Association analyses of the autosomal genome and mitochondrial DNA with accelerometry-derived sleep parameters in depressed UK biobank subjects

BACKGROUND: The bidirectional relationship between sleep disturbances and depression is well documented, yet the biology of sleep is not fully understood. Mitochondria have become of interest not only because of the connection between sleep and metabolism but also because of mitochondria's involvement in the production of reactive oxygen species, which are largely scavenged during sleep. METHODS: Genome-wide association studies (GWAS) of eight accelerometry-derived sleep measures were performed across both the autosomal and mitochondrial DNA (mtDNA) among two severity levels of depression in UK Biobank participants. We calculated SNP heritability for each of the sleep measures. Linear regression was performed to test associations and mitochondrial haplogroups. RESULTS: Variants included in the GWAS accounted for moderate heritability of bedtime (19.6%, p = 4.75 × 10^-7), sleep duration (16.6%, p = 8.58 × 10^-6) and duration of longest sleep bout (22.6%, p = 4.64 × 10^-4). No variants passed genome-wide significance in the autosomal genome. The top hit in the severe depression sample was rs145019802, near GOLGA8B, for sleep efficiency (p = 1.17 × 10^-7), and the top hit in the broad depression sample was rs7100859, an intergenic SNP, and nap duration (p = 1.25 × 10^-7). Top mtDNA loci were m.12633C > A of MT-ND5 with bedtime (p = 0.002) in the severe sample and m.16186C > T of the control region with nap duration (p = 0.002) in the broad sample. CONCLUSION: SNP heritability estimates support the involvement of common SNPs in specific sleep measures among depressed individuals. This is the first study to analyze mtDNA variance in sleep measures in depressed individuals. Our mtDNA findings, although nominally significant, provide preliminary suggestion that mitochondria are involved in sleep.