Global human frequencies of predicted nuclear pathogenic variants and the role played by protein hydrophobicity in pathogenicity potential

Mitochondrial proteins are coded by nuclear (nDNA) and mitochondrial (mtDNA) genes, implying a complex cross-talk between the two genomes. Here we investigated the diversity displayed in 104 nuclear-coded mitochondrial proteins from 1,092 individuals from the 1000 Genomes dataset, in order to evaluate if these genes are under the effects of purifying selection and how that selection compares with their mitochondrial encoded counterparts. Only the very rare variants (frequency < 0.1%) in these nDNA genes are indistinguishable from a random set from all possible variants in terms of predicted pathogenicity score, but more frequent variants display distinct signs of purifying selection. Comparisons of selection strength indicate stronger selection in the mtDNA genes compared to this set of nDNA genes, accounted for by the high hydrophobicity of the proteins coded by the mtDNA. Most of the predicted pathogenic variants in the nDNA genes were restricted to a single continental population. The proportion of individuals having at least one potential pathogenic mutation in this gene set was significantly lower in Europeans than in Africans and Asians. This difference may reflect demographic asymmetries, since African and Asian populations experienced main expansions in middle Holocene, while in Europeans the main expansions occurred earlier in the post-glacial period.