Abstract
AbstractBackground/ObjectivesWhole genome sequencing (WGS) from large cohorts enables the study of mitochondrial DNA (mtDNA) variation on human health. We aimed to investigate the influence of common, rare, and pathogenic mtDNA variants on 15 mitochondrial disease-related phenotypes.MethodsUsing WGS from 179,862 individuals from in the UK Biobank, we identified mtDNA variants using MitoHPC. We performed extensive association analyses with 15 mitochondrial disease-relevant phenotypes. We compared the results for the m.3243A>G variant with those from a clinically referred patient cohort.ResultsOf 15,881 mtDNA variants, 12 homoplasmic and one heteroplasmic variant had genome-wide significant associations. All homoplasmic variants increased aspartate aminotransferase level and three were novel, low frequency, variants (MAF∼0.002 and beta∼0.3 SD). Only m.3243A>G (MAF=0.0002) associated with diabetes (OR=5.6, 95%CI [3.2-9.9]), deafness (OR=12.3, 95%CI [6.2-24.4]) and heart failure (OR=39.5, 95%CI [9.76-160.1]). Multi-system disease risk and penetrance of all three traits increased with m.3243A>G level. Diabetes risk was further influenced by common nuclear genome variation. The penetrance of diabetes with m.3243A>G in the UK Biobank was lower than clinically referred patients, partly attributed to lower heteroplasmy. Of 73 pathogenic mitochondrial disease variants, most were rare in the population with low penetrance.ConclusionOur study highlights the utility of WGS for investigating mitochondrial genetics within a large, unselected population. We identified novel associations and demonstrated that pathogenic mitochondrial variants have lower penetrance in clinically unselected than clinically referred settings. m.3243A>G associated with mitochondrial-related phenotypes at higher heteroplasmy. Our findings suggest potential benefits of reporting incidentally identified m.3243A>G at high heteroplasmy levels.
Publisher
Cold Spring Harbor Laboratory