Large, three-generation CEPH families reveal post-zygotic mosaicism and variability in germline mutation accumulation

Abstract

AbstractThe number of de novo mutations (DNMs) found in an offspring’s genome is known to increase with both paternal and maternal age. But does the rate of mutation accumulation in parental gametes differ across families? To answer this question, we analyzed DNMs in 33 large, three-generation families collected in Utah by the Centre d’Etude du Polymorphisme Humain (CEPH) consortium. We observed significant variability in parental age effects on DNM counts across families, ranging from 0.24 to 3.33 additional DNMs per year. Using up to 14 grandchildren in these families, we find that 3% of DNMs originated following primordial germ cell specification (PGCS) in a parent, and differ from non-mosaic germline DNMs in their mutational spectra. We also identify a median of 3 gonosomal mutations per sample in the F1 generation, which, along with post-PGCS DNMs, occur at equivalent frequencies on the paternal and maternal haplotypes. These results demonstrate that the rate of germline mutation accumulation varies among families with similar ancestry, and confirm that parental mosaicism is a substantial source of de novo mutations in children.Data and code availabilityCode used for statistical analysis and figure generation has been deposited on GitHub as a collection of annotated Jupyter Notebooks: https://github.com/quinlan-lab/ceph-dnm-manuscript. Data files containing germline de novo mutations, as well as the gonosomal and post-primordial germ cell specification (PGCS) mosaic mutations, are included with these Notebooks. To mitigate compatibility/version issues, we have also made all notebooks available in a Binder environment, accessible at the above GitHub repository.