Identifying structural variants influencing human health in population cohorts

Brigham And Women'S Hospital, Boston MA

Investigators

Linked publications, trials & patents

Abstract

Project Summary/Abstract Large-scale biobank resources of genetic and phenotypic data hold great promise for revealing insights into disease genetics and enabling genetically-informed, targeted therapeutics. To more fully realize this potential, new statistical methods are needed to recover latent information about genomic structural variants – i.e., polymorphisms modifying >50 base pairs of DNA sequence – within these data sets. Because of their large size, structural variants collectively contribute more base pairs of variation within an individual’s genome than single-nucleotide polymorphisms (SNPs) or short indels. However, structural variants have been difficult to identify and genotype from the SNP-array and short-read sequencing data generated by biobanks to date. We will undertake a research program to develop a new suite of “haplotype-informed” statistical algorithms designed to accurately and efficiently genotype structural variants in large biobank data sets. This approach will leverage the fact that population-polymorphic genetic variants are typically carried by multiple individuals within a large cohort who co-inherited an extended SNP-haplotype. Identification of such shared haplotypes will enable information about a structural variant carried by one individual to inform detection of the same variant carried by other distantly related individuals, simultaneously facilitating structural variant genotyping, variant harmonization, and haplotype-resolved analysis. This project will have three specific aims. First, we will develop haplotype-informed computational methods that improve detection sensitivity and genotyping accuracy for several classes of structural variation using short- read sequencing data. These methods will be particularly helpful for analysis of short copy-number variants (CNVs) from exome-sequencing data and for analysis of multi-allelic CNVs and large repeats from exome- or genome-sequencing data. Second, we will develop methods for imputing structural variants into genotype- phenotype association data sets – a statistical approach that has been extremely effective in genome-wide association studies (GWAS) of SNPs and indels but has been difficult to apply to structural variants. We will develop new methods to impute structural variants from short- or long-read-based reference panels and will also develop a pipeline for imputing and fine-mapping structural variant associations into GWAS summary statistics. Third, we will genotype structural variants in multiple large genetic biobank data sets and identify associated health outcomes. We will return haplotype-resolved structural variant call sets for use by other researchers. We anticipate that these efforts will reveal new structural variant polymorphisms with large phenotypic effects, augment existing biobank resources, and enable imputation into further data sets.