Population-wide copy number variation calling using variant call format files from 6,898 individuals

Area of Science:

• Genomics
• Human Genetics
• Bioinformatics

Background:

• Accurate detection of copy number variants (CNVs) is crucial for understanding human diseases.
• Current CNV detection methods are computationally intensive, often relying on raw read alignments.
• Germline CNVs are a significant, yet challenging, class of genetic variation to identify.

Purpose of the Study:

• To develop and apply an efficient, regression tree-based method for germline CNV calling from existing whole-genome sequencing (WGS) variant call sets.
• To characterize the landscape of germline CNVs in a large European population cohort.
• To assess the functional impact of detected CNVs by associating them with protein quantitative trait loci (pQTLs).

Main Methods:

• Utilized a regression tree approach for germline CNV detection from WGS variant call sets (>18x coverage) across 6,898 individuals.
• Analyzed 1,320 identified CNVs, comparing them against the Database of Genomic Variants for known events.
• Performed association testing between deletions and 275 protein levels in 1,457 individuals to identify potential clinical impacts.

Main Results:

• Identified a substantial landscape of 1,320 germline CNVs, with 81% previously reported.
• Confirmed known deletions (e.g., GSTM1, RHD) and found 23% of high-quality deletions affect entire genes.
• Discovered significant associations between CNVs and protein levels, including a novel cis-association for NOMO1 deletions with NOMO1 protein levels.

Conclusions:

• Existing population-wide WGS data can be efficiently mined for germline CNVs with minimal computational cost.
• The developed method provides valuable insights into the impact of CNVs on human health and disease.
• This approach enhances the study of CNVs, a critical but often under-investigated class of genetic variation.