Most of the phenomena Charney terms “neogenetic” entail changes to the genetic substrate and are therefore classifiable as kinds of mutation. Epigenetic marks differ because they influence gene expression, not sequence. Somatic mutations during development can lead to genetic mosaicism, while epigenetic modifications presumably underlie cellular differentiation (Ng & Gurdon Reference Ng and Gurdon2008). Mutations in the germline can lead to non-Mendelian inheritance of portions of the genome (reviewed in Burt & Trivers Reference Burt and Trivers2006), and epigenetic changes can create unusual patterns of expression such as polar overdominance (Cockett et al. Reference Cockett, Jackson, Shay, Farnir, Berghmans, Snowder, Nielsen and Georges1996). Charney argues that these peculiarities undermine genotype to phenotype maps implicit in the statistical frameworks of behavior genetics. Rather than venture criticism, I here expand on the theoretical challenge presented by one aspect of mutation, simply that of its ubiquity.

Mutations can be understood as products either of necessity or chance. Charney's emphasis on adaptive phenotypic plasticity is apposite. To the extent that mutations are functional, they are instances of adaptedness per se, rather than drivers of adaptation (for an explanation of this perspective, see Dickins & Dickins Reference Dickins and Dickins2008). Charney also describes well the manner in which adventitious changes are harnessed through somatic hypermutation in the immune system. But he says “stochasticity continually threatens to undermine the order imposed by biological systems.” This is surely so: Severe and multifarious mutations affecting the genome within lineages (at least some of which are associated with behavioral dysfunction: e.g., Stewart et al. Reference Stewart, Hall, Kang, Shaw and Beaudet2011) would seem to threaten extinction. How, for example, does a brain maintain its function or a population of organisms preserve its genomic integrity, despite frequent aneuploidy?

Let us consider only “conventional” mutations, namely, single nucleotide changes or indels (insertions and deletions), of the kind that can be associated with Mendelian disorders. Many such mutations are deleterious, and even mildly deleterious mutations can lead to extinction when they accumulate under the influence of genetic drift (Muller Reference Muller1964). Even assuming an infinite population size, mutation accumulation can lead to population extinction with a probability that depends not on the average effect size of mutations but on their rate of occurrence and on the intrinsic fecundity of individuals (Bull et al. Reference Bull, Sanjuán and Wilke2007). Humans have slow life histories (Robson & Wood Reference Robson and Wood2008), a low effective population size (Yu et al. Reference Yu, Jensen-Seaman, Chemnick, Ryder and Li2004), and a relatively high mutation rate (Kondrashov Reference Kondrashov2003; Lynch Reference Lynch2010), although they benefit from recombination. Given the proportion of de novo mutations expected to be deleterious, and with advances in medical care that plausibly entail a relaxation of negative selection, these parameters have led to concerns about population fitness in the medium term (Lynch Reference Lynch2010; but see Keightley Reference Keightley2012). Low frequency alleles with relatively large effect sizes may also underlie at least some of the missing heritability in genome-wide association studies (Manolio et al. Reference Manolio, Collins, Cox, Goldstein, Hindorff, Hunter, McCarthy, Ramos, Cardon, Chakravarti, Cho, Guttmacher, Kong, Kruglyak, Mardis, Rotimi, Slatkin, Valle, Whittemore, Boehnke, Clark, Eichler, Gibson, Haines, Mackay, McCarroll and Visscher2009), thereby contributing to the burden of disease at this time.

The reduction in the mean fitness of a population caused by mutations (the mutation load) is attenuated when the fitness of an individual depends on others around it; this is so-called soft selection (Wallace Reference Wallace1975). Alternatively, germline viability selection can expurgate deleterious alleles, an observation made plausible in humans by the seemingly high rate of “occult” pregnancies (Edmonds et al. Reference Edmonds, Lindsay, Miller, Williamson and Wood1982), with many aborted concepti manifesting aneuploidies (Macklon et al. Reference Macklon, Geraedts and Fauser2002). The occurrence of mitotic and meiotic cell divisions in the germline of “sexual” species has significant population genetic consequences (Hastings Reference Hastings1991). For example, when germline selection is soft, this can favor the evolution of “anti-robustness” in which genotypes readily suffer reduced fitness when mutated (Archetti Reference Archetti2009). Anti-robust genotypes are also expected in regenerative tissues for theoretical reasons (Krakauer & Plotkin Reference Krakauer and Plotkin2002) paradoxically contributing to robustness at the level of the organism.

Emerging evidence supports purifying selection in the mammalian mitochondrial genome (Fan et al. Reference Fan, Waymire, Narula, Li, Rocher, Coskun, Vannan, Narula, Macgregor and Wallace2008; Stewart et al. Reference Stewart, Freyer, Elson, Wredenberg, Cansu, Trifunovic and Larsson2008), which is probably facilitated by a germline bottleneck in copy number. We are beginning to understand how cell lineages behave in mammalian oogenesis (Reizel et al. Reference Reizel, Itzkovitz, Adar, Elbaz, Jinich, Chapal-Ilani, Maruvka, Nevo, Marx, Horovitz, Wasserstrom, Mayo, Shur, Benayahu, Skorecki, Segal, Dekel and Shapiro2012), but important details are unresolved. In the male germline, recent evidence suggests positive selection for specific genetic disorders (Choi et al. Reference Choi, Yoon, Calabrese and Arnheim2012; Goriely et al. Reference Goriely, McVean, Röjmyr, Ingemarsson and Wilkie2003). One response to these preliminary data is to conceive of disease as an outcome of a breakdown in the regulation and control of deleterious mutations. Focusing on the regulation of mitochondrial function, one colorful review elaborates such a “quality-control” perspective (Braschi & McBride Reference Braschi and McBride2010). Although Charney suggests that transposon activation in the brain might be positively associated with neural plasticity and flexibility, it may prove necessary to consider how neural networks buffer themselves from the deleterious effects of mutations or even how behavior itself might modulate mutational effects.

Some of the phenomena Charney describes might threaten stability, not because of their passive accumulation but because they are selected for independently of their effect on the rest of the genome. For example, transposons active in the germline increase their contribution to posterity by overreplication, but can damage genes if they “jump” close by or into them. When intragenomic conflict occurs, and a component of the genome acts against the wider coalition, the evolution of repressors is favored by natural selection (Burt & Trivers Reference Burt and Trivers2006). Aspects of meiosis such as reduction division and recombination may have evolved to restrain selfish genetic elements.

Many forms of epigenetic regulation were revealed during research into genomic imprinting. Imprinted gene expression can be explained by conflict between paternally and maternally derived alleles within offspring (Haig & Westoby Reference Haig and Westoby1989) or as a consequence of maternal–fetal co-adaptation (Keverne & Curley Reference Keverne and Curley2008). From the conflict perspective, imprinted gene expression does not benefit the organism, but potentially creates problems in the event of epigenetic dsyregulation or mutation (due to haploinsufficiency at the expressing allele). Aberrant patterns of imprinted gene expression are implicated in intrauterine growth restriction (Tycko Reference Tycko2006), but recent screens (reviewed in Kelsey Reference Kelsey2011) have suggested abundant imprinted expression in the brain. Trivers (Reference Trivers2000) has also outlined ways in which intragenomic conflict can manifest itself in the behavior of individuals.