Abstract

Eighty per cent of de novo mutations passed on from parents to offspring occur in the male germline. We used PacBio HiFi long-read sequencing to identify mutations in testis and sperm samples from nine primate species: human, chimpanzee, bonobo, gorilla, orangutan, gibbon, baboon, macaque, and marmoset. Based on high-quality scores and auxiliary information from PacBio sequencing, we can confidently identify mutations, short indels, and evidence for genetic exchange (recombination and gene conversion). Furthermore, we developed a likelihood-based approach to classify sequencing reads into cell types based on their CpG methylation patterns, allowing most testis reads to be classified as either germline or somatic. In human sperm samples, we find the mutation rate, mutation spectrum and paternal age effect expected from trio studies. In other primate sperm samples, we find changes to both the spectrum and the rate, providing insights into the evolution of germline mutations. The mutational spectrum of germline mutations in the testis reflects that of the sperm but reveals more non-synonymous changes, suggesting natural selection against mutations in the testis. The somatic mutation rate in the testis is about twice that of the germline. By calling recombination in the same samples, we identify simple crossovers and gene conversions. The gene conversion tract lengths are short (40-100 bp) across species. However, we also observe complex gene conversions across all species. We interpret these as resulting from homologous repair before meiosis, and they typically move several SNVs between haplotypes over 10 kb. Intersecting recombination calling with de novo mutation calling, we directly observe a mutagenic effect of recombination, with the rate of complex gene conversion higher than that of simple gene conversions and crossovers.