The spectrum of copy number variation in the Pan-Canadian HostSeq databank

The integration of copy number variant (CNV) workflows into genome sequencing (GS) analysis pipelines allows for the identification of CNVs implicated in disease. Here, we generated a novel resource of CNVs identified in the HostSeq population cohort from Canada, identified the prevalence of recurrent CNVs associated with neurodevelopmental disorders, and determined CNVs of potential clinical relevance for reproductive planning and personal disease risk. GS data and CNV calls were generated for 10,488 participants from across Canada as part of the HostSeq initiative. The CNV calls were filtered to generate a rare dataset, which was further filtered into the OMIM morbid, ClinGen dosage, and DECIPHER datasets. CNV deletions were stratified into either Tier 1, 2, or 3 based on the inheritance pattern of the genes involved. A putatively pathogenic dataset was generated by identifying CNVs in the ClinGen dosage and DECIPHER datasets with at least 80% overlap with previously identified pathogenic CNVs. A total of 8,543,334 CNV calls were generated. Filtering for rare variants yielded 36,631 CNVs, of which 9,922 (27.08%) encompassed at least one OMIM gene, 728 (1.99%) had at least 10% overlap with a DECIPHER region, and 1,136 (3.10%) encompassed at least one ClinGen dosage-sensitive gene. CNV deletions were stratified into 1,833 Tier 1 deletions, 32 Tier 2 deletions, and 184 Tier 3 deletions. There were 82 CNVs identified in regions associated with neurodevelopmental disorders. Of the genes with either a ClinGen dosage sensitivity score or overlap with a DECIPHER region, 206 CNVs were deemed as putatively pathogenic. We were able to detect a wide range of CNVs, highlighting the use of integrating CNV workflows into the analysis pipeline, and generated a data resource for medical genomics for the Canadian population encompassing the full spectrum of CNVs.