Biology subjects

Capra, J. A.

Publications and source records attributed to Capra, J. A..

54 records · Page 3Linked to original sources

bioRxiv2020

Predicting changes in protein thermostability upon point mutation with deep 3D convolutional neural networks

Predicting mutation-induced changes in protein thermodynamic stability ({triangleup}{triangleup}G) is of great interest in protein engineering, variant interpretation, and understanding protein biophysics. We introduce ThermoNet, a deep, 3D-convolutional neural network designed for structure-based prediction of {triangleup}{triangleup}Gs upon point mutation. To leverage the image-processing power inherent in convolutional neural networks, we treat protein structures as if they were multi-channel 3D images. In particular, the inputs to ThermoNet are uniformly constructed as multi-channel voxel grids based on biophysical properties derived from raw atom coordinates. We train and evaluate ThermoNet with a curated data set that accounts for protein homology and is balanced with direct and reverse mutations; this provides a framework for addressing biases that have likely influenced many previous {triangleup}{triangleup}G prediction methods. ThermoNet demonstrates performance comparable to the best available methods on the widely used Ssym test set. However, ThermoNet accurately predicts the effects of both stabilizing and destabilizing mutations, while most other methods exhibit a strong bias towards predicting destabilization. We further show that homology between Ssym and widely used training sets like S2648 and VariBench has likely led to overestimated performance in previous studies. Finally, we demonstrate the practical utility of ThermoNet in predicting the {triangleup}{triangleup}Gs for two clinically relevant proteins, p53 and myoglobin, and for pathogenic and benign missense variants from ClinVar. Overall, our results suggest that 3D convolutional neural networks can model the complex, non-linear interactions perturbed by mutations, directly from biophysical properties of atoms. Author SummaryThe thermodynamic stability of a protein, usually represented as the Gibbs free energy for the biophysical process of protein folding ({triangleup}G), is a fundamental thermodynamic quantity. Predicting mutation-induced changes in protein thermodynamic stability ({triangleup}{triangleup}G) is of great interest in protein engineering, variant interpretation, and understanding protein biophysics. However, predicting {triangleup}{triangleup}Gs in an accurate and unbiased manner has been a long-standing challenge in the field of computational biology. In this work, we introduce ThermoNet, a deep, 3D-convolutional neural network designed for structure-based {triangleup}{triangleup}G prediction. To leverage the image-processing power inherent in convolutional neural networks, we treat protein structures as if they were multi-channel 3D images. ThermoNet demonstrates performance comparable to the best available methods. However, ThermoNet accurately predicts the effects of both stabilizing and destabilizing mutations, while most other methods exhibit a strong bias towards predicting destabilization. We also demonstrate that the presence of homologous proteins in commonly used training and testing sets for {triangleup}{triangleup}G prediction methods has likely influenced previous performance estimates. Finally, we highlight the practical utility of ThermoNet by applying it to predicting the {triangleup}{triangleup}Gs for two clinically relevant proteins, p53 and myoglobin, and for pathogenic and benign missense variants from ClinVar.

bioinformatics↗

bioRxiv2020

Topologically associating domain (TAD) boundaries stable across diverse cell types are evolutionarily constrained and enriched for heritability

Topologically associating domains (TADs) are fundamental units of three-dimensional (3D) nuclear organization. The regions bordering TADs--TAD boundaries--contribute to the regulation of gene expression by restricting interactions of cis-regulatory sequences to their target genes. TAD and TAD boundary disruption have been implicated in rare disease pathogenesis; however, we have a limited framework for integrating TADs and their variation across cell types into the interpretation of common trait-associated variants. Here, we investigate an attribute of 3D genome architecture--the stability of TAD boundaries across cell types--and demonstrate its relevance to understanding how genetic variation in TADs contribute to complex disease. By synthesizing TAD maps across 37 diverse cell types with 41 genome-wide association studies (GWAS), we investigate the differences in functionality and evolutionary pressure on variation in TADs versus TAD boundaries. We quantify their contribution to trait heritability and sequence-level evolutionary constraint and demonstrate that genetic variation in TAD boundaries contributes more to complex trait heritability, especially for immunologic, hematologic, and metabolic traits. We also show that TAD boundaries are more evolutionarily constrained than TADs. Next, stratifying boundaries by their stability across cell types, we find substantial differences. Boundaries stable across cell types are further enriched for complex trait heritability, evolutionary constraint, CTCF binding, and housekeeping genes compared to boundaries unique to a specific cell type. This suggests greater functional importance for stable boundaries. Thus, considering TAD boundary stability across cell types provides valuable context for understanding the genomes functional landscape and enabling 3D-structure aware variant interpretation.

genomics↗

bioRxiv2019

LEARNING AND INTERPRETING THE GENE REGULATORY GRAMMAR IN A DEEP LEARNING FRAMEWORK

Deep neural networks (DNNs) have achieved state-of-the-art performance in identifying gene regulatory sequences, but they have provided limited insight into the biology of regulatory elements due to the difficulty of interpreting the complex features they learn. Several models of how combinatorial binding of transcription factors, i.e. the regulatory grammar, drives enhancer activity have been proposed, ranging from the flexible TF billboard model to the stringent enhanceosome model. However, there is limited knowledge of the prevalence of these (or other) sequence architectures across enhancers. Here we perform several hypothesis-driven analyses to explore the ability of DNNs to learn the regulatory grammar of enhancers. We created a synthetic dataset based on existing hypotheses about combinatorial transcription factor binding site (TFBS) patterns, including homotypic clusters, heterotypic clusters, and enhanceosomes, from real TF binding motifs from diverse TF families. We then trained deep residual neural networks (ResNets) to model the sequences under a range of scenarios that reflect real-world multi-label regulatory sequence prediction tasks. We developed a gradient-based unsupervised clustering method to extract the patterns learned by the ResNet models. We demonstrated that simulated regulatory grammars are best learned in the penultimate layer of the ResNets, and the proposed method can accurately retrieve the regulatory grammar even when there is heterogeneity in the enhancer categories and a large fraction of TFBS outside of the regulatory grammar. However, we also identify common scenarios where ResNets fail to learn simulated regulatory grammars. Our results provide a framework for interpreting the regulatory rules learned by ResNets, and they demonstrate that the ability and efficiency of ResNets in learning the regulatory grammar highly depends on the nature of the prediction task.

bioinformatics↗

bioRxiv2019

High-throughput reclassification of SCN5A variants

RationalePartial or complete loss of function variants in SCN5A are the most common genetic cause of the arrhythmia disorder Brugada Syndrome (BrS1). However, the pathogenicity of SCN5A variants is often unknown or disputed; 80% of the 1,390 SCN5A missense variants observed in at least one individual to date are variants of uncertain significance (VUS). The designation of VUS is a barrier to the use of sequence data in clinical care. ObjectiveWe selected 83 variants for study: 10 previously studied control variants, 10 suspected benign variants, and 63 suspected Brugada Syndrome-associated variants, selected on the basis of their frequency in the general population and in patients with Brugada Syndrome. We used high-throughput automated patch clamping to study the function of the 83 variants, with the goal of reclassifying variants with functional data. Methods and ResultsTen previously studied variants had functional properties concordant with published manual patch clamp data. All 10 suspected benign variants had wildtype-like function. 22 suspected BrS variants had loss of channel function (<10% normalized peak current) and 23 variants had partial loss of function (10-50% normalized peak current). The 73 previously unstudied variants were initially classified as likely benign (n=2), likely pathogenic (n=11), or VUS (n=60). After the patch clamp studies, 16 variants were benign/likely benign, 47 were pathogenic/likely pathogenic, and only 10 were still VUS. 8/22 loss of function variants were partially rescuable by incubation at lower temperature or pretreatment with a sodium channel blocker. Structural modeling identified likely mechanisms for loss of function including altered thermostability, and disruptions to alpha helices, disulfide bonds, or the permeation pore. ConclusionsHigh-throughput automated patch clamping enabled the reclassification of the majority of tested VUSs in SCN5A.

genetics↗

bioRxiv2019

Unravelling the genetic architecture of musical rhythm

Moving in synchrony to the beat is a fundamental component of musicality. Here, we conducted a genome-wide association study (GWAS) to identify common genetic variants associated with beat synchronization in 606,825 individuals. Beat synchronization exhibited a highly polygenic architecture, with sixty-nine loci reaching genome-wide significance (p<5x10-8) and SNP-based heritability (on the liability scale) of 13%-16%. Heritability was enriched for genes expressed in brain tissues, and for fetal and adult brain-specific gene regulatory elements, underscoring the role of central nervous system-expressed genes linked to the genetic basis of the trait. We performed validations of the self-report phenotype (through internet-based experiments) and of the GWAS (polygenic scores for beat synchronization were associated with patients algorithmically classified as musicians in medical records of a separate biobank). Genetic correlations with breathing function, motor function, processing speed, and chronotype suggest shared genetic architecture with beat synchronization and provide avenues for new phenotypic and genetic explorations.

genetics↗

bioRxiv2019

Accounting for diverse evolutionary forces reveals the mosaic nature of selection on genomic regions associated with human preterm birth

Human pregnancy requires the coordinated function of multiple tissues in both mother and fetus and has evolved in concert with major human adaptations. As a result, pregnancy-associated phenotypes and related disorders are genetically complex and have likely been sculpted by diverse evolutionary forces. However, there is no framework to comprehensively evaluate how these traits evolved or to explore the relationship of evolutionary signatures on trait-associated genetic variants to molecular function. Here we develop an approach to test for signatures of diverse evolutionary forces, including multiple types of selection, and apply it to genomic regions associated with spontaneous preterm birth (sPTB), a complex disorder of global health concern. We find that sPTB-associated regions harbor diverse evolutionary signatures including evolutionary sequence conservation (consistent with the action of negative selection), excess population differentiation (local adaptation), accelerated evolution (positive selection), and balanced polymorphism (balancing selection). Furthermore, these genomic regions show diverse functional characteristics which enables us to use evolutionary and molecular lines of evidence to develop hypotheses about how these genomic regions contribute to sPTB risk. In summary, we introduce an approach for inferring the spectrum of evolutionary forces acting on genomic regions associated with complex disorders. When applied to sPTB-associated genomic regions, this approach both improves our understanding of the potential roles of these regions in pathology and illuminates the mosaic nature of evolutionary forces acting on genomic regions associated with sPTB.

evolutionary biology↗

bioRxiv2019

Do available protein 3D structures reflect human genetic and functional diversity?

Genomic databases are substantially biased towards European ancestry populations, and this bias contributes to health disparities. Here, we quantify how well 66,971 experimentally characterized human protein 3D structures represent the diversity of protein sequences observed across the 1000 Genomes Project. More than 85% of available structures do not match a sequence observed in at least one individual, and on average structures match the sequence of 74% of individuals. Nearly 23% of human structures do not match any observed sequences; however, after masking engineered/known mutations, this decreases to ~4%. African ancestry sequences are modestly, but significantly, less likely to be represented by structures (73.5% vs. 74.0%). These differences are mainly driven by the greater genetic diversity of African populations. We identify thousands of variants unrepresented in available structures that influence protein structure and function. Thus, the use of a single structure as representative of \"the wild type\" protein will often bias results against many individuals. The diversity of protein sequence and structure must be considered to enable accurate, reproducible, and generalizable conclusions from structural analyses.

genetics↗

bioRxiv2019

Neanderthal introgression reintroduced functional alleles lost in the human out of Africa bottleneck

Neanderthal ancestry remains across modern Eurasian genomes, and introgressed sequences influence diverse phenotypes, including immune, skin, and neuropsychiatric diseases. Interpretation of introgressed sequences has focused on alleles derived in the Neanderthal lineage. Here, we demonstrate that Neanderthal introgression also reintroduced thousands of ancestral hominin alleles lost in the Eurasian out of Africa bottleneck. Combining evolutionary simulations, expression quantitative trait loci (eQTL), massively parallel reporter assay (MPRA) data, and in vitro validation, we show that reintroduced alleles (RAs) have different fitness effects than Neanderthal-derived alleles (NDAs) and that some RAs regulate gene expression independent of NDAs. Illustrating the broad potential influence of RAs, we find that over 70% of known phenotype associations with NDAs are equally associated with RAs. Finally, we discover enrichment for RA eQTL activity in several tissues, with strongest enrichment in the brain. In summary, our study reveals that Neanderthal introgression supplied Eurasians with many lost functional variants and demonstrates that RAs must be considered when evaluating the effects of introgression. ONE SENTENCE SUMMARYNeanderthal interbreeding with modern humans restored to Eurasians, hundreds of thousands of ancient alleles that were lost in the out of Africa bottleneck.

evolutionary biology↗

bioRxiv2019

Signatures of recent positive selection in enhancers across 41 human tissues

Evolutionary changes in enhancers are widely associated with variation in human traits and diseases. However, studies comprehensively quantifying levels of selection on enhancers at multiple evolutionary time points during recent human evolution and how enhancer evolution varies across human tissues are lacking. To address these questions, we integrated a dataset of 41,561 transcribed enhancers active in 41 different human tissues (FANTOM Consortium) with whole genome sequences of 1,668 individuals from the African, Asian, and European populations (1000 Genomes Project). Our analyses based on four different metrics (Tajimas D, FST, H12, nSL) showed that ~5.90% of enhancers considered showed evidence of recent positive selection and that genes associated with enhancers under positive selection are enriched for diverse immune-related functions. The distributions of these metrics for brain and testis enhancers were often statistically significantly different compared to those of other tissues; the same was true for brain and testis enhancers that are tissue-specific compared to those that are tissue-broad and for testis enhancers associated with tissue-enriched and non-tissue-enriched genes. These differences varied considerably across metrics and tissues and were generally due to changes in distributions shapes rather than shifts in their values. These results suggest that many human enhancers experienced recent positive selection throughout multiple time periods in human evolutionary history, that this selection occurred in a tissue-dependent and immune-related functional context, and that much like the evolution of their coding counterparts, the evolution of brain and testis enhancers has been markedly different from that of enhancers in other tissues.

evolutionary biology↗

bioRxiv2018

The impact of natural selection on the evolution and function of placentally expressed galectins

Immunity genes have repeatedly experienced natural selection during mammalian evolution. Galectins are carbohydrate-binding proteins that regulate diverse immune responses, including maternal-fetal immune tolerance in placental pregnancy. Seven human galectins, four conserved across vertebrates and three specific to primates, are involved in placental development. To comprehensively study the molecular evolution of these galectins both across mammals and within humans, we conducted a series of between-and within-species evolutionary analyses. By examining patterns of sequence evolution between species, we found that primate-specific galectins showed uniformly high substitution rates, whereas two of the four other galectins experienced accelerated evolution in primates. By examining human population genomic variation, we found that galectin genes and variants, including variants previously linked to immune diseases, showed signatures of recent positive selection in specific human populations. By examining one nonsynonymous variant in Galectin-8 previously associated with autoimmune diseases, we further discovered that it is tightly linked to three other nonsynonymous variants; surprisingly, the global frequency of this four-variant haplotype is [~]50%. To begin understanding the impact of this major haplotype on Galectin-8 protein structure, we modeled its 3D protein structure and found that it differed substantially from the reference protein structure. These results suggest that placentally expressed galectins experienced both ancient and more recent selection in a lineage-and population-specific manner. Furthermore, our discovery that the major Galectin-8 haplotype is structurally distinct from and more commonly found than the reference haplotype illustrates the significance of understanding the evolutionary processes that sculpted variants associated with human genetic disease.

evolutionary biology↗

bioRxiv2018

Genome-wide maps of distal gene regulatory regions active in the human placenta

Placental dysfunction is implicated in many pregnancy complications, including preeclampsia and preterm birth (PTB). While both these syndromes are influenced by environmental risk factors, they also have a substantial genetic component that is not well understood. Precisely controlled gene expression during development is crucial to proper placental function and often mediated through gene regulatory enhancers. However, we lack accurate maps of placental enhancer activity due to the challenges of assaying the placenta and the difficulty of comprehensively identifying enhancers. To address the gap in our knowledge of gene regulatory elements in the placenta, we used a two-step machine learning pipeline to synthesize existing functional genomics studies, transcription factor (TF) binding patterns, and evolutionary information to predict placental enhancers. The trained classifiers accurately distinguish enhancers from the genomic background and placental enhancers from enhancers active in other tissues. Genomic features collected from tissues and cell lines involved in pregnancy are the most predictive of placental regulatory activity. Applying the classifiers genome-wide enabled us to create a map of 33,010 predicted placental enhancers, including 4,562 high-confidence enhancer predictions. The genome-wide placental enhancers are significantly enriched nearby genes associated with placental development and birth disorders and for SNPs associated with gestational age. These genome-wide predicted placental enhancers provide candidate regions for further testing in vitro, will assist in guiding future studies of genetic associations with pregnancy phenotypes, and aid interpretation of potential mechanisms of action for variants found through genetic studies.

genomics↗

bioRxiv2018

integRATE: a desirability-based data integration framework for the prioritization of candidate genes across heterogeneous omics and its application to preterm birth

BackgroundThe integration of high-quality, genome-wide analyses offers a robust approach to elucidating genetic factors involved in complex human diseases. Even though several methods exist to integrate heterogeneous omics data, most biologists still manually select candidate genes by examining the intersection of lists of candidates stemming from analyses of different types of omics data that have been generated by imposing hard (strict) thresholds on quantitative variables, such as P-values and fold changes, increasing the chance of missing potentially important candidates.\n\nMethodsTo better facilitate the unbiased integration of heterogeneous omics data collected from diverse platforms and samples, we propose a desirability function framework for identifying candidate genes with strong evidence across data types as targets for follow-up functional analysis. Our approach is targeted towards disease systems with sparse, heterogeneous omics data, so we tested it on one such pathology: spontaneous preterm birth (sPTB).\n\nResultsWe developed the software integRATE, which uses desirability functions to rank genes both within and across studies, identifying well-supported candidate genes according to the cumulative weight of biological evidence rather than based on imposition of hard thresholds of key variables. Integrating 10 sPTB omics studies identified both genes in pathways previously suspected to be involved in sPTB as well as novel genes never before linked to this syndrome. integRATE is available as an R package on GitHub (https://github.com/haleyeidem/integRATE).\n\nConclusionsDesirability-based data integration is a solution most applicable in biological research areas where omics data is especially heterogeneous and sparse, allowing for the prioritization of candidate genes that can be used to inform more targeted downstream functional analyses.

genomics↗

bioRxiv2018

Genome wide association analysis identifies genetic variants associated with reproductive variation across domestic dog breeds and uncovers links to domestication

The diversity of eutherian reproductive strategies has led to variation in many traits, such as number of offspring, age of reproductive maturity, and gestation length. While reproductive trait variation has been extensively investigated and is well established in mammals, the genetic loci contributing to this variation remain largely unknown. The domestic dog, Canis lupus familiaris is a powerful model for studies of the genetics of inherited disease due to its unique history of domestication. To gain insight into the genetic basis of reproductive traits across domestic dog breeds, we collected phenotypic data for four traits - cesarean section rate (n = 97 breeds), litter size (n = 60), stillbirth rate (n = 57), and gestation length (n = 23) - from primary literature and breeders handbooks. By matching our phenotypic data to genomic data from the Cornell Veterinary Biobank, we performed genome wide association analyses for these four reproductive traits, using body mass and kinship among breeds as co-variates. We identified 14 genome-wide significant associations between these traits and genetic loci, including variants near CACNA2D3 with gestation length, MSRB3 with litter size, SMOC2 with cesarean section rate, MITF with litter size and still birth rate, KRT71 with cesarean section rate, litter size, and stillbirth rate, and HTR2C with stillbirth rate. Some of these loci, such as CACNA2D3 and MSRB3, have been previously implicated in human reproductive pathologies. Many of the variants that we identified have been previously associated with domestication-related traits, including brachycephaly (SMOC2), coat color (MITF), coat curl (KRT71), and tameness (HTR2C). These results raise the hypothesis that the artificial selection that gave rise to dog breeds also shaped the observed variation in their reproductive traits. Overall, our work establishes the domestic dog as a system for studying the genetics of reproductive biology and disease.

evolutionary biology↗

bioRxiv2017

Polygenic selection underlies evolution of human brain structure and behavioral traits

Seemingly paradoxical characteristics of psychiatric disorders, including moderate to high prevalence, reduced fecundity, and high heritability have motivated explanations for the persistence of common risk alleles for severe psychiatric phenotypes throughout human evolution. Proposed mechanisms include balancing selection, drift, and weak polygenic adaptation acting either directly, or indirectly through selection on correlated traits. While many mechanisms have been proposed, few have been empirically tested. Leveraging publicly available data of unprecedented sample size, we studied twenty-five traits (i.e., ten neuropsychiatric disorders, three personality traits, total intracranial volume, seven subcortical brain structure volume traits, and four complex traits without neuropsychiatric associations) for evidence of several different signatures of selection over a range of evolutionary time scales. Consistent with the largely polygenic architecture of neuropsychiatric traits, we found no enrichment of trait-associated single-nucleotide polymorphisms (SNPs) in regions of the genome that underwent classical selective sweeps (i.e., events which would have driven selected alleles to near fixation). However, we discovered that SNPs associated with some, but not all, behaviors and brain structure volumes are enriched in genomic regions under selection since divergence from Neanderthals ~600,000 years ago, and show further evidence for signatures of ancient and recent polygenic adaptation. Individual subcortical brain structure volumes demonstrate genome-wide evidence in support of a mosaic theory of brain evolution while total intracranial volume and height appear to share evolutionary constraints consistent with concerted evolution. We further characterized the biological processes potentially targeted by selection, through expression Quantitative Trait Locus (eQTL) and Gene Ontology (GO) enrichment analyses and found evidence for the role of regulatory functions among selected SNPs in immune and brain tissues. Taken together, our results suggest that alleles associated with neuropsychiatric, behavioral, and brain volume phenotypes have experienced both ancient and recent polygenic adaptation in human evolution, acting through neurodevelopmental and immune-mediated pathways.

evolutionary biology↗

bioRxiv2017

Genome-wide Enhancer Maps Differ Significantly in Genomic Distribution, Evolution, and Function

Non-coding gene regulatory enhancers are essential to transcription in mammalian cells. As a result, a large variety of experimental and computational strategies have been developed to identify cis-regulatory enhancer sequences. In practice, most studies consider enhancers identified by only a single method, and the concordance of enhancers identified by different methods has not been comprehensively evaluated. Here, we assess the similarities of enhancer sets identified by ten representative strategies in four biological contexts and evaluate the robustness of downstream conclusions to the choice of identification strategy. All pairs of enhancer sets we evaluated overlap significantly more than expected by chance; however, we also found significant dissimilarity between enhancer sets in their genomic characteristics, evolutionary conservation, and association with functional loci within each context. We find most regions identified as enhancers are supported by only one method. The disagreement is sufficient to influence interpretation of GWAS SNPs and eQTL, and to lead to disparate conclusions about enhancer biology and disease mechanisms. We also find only limited evidence that regions identified by multiple enhancer identification methods are better candidates than those identified by a single method. Our results highlight the inherent complexity of enhancer biology and argue that current approaches have yet to adequately account for enhancer diversity. As a result, we cannot recommend the use of any single enhancer identification strategy in isolation. To facilitate assessment of enhancer diversity on studies conclusions, we developed creDB, a database of enhancer annotations designed to integrate into bioinformatics workflows. While our findings highlight a major challenge to mapping the genetic architecture of complex disease and interpreting regulatory variants found in patient genomes, a systematic understanding of similarities and differences in enhancer identification methodology will ultimately enable robust inferences about gene regulatory sequences.

genetics↗

bioRxiv2017

Short DNA sequence patterns accurately identify broadlyactive human enhancers

Enhancers are DNA regulatory elements that influence gene expression. There is substantial diversity in enhancers activity patterns: some enhancers drive expression in a single cellular context, while others are active across many. Sequence characteristics, such as transcription factor (TF) binding motifs, influence the activity patterns of regulatory sequences; however, the regulatory logic through which specific sequences drive enhancer activity patterns is poorly understood. Recent analysis of Drosophila enhancers suggested that short dinucleotide repeat motifs (DRMs) are general enhancer sequence features that drive broad regulatory activity. However, it is not known whether the regulatory role of DRMs is conserved across species. We performed a comprehensive analysis of the relationship between short DNA sequence patterns, including DRMs, and human enhancer activity in 38,538 enhancers across 411 different contexts. In a machine-learning framework, the occurrence patterns of short sequence motifs accurately predicted broadly active human enhancers. However, DRMs alone were weakly predictive of broad enhancer activity in humans and showed different enrichment patterns than in Drosophila. In general, GC-rich sequence motifs were significantly associated with broad enhancer activity; consistent with this enrichment, broadly active human TFs recognize GC-rich motifs. Our results reveal the importance of specific sequence motifs in broadly active human enhancers, demonstrate the lack of evolutionary conservation of the role of DRMs, and provide a computational framework for investigating the logic of enhancer sequences.

genomics↗

bioRxiv2017

Sequence properties underlying gene regulatory enhancers are conserved across mammals

In mammals, genomic regions with enhancer activity turnover rapidly; in contrast, gene expression patterns and transcription factor binding preferences are largely conserved. Based on this conservation, we hypothesized that enhancers active in different mammals would exhibit conserved sequence patterns in spite of their different genomic locations. We tested this hypothesis by quantifying the conservation of sequence patterns underlying histone-mark defined enhancers across six diverse mammals in two machine learning frameworks. We first trained support vector machine (SVM) classifiers based on the frequency spectrum of short DNA sequence patterns. These classifiers accurately identified many adult liver, developing limb, and developing brain enhancers in each species. Then, we applied these classifiers across species and found that classifiers trained in one species and tested in another performed nearly as well as classifiers trained and tested on the same species. This indicates that the short sequence patterns predictive of enhancers are largely conserved. We also observed similar cross-species conservation when applying the models to human and mouse enhancers validated in transgenic assays. The sequence patterns most predictive of enhancers in each species matched the binding motifs for a common set of TFs enriched for expression in relevant tissues, supporting the biological relevance of the learned features. To test the conservation of more complex sequences patterns, we trained convolutional neural networks (CNNs) on enhancer sequences in each species. The CNNs demonstrated better performance overall, but worse cross-species generalization than SVMs, suggesting the importance of combinatorial interactions between motifs, but less conservation of these more complex sequence patterns. Thus, despite the rapid change of active enhancer locations between mammals, cross-species enhancer prediction is often possible. Furthermore, short sequence patterns encoding enhancer activity have been maintained across more than 180 million years of mammalian evolution, with evolutionary change in more complex sequence patterns.\n\nAuthor summaryAlterations in gene expression levels are a driving force of both speciation and complex disease; therefore, it is of great importance to understand the mechanisms underlying the evolution and function gene regulatory DNA sequences. Recent studies have revealed that while gene expression patterns and transcription factor binding preferences are broadly conserved across diverse animals, there is extensive turnover in distal gene regulatory regions, called enhancers, between closely related species. We investigate this seeming incongruence by analyzing genome-wide enhancer datasets from six diverse mammalian species. We trained two machine-learning classifiers--a k-mer spectrum support vector machine (SVM) and convolutional neural network (CNN)--to distinguish enhancers from the genomic background. The k-mer spectrum SVM models the occurrences of short sequence patterns while the CNN models both the short sequences patterns and their combinatorial patterns. Both the SVM and CNN enhancer prediction models trained in one species are able to predict enhancers in the same cellular context in other species. However, CNNs performed better at predicting enhancers in each species, but they generalize less well across species than the SVMs. This argues that the short sequence properties encoding regulatory activity are remarkably conserved across more than 180 million years of mammalian evolution with more evolutionary turnover in the more complex combinations of the conserved short sequence motifs.

evolutionary biology↗

bioRxiv2017

Comprehensive Analysis of Constraint on the Spatial Distribution of Missense Variants in Human Protein Structures

The spatial distribution of genetic variation within proteins is shaped by evolutionary constraint and thus can provide insights into the functional importance of protein regions and the potential pathogenicity of protein alterations. Here, we comprehensively evaluate the 3D spatial patterns of constraint on human germline and somatic variation in 4,568 solved protein structures. Different classes of coding variants have significantly different spatial distributions. Neutral missense variants exhibit a range of 3D constraint patterns, with a general trend of spatial dispersion driven by constraint on core residues. In contrast, germline and somatic disease-causing variants are significantly more likely to be clustered in protein structure space. We demonstrate that this difference in the spatial distributions of disease-associated and benign germline variants provides a signature for accurately classifying variants of unknown significance (VUS) that is complementary to current approaches for VUS classification. We further illustrate the clinical utility of our approach by classifying new mutations identified from patients with familial idiopathic pneumonia (FIP) that segregate with disease.

genetics↗