Evolution depends on de novo mutations (DNMs) in offspring of each generation

Children (i.e. the next generation) receive half their DNA from each parent; however, the inherited DNA is not identical to the corresponding half of the parents’ genetic material. Instead, both the egg and sperm — which combine to generate an embryo — carry so-called ‘germline de novo’ mutations that are not present in any of the parents’ cells. Although these de novo mutations (DNMs) are an important source of genetic diversity (i.e. this is important for evolution; survival of any species depends on mutations that help defend against changing environmental adversity) — they can also cause disease. Geneticists have a long-standing interest in how, when, and at what rate, germline DNMs arise. These questions have commonly been addressed by analyzing the DNA of large cohorts of two-generation families.

Authors [see first attached article] used the genetic data of 33 families (from Utah, U.S.), which all span three generations, to determine the rate at which DNMs appear. Their data revealed that, on average, each person has ~70 DNMs that were not present in either parent’s genetic code. Authors found that sperm and egg cells from older parents typically contain more DNMs. However, this effect varied substantially across families (in some, an increase of one year in the parents’ age resulted in >3 DNMs in their children; in others, the number of DNMs barely increased at all). Authors also found that almost 10% of DNMs do not occur in the parents’ sperm or eggs — but appear in the embryo very soon after fertilization; these mutations can lead to ‘mosaicism’ (resulting in a person having a mutation in some, but not all, of their organs and tissues). [In some cases, this could cause an unknown number of sperm and egg cells to carry a mutation that others do not. This would make it hard to predict how likely two or more siblings are to inherit the mutation.] This analysis reveals that parental age affects the number of DNMs in children, but this effect varies from family to family. This finding [see first attached article] could point to genetic or environmental factors that alter human mutation rates.

Authors [see second attached article] suggest they have found evidence for the environment playing a role in DNM rate. DNMs are implicated in many diseases — including rare genetic disorders (e.g. Hutchinson-Gilford progeria, fibrodysplasia ossificans progressiva, von Hippel-Lindau disease, microcephaly, Marfan syndrome) and common complex diseases (e.g. autism and schizophrenia). Recently, modern sequencing technologies have enabled the use of pedigree data to directly estimate the number of DNMs found across the genome; as described above, these pedigree-based studies have identified both paternal and maternal age effects (i.e. estimated contributions of 1.51 and 0.37 DNMs per year of paternal and maternal age, respectively). These effects have been variously attributed to DNA replication errors, DNA repair processes, specific mutation patterns such as C—>G transversions, distinct clusters associated with DNA recombination, chromatin structure, common locations of meiotic crossovers, and even stretches of DNA exhibiting “heterozygote instability.”

Because most DNM studies have used data from small cohorts of individuals of predominantly European ancestry — little is known about the role of DNMs in evolution and health of populations of predominantly non-European ancestry, i.e. might DNM rates vary across different human populations?). To address this, authors (see second & third attached articles) used a high-coverage whole-genome sequencing (WGS) dataset — to directly estimate and analyze DNM accumulation across numerous human ancestries and populations. After analyzing genome-wide patterns of mutation using a call set of 93,325 single-nucleotide variant (SNV) DNMs, across 1,465 trios from an array of diverse human populations, authors used them to directly estimate and analyze DNM counts, rates, and spectra.

Authors discovered a significant positive correlation between local recombination rate and local DNM rate, and that DNM rate explains a substantial portion (between 9% and 35%) of genome-wide variation from 41,000 unrelated samples. Genome-wide heterozygosity is correlated with DNM rate, but only explains <1% of variation. Whereas authors are underpowered to see small differences, they did not find significant differences in DNM rate between individuals of European, African, and Latino ancestry, nor across ancestrally distinct segments within admixed individuals. Intriguingly, authors did find statistically significantly fewer DNMs in Amish individuals — even when compared with other Europeans, and even after accounting for parental age and sequencing center. Specifically, authors found significant decreases in the number of C—>A and T—>C mutations in the Amish, which seem to underpin their overall reduction in DNMs. Lastly, authors (see second & third attached articles) calculated near-zero estimates of narrow-sense heritability (h2); this interesting finding suggests that variation in DNM rate is significantly shaped by non-additive genetic effects, plus the environment. 😊


eLife 2019; 8: e46922 and Proc Natl Acad Sci USA 4 Feb 2020 117: 2560-2569

This entry was posted in Center for Environmental Genetics. Bookmark the permalink.