DD-predominant scores, whereas narrowly expressed, brain-specific genes are skewed toward ASD-predominant scores. However, this difference was not statistically significant (Mann–Whitney, P = 0.168 Fig. 2 B). This observation is consistent with recent findings by Litman et al. and Russ et al., which similarly identify systematic differences among ASD-associated gene groups and their relationships to neurodevelopmental phenotypes [ 22 , 23 ]. The previous analyses revealed key functional and phenotypic differences between broadly and narrowly expressed LoF-intolerant genes targeted by DNVs in ASD cases compared to controls. To further refine disease-relevant gene sets and improve variant interpretation, we next focused on the tissue repeatedly and strongly implicated in ASD-specific phenotypes—the prefrontal cortex (PFC) [ 15 , 23 ]. Single-cell transcriptomics provides an unprecedented opportunity to resolve the specific cell types and developmental windows associated with the pathogenesis of complex diseases, such as ASD. To this end, we leveraged the comprehensive single-nucleus transcriptomics atlas of the human PFC generated by Herring et al. [ 25 ]. Using these gene sets, we extended our enrichment framework to test for DNV burden in ASD cases versus controls at high cellular and temporal resolution. In their study, Herring et al. profiled PFC samples spanning gestation through adulthood, identifying 14,984 development-associated differentially expressed genes (devDEGs). These were grouped into 14 major cell-type trajectories (4 excitatory projection neurons (PN), 6 inhibitory interneurons (IN), and 4 glial), which were further clustered into four meta developmental trends—down, transiently down, transiently up, and up. As displayed in Fig. 3 A, each meta trend encompassed distinct subsets of kinetic patterns (e.g., sustained vs. late-onset changes), together yielding 196 trajectory-specific kinetic gene sets (14 cell types x 14 kinetic patterns). We next tested these PFC developmental gene sets for enrichment of LoF-intolerant ASD genes with either broad or narrow expression, as defined in our earlier analyses. Interestingly, broadly expressed ASD genes showed marked enrichment in the down_1 kinetic pattern clusters across diverse PFC cell types (Fig. 3 B), with peak expression at the earliest developmental timepoint sampled (neonatal). In contrast, narrowly expressed ASD genes were most enriched in the transient_up1 kinetic pattern of layer 5–6 excitatory TLE4 neurons, with expression peaking in infancy (Fig. 3 C). This observation further supports the existence of two functionally distinct genetic programs contributing to ASD pathogenesis, differing in the cell types and developmental intervals in which they are active. To assess the specificity of this signal to ASD, we repeated the same analysis using gene sets associated with late-onset brain disorders. For Alzheimer’s and Parkinson’s diseases, only mild enrichments were observed, restricted to genes induced in the PFC during adulthood ( Fig. S3A–B ). As an additional negative control, genes associated with congenital heart defects (CHD) showed no significant enrichment in any gene set linked to PFC maturation ( Fig. S3C ). Together, these findings underscore both the developmental and disease-specific nature of the ASD enrichments observed in this PFC-based framework. Next, we turned to investigate the potential contribution of noncoding DNVs to ASD pathogenesis. We first evaluated the enrichment of noncoding DNVs near ASD-associated genes using a curated list of 1,152 high-confidence genes from the SFARI database, which includes genes strongly implicated in autism spectrum disorder risk. Noncoding DNVs were collected from two large-scale studies: Zhou et al., and Nakamura et al., totaling 138,060 variants, including 71,405 from 5,860 probands and 66,655 from 4,979 unaffected siblings [ 26 , 27 ]. For each gene, promoter regions of varying sizes (2.5 kb, 10 kb, 50 kb, 75 kb, and 100 kb upstream of the MANE transcription start site) were defined to capture proximal regulatory elements. Corroborating previous studies that indicated involvement of promoter DNVs in ASD etiology, our analysis revealed significant enrichment of noncoding DNVs in ASD probands compared to controls across all tested promoter sizes, with the strongest enrichment observed for larger promoter windows (75–100 kb) (Fig. 4 ). Last, we asked whether noncoding DNVs in ASD individuals preferentially affect promoters active in specific developmental trajectories and kinetic patterns in the human PFC. To this end, we assembled a joint call set of noncoding promoter DNVs by restricting Zhou et al. variants to 2-kb windows upstream of the canonical TSS (GENCODE v46) and merging them with promoter DNVs from Nakamura et al. [ 26 , 27 ]. Across studies, the cohorts comprised 1,790 probands and 1,781 unaffected siblings from the SSC (Zhou et al.) and 5,044 probands and 4,095 unaffected siblings from SPARK (Nakamura et al.). After filtering, the merged promoter set contained 21,370 noncoding DNVs—11,641 from ASD probands and 9,729 from unaffected siblings/controls. We mapped the promoter DNVs detected in ASD cases and controls to their target genes and intersected them with the Herring et al. gene clusters spanning the 14 cell-type trajectories and 14 kinetic patterns in the PFC. Interestingly, contrasting ASD cases and controls, we found that promoter DNVs from the cases were significantly enriched in a single gene cluster: the TLE4 neuron trajectory in cortical layers 5–6 with the late-down kinetic pattern (Fig. 5 ). Notably, this is the same trajectory—layers 5–6 TLE4 neurons—that showed the strongest enrichment of coding DNVs in narrowly expressed ASD genes. However, whereas the coding DNVs were enriched in genes transiently induced between infancy and childhood in this trajectory (‘transient_up1’ pattern) (Fig. 3 C), the promoter ASD DNVs were enriched in genes peaking prenatally, with expression declining thereafter (‘down-4’ pattern). This suggests that variants contributing to ASD risk converge on the same L5-6_TLE4 neuronal lineage in the PFC but target different sets of genes with distinct developmental timing—coding DNVs acting on postnatal programs of L5-6_TLE4 neuron maturation, and promoter DNVs acting on fetal programs that shape their early development. Methods ASD coding de novo variants (DNVs) . We analyzed the set of coding DNVs detected in ASD cases and healthy controls reported by Satterstrom et al. [ 9 ]. A total of 8,031 coding DNVs from 5,421 ASD-diagnosed children and 2,515 DNVs from 1,467 unaffected siblings were included in our analysis. Variants were restricted to coding regions and categorized based on their VEP annotation as either loss-of-function (LoF) or synonymous. We used the mapping of these variants to genes and assigned the variants accordingly to LOEUF deciles. The enrichment factor for ASD coding DNVs in each LOEUF decile was calculated by counting the number of DNVs from ASD cases and from controls, normalized by the number of individuals in each phenotypic group (5,421 ASD cases and 1,467 controls). The enrichment factor is the ratio between these two normalized counts. Enrichment of broadly vs. narrowly expressed ASD-implicated genes. Using the Satterstrom dataset, we collected all genes within the most constrained LOEUF decile harboring LoF DNVs in ASD cases [ 9 ]. Each LoF-implicated gene was then assigned to an expression category based on GTEx tissue breadth: broadly expressed if detected in greater than or equal to 35 tissues, otherwise narrowly expressed . The threshold of 35 was selected based on the analysis in Fig. 1 A, in which genes falling in the most constrained bin were expressed in an average of 35 tissues. The broad and narrow LoF gene sets were analyzed separately ( Table S1 ). Enrichment was evaluated within the developmental gene clusters defined by Herring et al., which group genes by tissue trajectory and kinetic trend [ 25 ]. All unique genes present in the Herring clusters defined the background universe. For each cluster and for each ASD LoF gene set (broad or narrow), we tested the overlap using a one-sided hypergeometric test. We applied Benjamini–Hochberg FDR correction across clusters and reported results as –log10(adjusted P-values). Heatmaps summarize enrichment across tissue trajectories and kinetic trends, with clusters meeting FDR 2) highlighted. ASD noncoding DNVs . We analyzed noncoding DNVs from ASD cases and healthy siblings from two resources: First, we used variant data from Zhou et al. (1,790 probands; 1,781 unaffected siblings) [ 27 ]. All variants with any coding annotation by VEP in this dataset were excluded. After filtering, we retained 122,947 autosomal non-coding DNVs for further analyses: 62,948 from 1,782 ASD cases and 59,999 from 1,737 controls. Second, we used the set of promoter DNVs defined by Nakamura et al. (5,044 ASD probands; 4,095 unaffected siblings). For the analysis of ASD noncoding promoter DNVs, we assembled a joint callset by merging the promoter DNVs from Nakamura et al. with a promoter-filtered subset from Zhou et al.[ 26 , 27 ]. Promoters were defined as the 2-kb upstream window of the canonical transcription start site (TSS) in GENCODE v46, and variants were assigned to genes by the nearest TSS. After filtering, the merged promoter set comprised 21,370 noncoding DNVs—11,641 in ASD probands and 9,729 in unaffected siblings/controls. Comparison of ASD vs. DD Predominance Scores in Broadly and Narrowly Expressed Genes To test whether expression breadth relates to phenotypic specificity, we used gene-level posterior probabilities for ASD versus DD predominance reported by Fu et al. [ 21 ]. Analyses were restricted to LoF-intolerant genes (lowest LOEUF decile) that carried coding DNVs in the ASD cohort. Genes were stratified into broadly expressed (greater than or equal to 35 tissues) and narrowly expressed (< 35 tissues) categories using GTEx-based annotations. Posterior probabilities were compared between broad and narrow gene sets using Mann-Whitney U test. Gene scores for tolerance to LoF mutations . We used LOEUF scores calculated based on gnomAD v2 [ 17 ]. All protein-coding genes were binned into LOEUF deciles. Since LOEUF scores are transcript-specific, we assigned each gene the lowest LOEUF decile among its transcripts. Genes’ breadth of expression over human tissues . We used GTEx v8 expression matrix (gene-level median TPM values in each tissue). We included in our analysis the 39 tissues with at least 100 samples. We counted for each gene the number of tissues in which it was expressed, setting a threshold of 0.3 TPM for expression. Gene sets associated with human diseases . Alzheimer’s disease genes (n = 97) were obtained from the Alzheimer’s Disease Sequencing Project (ADSP) “Top Hits” gene list, available via the NIAGADS Genomics Database (n = 97, [ 28 ], accessed March, 2025). Parkinson’s disease genes (n = 602) were curated from the Gene4PD database, a comprehensive resource of genetic associations in Parkinson’s disease [ 29 ]. Genes associated with congenital heart defects (CHD, n = 210) were obtained from Richter et al. [ 30 ]. Discussion Our analyses indicate that different classes of genetic variation contribute to ASD through partially distinct but converging biological mechanisms. Coding DNVs in ASD cases are enriched in loss-of-function–intolerant genes, which subdivide into broadly expressed versus brain-preferential subsets. Broadly expressed intolerant genes participate in fundamental cellular processes such as transcriptional regulation and chromatin remodeling, whereas narrowly expressed genes are enriched for synaptic pathways. Correspondingly, coding DNVs in the former appear to confer broad neurodevelopmental vulnerability, while those in the latter are more specifically associated with ASD-predominant phenotypes. When viewed in the context of prefrontal cortical development, these gene groups reveal complementary trajectories. Broadly expressed, loss-of-function–intolerant ASD genes are enriched in early genetic programs downregulated across development, consistent with system-wide cellular regulation, whereas narrowly expressed ASD genes are enriched in transiently upregulated programs during infancy of layer 5–6 excitatory neurons, reflecting postnatal circuit maturation. Extending this framework, promoter DNVs in ASD probands are enriched in the same deep-layer (L5–6_TLE4) trajectory but within earlier prenatal-stage programs, suggesting that coding and regulatory variants act at different developmental windows within the same neuronal lineage. These findings align with recent studies partitioning ASD risk into temporally distinct axes: Litman et al. identified a prenatal progenitor and glial axis versus a postnatal neuronal axis, while Russ et al. observed an early chromatin/transcriptional axis and a later synaptic neuronal axis, highlighting deep-layer excitatory neurons as a key site of ASD risk [ 22 , 23 ]. Together, these convergent results reinforce a prenatal–postnatal dichotomy and sharpen the view of layer 5–6 excitatory neurons as a core substrate of ASD, bridging early system-wide regulation with later neuron-specific synaptic biology. In summary, our findings provide a unified model in which variant class, gene constraint, and developmental timing together shape ASD risk. Coding and regulatory variants perturb distinct molecular programs but converge on a shared cellular lineage and functional outcome, offering a mechanistic explanation for how diverse genetic perturbations ultimately manifest in a common neurodevelopmental disorder. Declarations Acknowledgements This study was supported by a grant from Tel Aviv University Center for AI and Data Science (TAD) to R.E. R.E. is a Faculty Fellow of the Edmond J. Safra Center for Bioinformatics at Tel Aviv University. E.E. and S.S. were partially supported by fellowships from the Edmond J. Safra Center for Bioinformatics at Tel Aviv University. Conflict of interest The authors declare no conflict of interest References Lord C, Elsabbagh M, Baird G, Veenstra-Vanderweele J. Autism spectrum disorder. Lancet. 2018;392:508–520. Maenner MJ, Shaw KA, Bakian A V, Bilder DA, Durkin MS, Esler A, et al. Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years - Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2018. MMWR Surveill Summ. 2021;70:1–16. Lord C, Brugha TS, Charman T, Cusack J, Dumas G, Frazier T, et al. Autism spectrum disorder. Nat Rev Dis Primers. 2020;6:5. Zeidan J, Fombonne E, Scorah J, Ibrahim A, Durkin MS, Saxena S, et al. Global prevalence of autism: A systematic review update. Autism Res. 2022;15:778–790. Colvert E, Tick B, McEwen F, Stewart C, Curran SR, Woodhouse E, et al. Heritability of Autism Spectrum Disorder in a UK Population-Based Twin Sample. JAMA Psychiatry. 2015;72:415–423. Tick B, Bolton P, Happé F, Rutter M, Rijsdijk F. Heritability of autism spectrum disorders: a meta-analysis of twin studies. J Child Psychol Psychiatry. 2016;57:585–595. Iossifov I, O’Roak BJ, Sanders SJ, Ronemus M, Krumm N, Levy D, et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature. 2014;515:216–221. Sanders SJ, He X, Willsey AJ, Ercan-Sencicek AG, Samocha KE, Cicek AE, et al. Insights into Autism Spectrum Disorder Genomic Architecture and Biology from 71 Risk Loci. Neuron. 2015;87:1215–1233. Satterstrom FK, Kosmicki JA, Wang J, Breen MS, De Rubeis S, An J-Y, et al. Large-Scale Exome Sequencing Study Implicates Both Developmental and Functional Changes in the Neurobiology of Autism. Cell. 2020;180:568–584.e23. Krumm N, Turner TN, Baker C, Vives L, Mohajeri K, Witherspoon K, et al. Excess of rare, inherited truncating mutations in autism. Nat Genet. 2015;47:582–588. Wilfert AB, Turner TN, Murali SC, Hsieh P, Sulovari A, Wang T, et al. Recent ultra-rare inherited variants implicate new autism candidate risk genes. Nat Genet. 2021;53:1125–1134. An J-Y, Lin K, Zhu L, Werling DM, Dong S, Brand H, et al. Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder. Science. 2018;362. Brandler WM, Antaki D, Gujral M, Kleiber ML, Whitney J, Maile MS, et al. Paternally inherited cis-regulatory structural variants are associated with autism. Science. 2018;360:327–331. Willsey AJ, Sanders SJ, Li M, Dong S, Tebbenkamp AT, Muhle RA, et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell. 2013;155:997–1007. Parikshak NN, Luo R, Zhang A, Won H, Lowe JK, Chandran V, et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell. 2013;155:1008–1021. De Rubeis S, He X, Goldberg AP, Poultney CS, Samocha K, Cicek AE, et al. Synaptic, transcriptional and chromatin genes disrupted in autism. Nature. 2014;515:209–215. Karczewski KJ, Francioli LC, Tiao G, Cummings BB, Alföldi J, Wang Q, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581:434–443. Levi H, Elkon R, Shamir R. DOMINO: a network-based active module identification algorithm with reduced rate of false calls. Mol Syst Biol. 2021;17:e9593. Doshi-Velez F, Ge Y, Kohane I. Comorbidity clusters in autism spectrum disorders: an electronic health record time-series analysis. Pediatrics. 2014;133:e54-63. Sanders SJ, Sahin M, Hostyk J, Thurm A, Jacquemont S, Avillach P, et al. A framework for the investigation of rare genetic disorders in neuropsychiatry. Nat Med. 2019;25:1477–1487. Fu JM, Satterstrom FK, Peng M, Brand H, Collins RL, Dong S, et al. Rare coding variation provides insight into the genetic architecture and phenotypic context of autism. Nat Genet. 2022;54:1320–1331. Russ JB, Stone AC, Maney K, Morris LC, Wright CF, Hurst JH, et al. Cell-specific expression biases in human cortex of genes associated with neurodevelopmental disorders. Sci Rep. 2025;15:23172. Litman A, Sauerwald N, Green Snyder L, Foss-Feig J, Park CY, Hao Y, et al. Decomposition of phenotypic heterogeneity in autism reveals underlying genetic programs. Nat Genet. 2025;57:1611–1619. Velmeshev D, Schirmer L, Jung D, Haeussler M, Perez Y, Mayer S, et al. Single-cell genomics identifies cell type-specific molecular changes in autism. Science. 2019;364:685–689. Herring CA, Simmons RK, Freytag S, Poppe D, Moffet JJD, Pflueger J, et al. Human prefrontal cortex gene regulatory dynamics from gestation to adulthood at single-cell resolution. Cell. 2022;185:4428–4447.e28. Nakamura T, Ueda J, Mizuno S, Honda K, Kazuno A-A, Yamamoto H, et al. Topologically associating domains define the impact of de novo promoter variants on autism spectrum disorder risk. Cell Genomics. 2024;4:100488. Zhou J, Park CY, Theesfeld CL, Wong AK, Yuan Y, Scheckel C, et al. Whole-genome deep-learning analysis identifies contribution of noncoding mutations to autism risk. Nat Genet. 2019;51:973–980. NIAGADS Genomics Database. https://adsp.niagads.org/gvc-top-hits-list/ . Li B, Zhao G, Zhou Q, Xie Y, Wang Z, Fang Z, et al. Gene4PD: A Comprehensive Genetic Database of Parkinson’s Disease. Front Neurosci. 2021;15:679568. Richter F, Morton SU, Kim SW, Kitaygorodsky A, Wasson LK, Chen KM, et al. Genomic analyses implicate noncoding de novo variants in congenital heart disease. Nat Genet. 2020;52:769–777. Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files SupplementaryTableS1.xlsx Supplementary Table 1 SupplementaryFigureLegends.docx SuppFigures121025.pdf Supplementary Figures Cite Share Download PDF Status: Posted 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-7841641","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":545150490,"identity":"5106b522-aba3-4443-95f4-c61ce31f5796","order_by":0,"name":"Ran Elkon","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3klEQVRIie3PIQvCQBjG8WcIJsF6QeZXuCFoEPGr3DHQMswGw5IWsQviZ1gyv+NgK2dfc3bDJTGIaNCg4M5ouD9cOd4f9x7gcv1r5fTjgkobEZo/R19E2IicfxBUkd5in5Zye23zfMlrZqbQXJBX+UpLT0IudzxIdCZAmQLTonoxhqjLHsRLipBAdQUUlr+w5ql3kRs+TA7HGHRTaFsJi7qQMZdJUSOkcwVuJ6cOE1knXOuRoP1q3Ai0jC2LRYExM3+wynVQTs9938+VMqaCvEWP0wC8+Ffgcrlcri/dAS6YU+Gp4/WQAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-3440-1286","institution":"Tel Aviv University","correspondingAuthor":true,"prefix":"","firstName":"Ran","middleName":"","lastName":"Elkon","suffix":""},{"id":545150491,"identity":"38996082-d2e3-4e80-8f6f-4c6bc7523068","order_by":1,"name":"Eleina England","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Eleina","middleName":"","lastName":"England","suffix":""},{"id":545150492,"identity":"46382473-b9ca-43ba-b701-7602230bb278","order_by":2,"name":"Sapir Shemesh","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Sapir","middleName":"","lastName":"Shemesh","suffix":""}],"badges":[],"createdAt":"2025-10-12 15:25:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7841641/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7841641/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96745359,"identity":"bba50679-a2d6-4703-9330-fb787bb05081","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":277879,"visible":true,"origin":"","legend":"","description":"","filename":"correctedreffinalASDMS.docx","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/995043d019f9b8299e93db22.docx"},{"id":96914436,"identity":"8b6e5db8-9373-495b-94e5-617f864f7f75","added_by":"auto","created_at":"2025-11-27 14:05:55","extension":"json","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5339,"visible":true,"origin":"","legend":"","description":"","filename":"2025MP002505.json","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/8f6a031064f24c0784eec7c0.json"},{"id":96745365,"identity":"84465740-8ed2-4519-97d2-1983b94fc0b7","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1088058,"visible":true,"origin":"","legend":"","description":"","filename":"SuppFigures121025.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/90b97d745a0ab906368d9e1e.pdf"},{"id":96745362,"identity":"4adb07a3-ee98-450b-ae24-65a185526a6a","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16330,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTableS1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/069ac059174636ec765ecd3b.xlsx"},{"id":96914104,"identity":"b85c5238-4b2b-444e-9d7d-8bb97957973c","added_by":"auto","created_at":"2025-11-27 14:05:28","extension":"xml","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84175,"visible":true,"origin":"","legend":"","description":"","filename":"2025MP0025050enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/f5264c96dc436c6d9f129412.xml"},{"id":96745367,"identity":"b6d0a583-9170-4516-be1d-0fa3378ec7f8","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"pdf","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1077246,"visible":true,"origin":"","legend":"","description":"","filename":"Figures121025.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/e213abe8f149669f1a409991.pdf"},{"id":96745363,"identity":"e9dc7a27-7126-40bd-b469-b687a706f2f7","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80008,"visible":true,"origin":"","legend":"","description":"","filename":"2025MP0025050structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/a8ae7ed408bd7786ce64c817.xml"},{"id":96745366,"identity":"bb8d456e-715e-41a9-93e7-0418ee190e5e","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"html","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":90470,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/fc6bb426928caeb7374fa2a6.html"},{"id":96745352,"identity":"63ef90ac-fd77-48c5-8302-81599adf47cc","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":319975,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation between expression breadth of protein-coding genes across human tissues and their tolerance to LoF variants. A. \u003c/strong\u003eNumber of tissues (out of 39) in which protein-coding genes are expressed, stratified by LOEUF decile. Red dots mark means, showing decreasing expression breadth with increasing LoF tolerance.\u003cstrong\u003e B. \u003c/strong\u003eConsidering genes with restricted expression (expressed in fewer than 35 tissues), we compared, for each tissue, the proportion of genes expressed in it among those belonging to the lowest and highest LOEUF deciles. (dashed boxes in A). LoF-intolerant genes with restricted expression showed marked enrichment for expression in brain regions. P-values below 1x10\u003csup\u003e-15\u003c/sup\u003e are displayed.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/c4951048580a35b9ffbaf840.png"},{"id":96745355,"identity":"14b7d089-65db-4bb5-bbbf-85b7968b5bea","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":228558,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEnrichment of ASD coding DNVs in LoF-intolerant genes with broad and narrow expression patterns. A\u003c/strong\u003e. Genes were divided within each LOEUF decile into broadly (greater than or equal to 35 tissues) and narrowly (\u0026lt;35 tissues) expressed categories. Stronger enrichment for ASD coding DNVs was observed in narrowly expressed LoF-intolerant genes. \u003cstrong\u003eB.\u003c/strong\u003e LoF-intolerant genes (lowest LOEUF decile) carrying coding DNVs were further stratified into broad vs. narrow expression subsets and compared for ASD- vs. DD-predominant probabilities [21]. Narrowly expressed genes showed a trend toward higher ASD-predominant probabilities, though not statistically significant (Mann–Whitney U test, P = 0.168).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/57839ec939174014ed8a5a2d.png"},{"id":96914370,"identity":"986b819c-7723-4e31-a86f-c3447b995169","added_by":"auto","created_at":"2025-11-27 14:05:49","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1084634,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation of ASD genes with cell-type maturation trajectories and kinetic patterns in the human prefrontal cortex. A.\u003c/strong\u003e Developmental kinetic programs were defined by Herring et al. [25] for 14,984 genes across 14 cell-type trajectories in the PFC, spanning gestation to adulthood. Furthermore, 14 kinetic sub-trends were detected and grouped into four meta-trends (Down, Transient Down, Transient Up, Up), each with distinct temporal patterns (schematic). The 14 maturation trajectories include: astrocytes (A_Astro), inhibitory neurons ID2-positive (B_ID2), excitatory neurons L2–3 CUX2 (C_L2-3_CUX2), excitatory neurons L4 RORB (D_L4_RORB), excitatory neurons L5–6 THEMIS (E_L5-6_THEMIS), excitatory neurons L5–6 TLE4 (F_L5-6_TLE4), inhibitory neurons LAMP5/NOS1 (G_LAMP5_NOS1), microglia (H_Micro), oligodendrocytes (I_Oligo), oligodendrocyte precursor cells (J_OPC), inhibitory neurons parvalbumin-positive (K_PV), excitatory pyramidal neurons SCUBE3 (L_PY_SCUBE3), inhibitory neurons somatostatin-positive (M_SST), and inhibitory neurons VIP-positive (N_VIP). \u003cstrong\u003eB. \u003c/strong\u003eEnrichment of broadly expressed LoF-intolerant genes (lowest LOEUF decile) carrying predicted LoF DNVs in ASD probands across clusters. \u003cstrong\u003eC. \u003c/strong\u003eSame analysis as in B, but for narrowly expressed genes (\u0026lt;35 GTEx tissues). P-values were computed by one-sided hypergeometric tests and FDR-corrected across 196 tests\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/0071f25bf9aa8e3a4ba53281.png"},{"id":96914020,"identity":"2efc2928-bb5e-4c99-b521-9fd5487fd87f","added_by":"auto","created_at":"2025-11-27 14:05:10","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":71550,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEnrichment of ASD \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003ede novo\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e noncoding variants in promoter regions of ASD genes\u003c/strong\u003e. Enrichment of noncoding DNVs in genomic intervals of varying lengths (2.5 kb, 10 kb, 50 kb, 75 kb, and 100 kb) upstream of the MANE transcription start sites (TSSs) of 1,152 SFARI ASD risk genes. Analysis was performed using all noncoding variants from probands and unaffected siblings. p-values were calculated using hypergeometric test (contrasting \u003cem\u003ede novo\u003c/em\u003enoncoding variants in the ASD and control groups) within each promoter window.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/327a80e7ee8674b1a73485d8.png"},{"id":96914758,"identity":"62edb039-d92b-47cf-98e4-e5f73d67bd9b","added_by":"auto","created_at":"2025-11-27 14:06:21","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":900938,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEnrichment of ASD promoter DNVs in a specific neuronal maturation trajectory\u003c/strong\u003e. \u003cem\u003eDe novo\u003c/em\u003e promoter variants from ASD cases and controls [26, 27] were mapped to canonical promoters and intersected with Herring’s developmental clusters. Enrichment was evaluated at the variant level, using the closest gene TSS from VEP annotation. This quantified enrichment as the number of variants mapping to each gene in cases versus controls, assessed by hypergeometric tests and corrected for multiple comparisons (196 tests) with FDR. Values represent –log10(FDR-adjusted P-values).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/5aeb8e85fd2fa3cef91332fb.png"},{"id":100547075,"identity":"8e4b664f-4037-4a89-88ec-ddafa2e5677b","added_by":"auto","created_at":"2026-01-19 08:14:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3348325,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/f7e7a2e2-18fb-4c85-a08d-8077e5e0cb94.pdf"},{"id":96745353,"identity":"2cc75264-35c0-4ade-8b70-2591b52dc045","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":16330,"visible":true,"origin":"","legend":"Supplementary Table 1","description":"","filename":"SupplementaryTableS1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/c73ad928c14dc725396637e9.xlsx"},{"id":96914138,"identity":"669e2c0b-f251-4064-a5c3-7f12e2e62431","added_by":"auto","created_at":"2025-11-27 14:05:31","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":23327,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigureLegends.docx","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/0b9e58f8405545c9322c2670.docx"},{"id":96745357,"identity":"ee35a7f3-d894-415d-a26b-a9b0e88838c5","added_by":"auto","created_at":"2025-11-25 15:48:42","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1088058,"visible":true,"origin":"","legend":"Supplementary Figures","description":"","filename":"SuppFigures121025.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7841641/v1/a8600bc568bdfe1e836bdb34.pdf"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Coding and noncoding de novo variation converge on developmental trajectories of cortical layer 5–6 neurons in autism spectrum disorder","fulltext":[{"header":"Background","content":"\u003cp\u003eAutism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition characterized by difficulties in social communication and interaction, accompanied by restricted and repetitive behaviors and interests. Symptoms typically arise in early childhood and persist across the lifespan, although their severity and impact on daily functioning vary widely among affected individuals. This clinical diversity reflects the complex biology underlying ASD [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. ASD is relatively common, with recent epidemiological studies estimating a prevalence of approximately 1\u0026ndash;2% worldwide. Reported rates have risen over the past few decades, partly due to broadened diagnostic criteria and increased awareness, and may also reflect the true global burden [\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eGenetic studies have provided some of the strongest insights into ASD biology. Family and twin studies estimate ASD heritability at ~\u0026thinsp;70\u0026ndash;90% [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Large-scale sequencing efforts\u0026mdash;particularly whole-exome sequencing (WES) and whole-genome sequencing (WGS)\u0026mdash;have transformed our understanding of ASD\u0026rsquo;s genetic architecture. WES studies in thousands of parent\u0026ndash;child trios demonstrate a key contribution of \u003cem\u003ede novo\u003c/em\u003e protein-truncating variants (PTVs) in genes intolerant to loss-of-function (LoF) mutations to ASD risk. Landmark studies identified dozens of high-confidence ASD risk genes, now recognized as critical nodes in neuronal development and synaptic signaling [\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eBeyond \u003cem\u003ede novo\u003c/em\u003e variants, WES has also revealed an excess burden of rare inherited damaging variants in ASD families, particularly in genes associated with neuronal communication [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. This suggests that both \u003cem\u003ede novo\u003c/em\u003e and inherited, rare variants act alongside common polygenic risk to shape ASD liability. More recently, WGS has further expanded these insights by uncovering noncoding contributors to ASD, including rare structural and regulatory variants in promoters and untranslated regions that disrupt gene expression during brain development [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eA key advance from these large-scale sequencing efforts is network-based analysis showing that ASD genes, although individually diverse, converge on interconnected biological modules. Co-expression networks from human brain transcriptomes reveal that \u003cem\u003ede novo\u003c/em\u003e ASD risk genes preferentially cluster in modules active during fetal cortical development and are enriched for neuronal differentiation and synaptic function [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. In addition, protein\u0026ndash;protein interaction (PPI) network analyses demonstrate that ASD genes form highly connected functional modules tied to chromatin remodeling, transcriptional regulation, synaptic signaling, and ubiquitin-mediated protein degradation [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eCollectively, WES, WGS, and network-based analyses indicate that ASD arises from a broad spectrum of genetic perturbations\u0026mdash;ranging from common polygenic influences to rare high-risk mutations\u0026mdash;that converge on neuronal and developmental pathways. Elucidating this convergence has provided a framework for understanding the biological underpinnings of ASD and offers a foundation for developing mechanism-based diagnostic and therapeutic strategies. Nevertheless, our understanding of the precise mechanisms by which mutations in highly heterogeneous genes converge on shared neurodevelopmental pathways, and of the specific brain cell types and developmental windows most affected by these genetic aberrations, remains limited. In this study, we integrated results from large-scale ASD WES/WGS studies and bulk and single-nucleus transcriptomic datasets to shed further light on these outstanding challenges.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eTo establish a foundation for our analyses, we first examined coding DNVs in ASD. Using data from Satterstrom et al. [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] we confirmed prior reports that such variants are significantly enriched in LoF-intolerant genes among ASD probands compared to their unaffected siblings (\u003cb\u003eFig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). We next characterized the relationship between genes’ LoF intolerance and their breadth of expression across human tissues. Using the GTEx expression atlas, we counted the number of tissues in which each gene is expressed. We recapitulated the results of Karczewski et al. and found a clear monotonic inverse relationship between tolerance to LoF mutations and expression breadth, with highly tolerant genes expressed in fewer tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Although genes in the lowest LOEUF decile generally showed the broadest expression patterns, a subset within this group displayed more restricted expression. Notably, these highly LoF-intolerant genes with restricted expression (expressed in fewer than 35 tissues) were significantly enriched across all brain regions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). For example, considering genes with restricted expression, more than 70% of the most LoF-intolerant genes (lowest LOEUF decile) are expressed in the hypothalamus, whereas fewer than 20% of the most tolerant genes (highest decile) show hypothalamic expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB; p \u0026lt; 9.31 × 10⁻\u003csup\u003e48\u003c/sup\u003e, chi-square test).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eSince coding DNVs from ASD cases are enriched in LoF-intolerant genes, and as these genes can be further divided into broadly versus narrowly expressed (with tendency toward brain expression) subsets, we next asked which of these two subsets primarily drives the observed enrichment. To address this, we stratified genes in the lowest LOEUF decile (most intolerant) by expression breadth, defining broadly expressed genes as those expressed in at least 35 tissues. Repeating the enrichment analysis separately for each of the two subsets revealed that the signal was markedly stronger for the LoF-intolerant genes with narrow, brain-preferential expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). To further dissect the functional roles of these two subsets, we examined the molecular pathways they contribute to. Previous studies have shown that ASD-associated genes are enriched for processes related to chromatin remodeling as well as synaptic development and function [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Here, we applied DOMINO—an algorithm we recently developed for identifying “active modules” within large protein–protein interaction networks [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]—to the set of highly LoF-intolerant genes carrying at least one DNV in the ASD cohort we analyzed, stratifying them into broadly expressed versus tissue-restricted subsets. This analysis refined the insight that the biological processes disrupted by coding ASD DNVs are driven by two key distinct components: (1) fundamental cellular functions—such as transcriptional regulation, chromatin organization, histone modification, and splicing—mediated by broadly expressed genes, and (2) synaptic processes mediated by genes with narrow, brain-specific expression (\u003cb\u003eFig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e\u003c/b\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe functional divergence between broadly and narrowly expressed LoF-intolerant ASD genes raised the question of whether these differences also manifest at the phenotypic level. Broadly expressed genes largely govern fundamental cellular processes, whereas narrowly expressed genes are more involved in brain-specific functions such as synaptic signaling. This suggests that mutations in these two sets may yield distinct clinical outcomes. Although many genes are implicated in both ASD and neurodevelopmental disorders (NDDs) more broadly, emerging evidence suggests that functional specialization may underlie some degree of phenotypic specificity [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. To test this idea, we examined whether broadly and narrowly expressed LoF-intolerant genes show distinct associations with ASD- versus DD-predominant gene categories. Fu et al. recently analyzed large ASD and DD cohorts and, based on enrichment for damaging DNVs, classified 466 genes as either DD- or ASD-predominant [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Intersecting the LoF-intolerant genes containing LoF DNVs in the ASD cohort we analyzed with this classification indicated that broadly expressed genes show a tendency toward DD-predominant scores, whereas narrowly expressed, brain-specific genes are skewed toward ASD-predominant scores. However, this difference was not statistically significant (Mann–Whitney, \u003cem\u003eP\u003c/em\u003e = 0.168 Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). This observation is consistent with recent findings by Litman et al. and Russ et al., which similarly identify systematic differences among ASD-associated gene groups and their relationships to neurodevelopmental phenotypes [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe previous analyses revealed key functional and phenotypic differences between broadly and narrowly expressed LoF-intolerant genes targeted by DNVs in ASD cases compared to controls. To further refine disease-relevant gene sets and improve variant interpretation, we next focused on the tissue repeatedly and strongly implicated in ASD-specific phenotypes—the prefrontal cortex (PFC) [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Single-cell transcriptomics provides an unprecedented opportunity to resolve the specific cell types and developmental windows associated with the pathogenesis of complex diseases, such as ASD. To this end, we leveraged the comprehensive single-nucleus transcriptomics atlas of the human PFC generated by Herring et al. [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Using these gene sets, we extended our enrichment framework to test for DNV burden in ASD cases versus controls at high cellular and temporal resolution. In their study, Herring et al. profiled PFC samples spanning gestation through adulthood, identifying 14,984 development-associated differentially expressed genes (devDEGs). These were grouped into 14 major cell-type trajectories (4 excitatory projection neurons (PN), 6 inhibitory interneurons (IN), and 4 glial), which were further clustered into four meta developmental trends—down, transiently down, transiently up, and up. As displayed in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, each meta trend encompassed distinct subsets of kinetic patterns (e.g., sustained vs. late-onset changes), together yielding 196 trajectory-specific kinetic gene sets (14 cell types x 14 kinetic patterns). We next tested these PFC developmental gene sets for enrichment of LoF-intolerant ASD genes with either broad or narrow expression, as defined in our earlier analyses. Interestingly, broadly expressed ASD genes showed marked enrichment in the \u003cem\u003edown_1\u003c/em\u003e kinetic pattern clusters across diverse PFC cell types (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), with peak expression at the earliest developmental timepoint sampled (neonatal). In contrast, narrowly expressed ASD genes were most enriched in the \u003cem\u003etransient_up1\u003c/em\u003e kinetic pattern of layer 5–6 excitatory TLE4 neurons, with expression peaking in infancy (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). This observation further supports the existence of two functionally distinct genetic programs contributing to ASD pathogenesis, differing in the cell types and developmental intervals in which they are active. To assess the specificity of this signal to ASD, we repeated the same analysis using gene sets associated with late-onset brain disorders. For Alzheimer’s and Parkinson’s diseases, only mild enrichments were observed, restricted to genes induced in the PFC during adulthood (\u003cb\u003eFig. S3A–B\u003c/b\u003e). As an additional negative control, genes associated with congenital heart defects (CHD) showed no significant enrichment in any gene set linked to PFC maturation (\u003cb\u003eFig. S3C\u003c/b\u003e). Together, these findings underscore both the developmental and disease-specific nature of the ASD enrichments observed in this PFC-based framework.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eNext, we turned to investigate the potential contribution of noncoding DNVs to ASD pathogenesis. We first evaluated the enrichment of noncoding DNVs near ASD-associated genes using a curated list of 1,152 high-confidence genes from the SFARI database, which includes genes strongly implicated in autism spectrum disorder risk. Noncoding DNVs were collected from two large-scale studies: Zhou et al., and Nakamura et al., totaling 138,060 variants, including 71,405 from 5,860 probands and 66,655 from 4,979 unaffected siblings [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. For each gene, promoter regions of varying sizes (2.5 kb, 10 kb, 50 kb, 75 kb, and 100 kb upstream of the MANE transcription start site) were defined to capture proximal regulatory elements. Corroborating previous studies that indicated involvement of promoter DNVs in ASD etiology, our analysis revealed significant enrichment of noncoding DNVs in ASD probands compared to controls across all tested promoter sizes, with the strongest enrichment observed for larger promoter windows (75–100 kb) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Last, we asked whether noncoding DNVs in ASD individuals preferentially affect promoters active in specific developmental trajectories and kinetic patterns in the human PFC. To this end, we assembled a joint call set of noncoding promoter DNVs by restricting Zhou et al. variants to 2-kb windows upstream of the canonical TSS (GENCODE v46) and merging them with promoter DNVs from Nakamura et al. [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Across studies, the cohorts comprised 1,790 probands and 1,781 unaffected siblings from the SSC (Zhou et al.) and 5,044 probands and 4,095 unaffected siblings from SPARK (Nakamura et al.). After filtering, the merged promoter set contained 21,370 noncoding DNVs—11,641 from ASD probands and 9,729 from unaffected siblings/controls. We mapped the promoter DNVs detected in ASD cases and controls to their target genes and intersected them with the Herring et al. gene clusters spanning the 14 cell-type trajectories and 14 kinetic patterns in the PFC. Interestingly, contrasting ASD cases and controls, we found that promoter DNVs from the cases were significantly enriched in a single gene cluster: the TLE4 neuron trajectory in cortical layers 5–6 with the late-down kinetic pattern (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). Notably, this is the same trajectory—layers 5–6 TLE4 neurons—that showed the strongest enrichment of coding DNVs in narrowly expressed ASD genes. However, whereas the coding DNVs were enriched in genes transiently induced between infancy and childhood in this trajectory (‘transient_up1’ pattern) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC), the promoter ASD DNVs were enriched in genes peaking prenatally, with expression declining thereafter (‘down-4’ pattern). This suggests that variants contributing to ASD risk converge on the same L5-6_TLE4 neuronal lineage in the PFC but target different sets of genes with distinct developmental timing—coding DNVs acting on postnatal programs of L5-6_TLE4 neuron maturation, and promoter DNVs acting on fetal programs that shape their early development.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eASD coding\u003c/span\u003e \u003cspan type=\"ItalicUnderline\" class=\"ItalicUnderline\" name=\"Emphasis\"\u003ede novo\u003c/span\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003evariants (DNVs)\u003c/span\u003e. We analyzed the set of coding DNVs detected in ASD cases and healthy controls reported by Satterstrom et al. [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. A total of 8,031 coding DNVs from 5,421 ASD-diagnosed children and 2,515 DNVs from 1,467 unaffected siblings were included in our analysis. Variants were restricted to coding regions and categorized based on their VEP annotation as either loss-of-function (LoF) or synonymous. We used the mapping of these variants to genes and assigned the variants accordingly to LOEUF deciles. The enrichment factor for ASD coding DNVs in each LOEUF decile was calculated by counting the number of DNVs from ASD cases and from controls, normalized by the number of individuals in each phenotypic group (5,421 ASD cases and 1,467 controls). The enrichment factor is the ratio between these two normalized counts.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eEnrichment of broadly vs. narrowly expressed ASD-implicated genes.\u003c/span\u003e Using the Satterstrom dataset, we collected all genes within the most constrained LOEUF decile harboring LoF DNVs in ASD cases [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Each LoF-implicated gene was then assigned to an expression category based on GTEx tissue breadth: \u003cem\u003ebroadly expressed\u003c/em\u003e if detected in greater than or equal to 35 tissues, otherwise \u003cem\u003enarrowly expressed\u003c/em\u003e. The threshold of 35 was selected based on the analysis in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, in which genes falling in the most constrained bin were expressed in an average of 35 tissues. The broad and narrow LoF gene sets were analyzed separately (\u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e).\u003c/p\u003e\u003cp\u003eEnrichment was evaluated within the developmental gene clusters defined by Herring et al., which group genes by tissue trajectory and kinetic trend [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. All unique genes present in the Herring clusters defined the background universe. For each cluster and for each ASD LoF gene set (broad or narrow), we tested the overlap using a one-sided hypergeometric test. We applied Benjamini–Hochberg FDR correction across clusters and reported results as –log10(adjusted P-values). Heatmaps summarize enrichment across tissue trajectories and kinetic trends, with clusters meeting FDR \u0026lt; 0.01 (i.e., –log10 \u0026gt; 2) highlighted.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eASD noncoding DNVs\u003c/span\u003e. We analyzed noncoding DNVs from ASD cases and healthy siblings from two resources: First, we used variant data from Zhou et al. (1,790 probands; 1,781 unaffected siblings) [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. All variants with any coding annotation by VEP in this dataset were excluded. After filtering, we retained 122,947 autosomal non-coding DNVs for further analyses: 62,948 from 1,782 ASD cases and 59,999 from 1,737 controls. Second, we used the set of promoter DNVs defined by Nakamura et al. (5,044 ASD probands; 4,095 unaffected siblings). For the analysis of ASD noncoding promoter DNVs, we assembled a joint callset by merging the promoter DNVs from Nakamura et al. with a promoter-filtered subset from Zhou et al.[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Promoters were defined as the 2-kb upstream window of the canonical transcription start site (TSS) in GENCODE v46, and variants were assigned to genes by the nearest TSS. After filtering, the merged promoter set comprised 21,370 noncoding DNVs—11,641 in ASD probands and 9,729 in unaffected siblings/controls.\u003c/p\u003e\n\u003ch3\u003eComparison of ASD vs. DD Predominance Scores in Broadly and Narrowly Expressed Genes\u003c/h3\u003e\n\u003cp\u003eTo test whether expression breadth relates to phenotypic specificity, we used gene-level posterior probabilities for ASD versus DD predominance reported by Fu et al. [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Analyses were restricted to LoF-intolerant genes (lowest LOEUF decile) that carried coding DNVs in the ASD cohort. Genes were stratified into broadly expressed (greater than or equal to 35 tissues) and narrowly expressed (\u0026lt;\u0026thinsp;35 tissues) categories using GTEx-based annotations. Posterior probabilities were compared between broad and narrow gene sets using Mann-Whitney U test.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGene scores for tolerance to LoF mutations\u003c/span\u003e. We used LOEUF scores calculated based on gnomAD v2 [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. All protein-coding genes were binned into LOEUF deciles. Since LOEUF scores are transcript-specific, we assigned each gene the lowest LOEUF decile among its transcripts.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGenes\u0026rsquo; breadth of expression over human tissues\u003c/span\u003e. We used GTEx v8 expression matrix (gene-level median TPM values in each tissue). We included in our analysis the 39 tissues with at least 100 samples. We counted for each gene the number of tissues in which it was expressed, setting a threshold of 0.3 TPM for expression.\u003c/p\u003e\u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGene sets associated with human diseases\u003c/span\u003e. Alzheimer\u0026rsquo;s disease genes (n\u0026thinsp;=\u0026thinsp;97) were obtained from the Alzheimer\u0026rsquo;s Disease Sequencing Project (ADSP) \u0026ldquo;Top Hits\u0026rdquo; gene list, available via the NIAGADS Genomics Database (n\u0026thinsp;=\u0026thinsp;97, [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], accessed March, 2025). Parkinson\u0026rsquo;s disease genes (n\u0026thinsp;=\u0026thinsp;602) were curated from the Gene4PD database, a comprehensive resource of genetic associations in Parkinson\u0026rsquo;s disease [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Genes associated with congenital heart defects (CHD, n\u0026thinsp;=\u0026thinsp;210) were obtained from Richter et al. [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e].\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur analyses indicate that different classes of genetic variation contribute to ASD through partially distinct but converging biological mechanisms. Coding DNVs in ASD cases are enriched in loss-of-function\u0026ndash;intolerant genes, which subdivide into broadly expressed versus brain-preferential subsets. Broadly expressed intolerant genes participate in fundamental cellular processes such as transcriptional regulation and chromatin remodeling, whereas narrowly expressed genes are enriched for synaptic pathways. Correspondingly, coding DNVs in the former appear to confer broad neurodevelopmental vulnerability, while those in the latter are more specifically associated with ASD-predominant phenotypes.\u003c/p\u003e\u003cp\u003eWhen viewed in the context of prefrontal cortical development, these gene groups reveal complementary trajectories. Broadly expressed, loss-of-function\u0026ndash;intolerant ASD genes are enriched in early genetic programs downregulated across development, consistent with system-wide cellular regulation, whereas narrowly expressed ASD genes are enriched in transiently upregulated programs during infancy of layer 5\u0026ndash;6 excitatory neurons, reflecting postnatal circuit maturation. Extending this framework, promoter DNVs in ASD probands are enriched in the same deep-layer (L5\u0026ndash;6_TLE4) trajectory but within earlier prenatal-stage programs, suggesting that coding and regulatory variants act at different developmental windows within the same neuronal lineage. These findings align with recent studies partitioning ASD risk into temporally distinct axes: Litman et al. identified a prenatal progenitor and glial axis versus a postnatal neuronal axis, while Russ et al. observed an early chromatin/transcriptional axis and a later synaptic neuronal axis, highlighting deep-layer excitatory neurons as a key site of ASD risk [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Together, these convergent results reinforce a prenatal\u0026ndash;postnatal dichotomy and sharpen the view of layer 5\u0026ndash;6 excitatory neurons as a core substrate of ASD, bridging early system-wide regulation with later neuron-specific synaptic biology.\u003c/p\u003e\u003cp\u003eIn summary, our findings provide a unified model in which variant class, gene constraint, and developmental timing together shape ASD risk. Coding and regulatory variants perturb distinct molecular programs but converge on a shared cellular lineage and functional outcome, offering a mechanistic explanation for how diverse genetic perturbations ultimately manifest in a common neurodevelopmental disorder.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by a grant from Tel Aviv University Center for AI and Data Science (TAD) to R.E. R.E. is a Faculty Fellow of the Edmond J. Safra Center for Bioinformatics at Tel Aviv University. E.E. and S.S. were partially supported by fellowships from the Edmond J. Safra Center for Bioinformatics at Tel Aviv University.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLord C, Elsabbagh M, Baird G, Veenstra-Vanderweele J. Autism spectrum disorder. Lancet. 2018;392:508\u0026ndash;520.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaenner MJ, Shaw KA, Bakian A V, Bilder DA, Durkin MS, Esler A, et al. Prevalence and Characteristics of Autism Spectrum Disorder Among Children Aged 8 Years - Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2018. MMWR Surveill Summ. 2021;70:1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLord C, Brugha TS, Charman T, Cusack J, Dumas G, Frazier T, et al. Autism spectrum disorder. Nat Rev Dis Primers. 2020;6:5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZeidan J, Fombonne E, Scorah J, Ibrahim A, Durkin MS, Saxena S, et al. Global prevalence of autism: A systematic review update. Autism Res. 2022;15:778\u0026ndash;790.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eColvert E, Tick B, McEwen F, Stewart C, Curran SR, Woodhouse E, et al. Heritability of Autism Spectrum Disorder in a UK Population-Based Twin Sample. JAMA Psychiatry. 2015;72:415\u0026ndash;423.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTick B, Bolton P, Happ\u0026eacute; F, Rutter M, Rijsdijk F. Heritability of autism spectrum disorders: a meta-analysis of twin studies. J Child Psychol Psychiatry. 2016;57:585\u0026ndash;595.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIossifov I, O\u0026rsquo;Roak BJ, Sanders SJ, Ronemus M, Krumm N, Levy D, et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature. 2014;515:216\u0026ndash;221.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSanders SJ, He X, Willsey AJ, Ercan-Sencicek AG, Samocha KE, Cicek AE, et al. Insights into Autism Spectrum Disorder Genomic Architecture and Biology from 71 Risk Loci. Neuron. 2015;87:1215\u0026ndash;1233.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSatterstrom FK, Kosmicki JA, Wang J, Breen MS, De Rubeis S, An J-Y, et al. Large-Scale Exome Sequencing Study Implicates Both Developmental and Functional Changes in the Neurobiology of Autism. Cell. 2020;180:568\u0026ndash;584.e23.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKrumm N, Turner TN, Baker C, Vives L, Mohajeri K, Witherspoon K, et al. Excess of rare, inherited truncating mutations in autism. Nat Genet. 2015;47:582\u0026ndash;588.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWilfert AB, Turner TN, Murali SC, Hsieh P, Sulovari A, Wang T, et al. Recent ultra-rare inherited variants implicate new autism candidate risk genes. Nat Genet. 2021;53:1125\u0026ndash;1134.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAn J-Y, Lin K, Zhu L, Werling DM, Dong S, Brand H, et al. Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder. Science. 2018;362.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrandler WM, Antaki D, Gujral M, Kleiber ML, Whitney J, Maile MS, et al. Paternally inherited cis-regulatory structural variants are associated with autism. Science. 2018;360:327\u0026ndash;331.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWillsey AJ, Sanders SJ, Li M, Dong S, Tebbenkamp AT, Muhle RA, et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell. 2013;155:997\u0026ndash;1007.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParikshak NN, Luo R, Zhang A, Won H, Lowe JK, Chandran V, et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell. 2013;155:1008\u0026ndash;1021.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDe Rubeis S, He X, Goldberg AP, Poultney CS, Samocha K, Cicek AE, et al. Synaptic, transcriptional and chromatin genes disrupted in autism. Nature. 2014;515:209\u0026ndash;215.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKarczewski KJ, Francioli LC, Tiao G, Cummings BB, Alf\u0026ouml;ldi J, Wang Q, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581:434\u0026ndash;443.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLevi H, Elkon R, Shamir R. DOMINO: a network-based active module identification algorithm with reduced rate of false calls. Mol Syst Biol. 2021;17:e9593.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDoshi-Velez F, Ge Y, Kohane I. Comorbidity clusters in autism spectrum disorders: an electronic health record time-series analysis. Pediatrics. 2014;133:e54-63.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSanders SJ, Sahin M, Hostyk J, Thurm A, Jacquemont S, Avillach P, et al. A framework for the investigation of rare genetic disorders in neuropsychiatry. Nat Med. 2019;25:1477\u0026ndash;1487.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFu JM, Satterstrom FK, Peng M, Brand H, Collins RL, Dong S, et al. Rare coding variation provides insight into the genetic architecture and phenotypic context of autism. Nat Genet. 2022;54:1320\u0026ndash;1331.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRuss JB, Stone AC, Maney K, Morris LC, Wright CF, Hurst JH, et al. Cell-specific expression biases in human cortex of genes associated with neurodevelopmental disorders. Sci Rep. 2025;15:23172.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLitman A, Sauerwald N, Green Snyder L, Foss-Feig J, Park CY, Hao Y, et al. Decomposition of phenotypic heterogeneity in autism reveals underlying genetic programs. Nat Genet. 2025;57:1611\u0026ndash;1619.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVelmeshev D, Schirmer L, Jung D, Haeussler M, Perez Y, Mayer S, et al. Single-cell genomics identifies cell type-specific molecular changes in autism. Science. 2019;364:685\u0026ndash;689.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHerring CA, Simmons RK, Freytag S, Poppe D, Moffet JJD, Pflueger J, et al. Human prefrontal cortex gene regulatory dynamics from gestation to adulthood at single-cell resolution. Cell. 2022;185:4428\u0026ndash;4447.e28.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNakamura T, Ueda J, Mizuno S, Honda K, Kazuno A-A, Yamamoto H, et al. Topologically associating domains define the impact of de novo promoter variants on autism spectrum disorder risk. Cell Genomics. 2024;4:100488.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhou J, Park CY, Theesfeld CL, Wong AK, Yuan Y, Scheckel C, et al. Whole-genome deep-learning analysis identifies contribution of noncoding mutations to autism risk. Nat Genet. 2019;51:973\u0026ndash;980.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNIAGADS Genomics Database. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://adsp.niagads.org/gvc-top-hits-list/\u003c/span\u003e\u003cspan address=\"https://adsp.niagads.org/gvc-top-hits-list/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLi B, Zhao G, Zhou Q, Xie Y, Wang Z, Fang Z, et al. Gene4PD: A Comprehensive Genetic Database of Parkinson\u0026rsquo;s Disease. Front Neurosci. 2021;15:679568.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRichter F, Morton SU, Kim SW, Kitaygorodsky A, Wasson LK, Chen KM, et al. Genomic analyses implicate noncoding de novo variants in congenital heart disease. Nat Genet. 2020;52:769\u0026ndash;777.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7841641/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7841641/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLarge-scale sequencing has greatly advanced our understanding of the genetic architecture of autism spectrum disorder (ASD). Whole-exome sequencing (WES) in thousands of trios revealed a major role for \u003cem\u003ede novo\u003c/em\u003e protein-truncating variants (PTVs) in loss-of-function (LoF) intolerant genes and identified dozens of high-confidence ASD genes, many central to neuronal development and synaptic signaling. Whole-genome sequencing (WGS) extended these findings by uncovering noncoding contributors, including rare structural and regulatory variants that disrupt gene expression during brain development. Despite these advances, the mechanisms by which diverse mutations converge on shared neurodevelopmental pathways, and the specific cell types and developmental windows most impacted, remain incompletely understood. To address this, we integrated results from large ASD WES/WGS studies with bulk and single-nucleus transcriptomic data. Stratifying LoF-intolerant ASD genes into broadly expressed versus brain-restricted subsets revealed distinct functional roles: broadly expressed genes regulate transcription, chromatin, histone modification, and splicing, whereas brain-restricted genes function mainly in synaptic processes. Clinically, the former tend to be linked to general neurodevelopmental disorders (NDDs), while the latter are more associated with ASD-predominant phenotypes. Intersecting coding and regulatory ASD variants with human prefrontal cortex (PFC) trajectories showed that both converge on the L5-6_TLE4 neuronal lineage, but at different stages: coding \u003cem\u003ede novo\u003c/em\u003e variants disrupt postnatal programs of neuronal maturation, while regulatory promoter variants act earlier, on fetal developmental programs. These findings highlight a framework in which distinct variant classes act within different subsets of LoF-intolerant genes, shaping ASD risk through cell type\u0026ndash; and stage-specific mechanisms.\u003c/p\u003e","manuscriptTitle":"Coding and noncoding de novo variation converge on developmental trajectories of cortical layer 5–6 neurons in autism spectrum disorder","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-25 15:48:37","doi":"10.21203/rs.3.rs-7841641/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1856950b-18ca-409d-a00b-98e285a90f32","owner":[],"postedDate":"November 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":58002665,"name":"Health sciences/Diseases/Psychiatric disorders/Autism spectrum disorders"},{"id":58002666,"name":"Biological sciences/Genetics"}],"tags":[],"updatedAt":"2026-01-16T16:43:26+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-25 15:48:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7841641","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7841641","identity":"rs-7841641","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}