Kondrashov has investigated the implications of synergistic epistasis with theoretical studies. Other researchers have taken the experimental route, trying to detect whether, in real life, mutations can interact with each other this way. Those tests yielded mixed results, though, perhaps because the effect would not have to be very large to keep a population from succumbing.
Now, however, Kondrashov and his co-authors have put together a statistical case, pulled from the genomes of about 2,000 people and about 300 wild fruit flies, that the effect has been quietly acting on us and other organisms all along. Drawing on knowledge of the species’ mutation rates and other factors, the scientists began by calculating what the distribution of mutations in populations of humans and flies ought to be in the absence of this purging effect. Certain numbers of individuals in the group, for example, ought to show 100, 50 or 30 mutations. Then the group of researchers turned to the genomic data, looking for the distribution of mutations in real-world populations.
What they found was that significantly fewer individuals than expected had large numbers of dangerous mutations. They are missing from the population, “suggesting that at the high end, at the end where people have many deleterious mutations, there’s stronger selection against these people,” said Arjan de Visser, an evolutionary geneticist at University of Wageningen who was not involved in the work. This observation fits well with what should happen if mutations are not acting independently.
That finding comes with some caveats. There does not seem to be any shrinkage in the number of individuals with less-than-devastating mutations, cautioned both Kondrashov and Shamil Sunyaev, a computational geneticist at Harvard Medical School and another senior author of the paper. “We don’t see it for the whole genome,” Sunyaev said, although the decrease is there “at least for mutations that are undoubtedly deleterious in effect.” The team would also like to get better data on the consequences of mutations in parts of the genome that don’t make proteins. That would let them run their statistical tests again with more confidence that the interactions are occurring more broadly.
Still, the evidence is provocative, and the idea elegant. “I always found it quite attractive, biologically,” said Brian Charlesworth, an evolutionary geneticist at University of Edinburgh who was not involved in the study. “If you think about someone getting hit on the head with a hammer, the first few blows might not do you too much harm, but after a while it will finish you off.” Of the new work, he said, “It’s really the first study which comes up with evidence from what’s going on actually out there in natural populations.”
Perhaps the most interesting corollary of this finding, however, is that it might help explain the persistence of sex. Among population geneticists, sexual reproduction is notoriously difficult to justify as an evolutionary strategy. As a sexual organism, even if everything goes well — if you manage to find a mate who accepts you, if you manage to conceive — you will still be passing on only half of your genes. An asexually reproducing organism, having daughters by making perfect copies of itself, gets double the benefit, none of the hassle. Yet clearly, sex continues.