Prioritizing non-coding regions based on human genomic constraint and sequence context with deep learning

Vitsios, Dimitrios; Dhindsa, Ryan S.; Middleton, Lawrence; Gussow, Ayal B.; Petrovski, Slavé

doi:10.1038/s41467-021-21790-4

Cited by 51 publications

(52 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For most regions, even nominal significance was not detected using burden testing indicating the likelihood of variants with bidirectional effects further complicating clinical interpretation. When burden signals were detected, observed effects were typically larger than common non-coding variants and less than rare coding variants, with the exception of LDLR , consistent with whole genome mutational constraint models 49, 50, 51 .…”

Section: Discussionsupporting

confidence: 60%

Whole genome sequence analysis of blood lipid levels in >66,000 individuals

Selvaraj

et al. 2021

Preprint

View full text Add to dashboard Cite

Plasma lipids are heritable modifiable causal factors for coronary artery disease, the leading cause of death globally. Despite the well-described monogenic and polygenic bases of dyslipidemia, limitations remain in discovery of lipid-associated alleles using whole genome sequencing, partly due to limited sample sizes, ancestral diversity, and interpretation of potential clinical significance. Increasingly larger whole genome sequence datasets with plasma lipids coupled with methodologic advances enable us to more fully catalog the allelic spectrum for lipids. Here, among 66,329 ancestrally diverse (56% non-European ancestry) participants, we associate 428M variants from deep-coverage whole genome sequences with plasma lipids. Approximately 400M of these variants were not studied in prior lipids genetic analyses. We find multiple lipid-related genes strongly associated with plasma lipids through analysis of common and rare coding variants. We additionally discover several significantly associated rare non-coding variants largely at Mendelian lipid genes. Notably, we detect rare LDLR intronic variants associated with markedly increased LDL-C, similar to rare LDLR exonic variants. In conclusion, we conducted a systematic whole genome scan for plasma lipids expanding the alleles linked to lipids for multiple ancestries and characterize a clinically-relevant rare non-coding variant model for lipids.

show abstract

Section: Discussionsupporting

confidence: 60%

Whole genome sequence analysis of blood lipid levels in >66,000 individuals

Selvaraj

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…We compared our metric with other four genome-wide predictive scores – Orion 13 , CDTS 14 , gwRVIS 17 , and JARVIS 17 ) – in their correlation with experimental measurements on 11 enhancers tested by MAVE 77 . Each predictive score was downloaded from the original study, lifted over to GRCh38 (for Orion), and applied to score enhancers by taking the average over corresponding genomic regions.…”

Section: Methodsmentioning

confidence: 99%

A genome-wide mutational constraint map quantified from variation in 76,156 human genomes

Chen

Francioli

Goodrich

et al. 2022

Preprint

289

257

View full text Add to dashboard Cite

The depletion of disruptive variation caused by purifying natural selection (constraint) has been widely used to investigate protein-coding genes underlying human disorders, but attempts to assess constraint for non-protein-coding regions have proven more difficult. Here we aggregate, process, and release a dataset of 76,156 human genomes from the Genome Aggregation Database (gnomAD), the largest public open-access human genome reference dataset, and use this dataset to build a mutational constraint map for the whole genome. We present a refined mutational model that incorporates local sequence context and regional genomic features to detect depletions of variation across the genome. As expected, protein-coding sequences overall are under stronger constraint than non-coding regions. Within the non-coding genome, constrained regions are enriched for known regulatory elements and variants implicated in complex human diseases and traits, facilitating the triangulation of biological annotation, disease association, and natural selection to non-coding DNA analysis. More constrained regulatory elements tend to regulate more constrained protein-coding genes, while non-coding constraint captures additional functional information underrecognized by gene constraint metrics. We demonstrate that this genome-wide constraint map provides an effective approach for characterizing the non-coding genome and improving the identification and interpretation of functional human genetic variation.

show abstract

“…Quantifying the depletion of natural variation in human populations provides a powerful approach to identify variants of large effect [1][2][3][4][5][6][7][8] . Since variants causing severe early-onset disorders are under selective pressure, they are observed less often than functionally neutral variants.…”

Section: Mainmentioning

confidence: 99%

“…Since variants causing severe early-onset disorders are under selective pressure, they are observed less often than functionally neutral variants. Such depletion of genetic variation (constraint) has been shown to provide strong evidence to prioritise diseaseassociated genes [1][2][3] , identify critical regions within genes 4,5 , and investigate the effect of non-coding variants [6][7][8] .…”

Section: Mainmentioning

confidence: 99%

Genetic constraint at single amino acid resolution improves missense variant prioritisation and gene discovery

Zhang

Theotokis

et al. 2022

Preprint

View full text Add to dashboard Cite

The clinical impact of most germline missense variants in humans remains unknown. Genetic constraint identifies genomic regions under negative selection, where variations likely have functional impacts, but the spatial resolution of existing constraint metrics is limited. Here we present the Homologous Missense Constraint (HMC) score, which measures genetic constraint at quasi single amino-acid resolution by aggregating signals across protein homologues. We identify one million possible missense variants under strong negative selection. HMC precisely distinguishes pathogenic variants from benign variants for both early-onset and adult-onset disorders. It outperforms existing constraint metrics and pathogenicity meta-predictors in prioritising de novo mutations from probands with developmental disorders (DD), and is orthogonal to these, adding power when used in combination. We demonstrate utility for gene discovery by identifying seven genes newly-significant associated with DD that could act through an altered-function mechanism. Overall, HMC is a novel and strong predictor to improve missense variant interpretation.

show abstract

Prioritizing non-coding regions based on human genomic constraint and sequence context with deep learning

Cited by 51 publications

References 49 publications

Whole genome sequence analysis of blood lipid levels in >66,000 individuals

Whole genome sequence analysis of blood lipid levels in >66,000 individuals

A genome-wide mutational constraint map quantified from variation in 76,156 human genomes

Genetic constraint at single amino acid resolution improves missense variant prioritisation and gene discovery

Contact Info

Product

Resources

About