2014/01/29 by Stephen Wright · 1 citation
Arts and Humanities · #Evolution and Science Education #Biology #Lens (geology) #Mutation #Through-the-lens metering #Evolutionary biology #Genetics #Paleontology #Gene
paper · pdf · doi:10.1111/evo.12369
published in Evolution 68(4), 1225-1227 (Oxford University Press)
openalex publication_date 2014/01/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/30
Since its inception, a key goal of evolutionary genetics has been to improve our understanding of the relative importance of natural selection, genetic drift, and mutation in structuring genetic variation within and between species. Investigators have also tried to establish the extent to which the rate of adaptation is limited by mutation, compared with natural selection acting on standing genetic variation. These fundamental questions have been among the most controversial in evolutionary biology, and remain so to this day. For much of the history of molecular population genetics, the prevailing view was that most of the genome conformed to Kimura's neutral theory, which proposes that the vast majority of the polymorphism and divergence across the genome is governed primarily by neutral mutation and genetic drift. According to this view, positive selection is assumed to be rare at the molecular level, and when it does occur it was assumed to be transient; a new mutation arising in a population spreads rapidly causing a “selective sweep” of neutral genetic variation. With the recent advent of genome-scale datasets on molecular variation, the field has experienced a significant change in perspective on the balance of these forces. For example, genome-wide population genomics data suggest a high incidence of positive selection on nonsynonymous and regulatory mutations in some taxa (Sella et al. 2009). Furthermore, there is growing evidence that a predominant mode of adaptation in many systems may reflect selection from preexisting (or readily available) variation, or “soft sweeps” rather than new mutations (Pritchard et al. 2010; Messer and Petrov 2013). Much of this selection on standing variation may be targeted to polygenic traits with complex genetic architecture. Although far from resolved, there is a growing shift in emphasis in the field of molecular evolution toward the quantitative geneticist's view of evolution, which incorporates both selection from standing variation and, once that variation is exhausted, new mutations. Given these changing perspectives, I was intrigued to review Mutation-Driven Evolution, the new book by Masatoshi Nei, one of the founders of the field of molecular population genetics. In contrast with the evolving perspectives discussed above, the principal goal of Nei's volume is to make a strong case for the preeminent role of mutation over natural selection in evolution. The main thesis of the book is that not only genomic but, more controversially, also phenotypic evolution are largely governed by neutral processes, and that adaptation rarely acts on preexisting variation but relies on new major mutations, where the genomic, developmental, and mutational details matter for guiding evolutionary trajectories. In the first several chapters of the book, Nei examines the history of evolutionary thought and attempts to question the widespread importance of natural selection by highlighting the challenges of measuring it. Temporal and spatial heterogeneity are said to complicate estimates of natural selection in the wild, and violate simple theories, such as Fisher's fundamental theorem, that predict a continuous increase in mean fitness over time. Nei dismisses the notion that natural selection is a creative process as a rhetorical argument without grounding in either theory or empirical evidence, flying in the face of the persuasive arguments to the contrary advanced a long time ago by Charles Darwin, R. A. Fisher, and others since. This section certainly highlights some limitations of simple models of constant directional selection with unlimited genetic variation, and of inferences about phenotypic selection in the wild based on such assumptions. There is no doubt that our picture of natural selection is incomplete, including a lack of detailed knowledge of the nature of selective agents, the extent of fluctuating selection, and the genetic basis of adaptation in other than a few cases. However, Nei does not discuss how far these limitations simply reflect the difficulty in obtaining the relevant data. He attempts to minimize the overwhelming body of evidence for natural selection acting on traits of organisms in the wild (e.g., Endler 1986), but fails to show how any process other than selection can explain the evolution of complex adaptations. Subsequent chapters incorporating evidence from genomics, developmental biology, population genetics, and speciation research are similarly unconvincing. Let's consider three examples. First, when reviewing the molecular population genetics literature, Nei highlights the key limitations and assumptions built into tests for positive selection. The test proposed by McDonald and Kreitman (1991), for example, can lead to spurious evidence for positive selection if populations have experienced population bottlenecks. If taxa have experienced recurrent bottlenecks, much of the signal for widespread positive selection may be erroneously inferred. However, Nei does not provide evidence for the near-universality of historical population bottlenecks in those species that show genome-wide positive selection signals. This seems especially implausible for taxa such as Drosophila, with their large population sizes, from which much of this evidence has been obtained. Again, I would agree that the limits and potential biases of tests of selection are very important to consider, but I cannot see why this should lead to a strong rejection in favor of neutrality and mutationism. Second, Nei emphasizes the importance of genes of major effect in developmental evolution, citing classic examples such as the MC1R locus in controlling coat color evolution in mammals. He again emphasizes the importance of the mutational details (i.e., the numbers of genes involved, their effect sizes, and their molecular basis) in our understanding of adaptation, but why they actually matter and how these examples demonstrate the preeminence of mutation is not made clear. Furthermore, Nei does not explore the fact that genes of major effect are the easiest to characterize, and ignores the possibility of the widespread importance of genes of small effect that may never be functionally identified (Rockman 2012). Strong polygenic selection on phenotypes can lead to subtle selection signals, and these signals may require large-scale datasets to detect (e.g., Turchin et al. 2012). Strong selection on phenotypes may often not translate into a strong genomic signal, but this does not imply the neutrality of the traits in question, and the widespread evidence of selection on phenotypic traits in the wild highlights this. Furthermore, even large-effect mutations still require the action of natural selection to spread. Similarly, the evidence for the preeminent role of selection in the speciation process is discussed but considered weak. But the only argument I can find for this relates to our limited knowledge of the genetics of speciation and the problems of molecular population genetics tests for selection. But limited knowledge and the limits of tests do not translate into a strong statement about the role of drift versus selection, and Nei avoids discussion of important evidence for the role of ecological and behavioral factors in driving pre- and postzygotic isolation (Schluter 2009). In the concluding chapters, Nei attempts to solidify the argument for mutation-driven evolution. The argument being put forward follows two main paths. First, evolution cannot occur without mutation, but it can proceed without selection. While this is clearly the case, the same cannot be said for adaptation and, despite claims to the contrary, I fail to identify any example in the book of the evolution of adaptation without the action of natural selection. Second, selection plays a relatively minor role by eliminating individuals with the less fit genotypes. Why is this minor? This is exactly the force that causes allele frequency changes of beneficial mutations; since mutation rates are typically low and population sizes of species are often large, only selection can cause allele frequencies and trait means to change over short timescales. Nei's viewpoint is essentially a philosophical position, which puts the source of the raw materials (mutation) as primary over the agent that drives the evolution of a highly functioning organism (selection). But by analogy, if we were to ask why houses are found on a street, the availability of bricks cannot but be a secondary explanation compared with the need for shelter. In general, Nei emphasizes that we have an incomplete picture of the role of natural selection at the genome level, in trait evolution, and in speciation. In Bayesian terms, Nei's prior probability is strongly in favor of mutationism, and the evidence for natural selection is seen as too weak for the posterior to change from his prior. The limitations of models and tests of natural selection are clearly important, and this book highlights the importance of addressing these. The exciting challenge over the next decade will be to improve our understanding of rates of adaptation, including the relative role of selection on new mutations versus standing variation, the role of small population size in limiting adaptation, and the extent to which the rate of adaptation is governed by mutation. But discussion and exploration of these limitations seems a far cry from convincingly putting mutation into the driver's seat of evolution, and the challenging task of genome-wide quantification of the balance of evolutionary forces remains before us. I am grateful to S. C. H. Barrett, B. Charlesworth, D. Charlesworth, G. Coop, C. Bergman, and A. Ågren for important comments on this review. Associate Editor: M. Wade