Low-coverage genome sequencing for the detection of clinically relevant copy-number and mtDNA variants

Abstract

Background: Compared to exome sequencing, genome sequencing is widely appreciated for its superior ability to detect a wide range of genetic variations including copy-number variants (CNVs) and mitochondrial (mtDNA) variants. We assessed whether low-coverage genome sequencing, a considerably cheaper approach, would detect clinically relevant CNVs and mtDNA variants and would thus be a cost-efficient supplement to exome sequencing in rare disease diagnostics. Methods: To assess the level of sequencing depth needed for variant detection, first, 30x mean coverage genome sequencing data were subsampled to 0.5x, 1x, 2x, and 4x coverage files in silico followed by CNV and mtDNA detection. Based on the analysis, 2x short-read sequencing was selected to be performed in 16 patients with putatively pathogenic CNVs or mtDNA variants to assess the empirical sensitivity. Results: For CNV calling, 2x coverage was sufficient to detect all heterozygous CNVs greater than 10kb in size from in silico subsampled data. In experimental data, the results were similar, although a 16kb heterozygous deletion was once not detected. Regarding mtDNA variants, 2x coverage sufficed for variant confident variant calling and heteroplasmy assessment for all samples. Conclusions: Low-coverage genome sequencing may be used to complement exome sequencing for simultaneous mtDNA variant and CNV detection.