History, chance and selection during phenotypic and genomic experimental evolution: replaying the tape of life at different levels

摘要

Ever since Darwin, understanding evolutionary processes and patterns have been major scientific quests. In the Origin of Species , Darwin explained both adaptation and diversity, and most of his arguments were based on indirect evidence, including comparative approaches. These findings led Darwin to defend that evolution in nature is extremely slow and gradual, hardly being directly observable at the scale of a human generation. Artificial selection, in contrast, was used by Darwin to illustrate the efficacy of natural selection (Darwin, 1859 ). During the last decades, evolution has been observed in real time. This opened new research possibilities and gave rise to Experimental Evolution, a rapidly expanding field that covers many topics and organisms (Garland and Rose, 2009 ; Kawecki et al., 2012 ). The joint power of experimental evolution and recently developed genome-wide tools may now lead us a step further in understanding real-time evolutionary dynamics of populations, both at phenotypic and genomic levels (Baldwin-Brown et al., 2014 ; Schl?tterer et al., 2014 ). Our contribution to this special issue of Frontiers in Genetics focuses on the power of these approaches to assess the role of historical contingencies during adaptation to novel environments, a fundamental subject that has been neglected. Laboratory experimental evolution studies are powerful because we can follow evolutionary trajectories of independent replicates for one or more traits while controlling for all but the factors under study. For instance this tool is particularly suited to analyze temporal changes using ancestral populations as baselines, generate contrasting phenotypes by divergent selection, and test for predictability of evolution (Kawecki et al., 2012 ). With such approach, essential questions can be addressed: (1) What is the adaptive potential of populations to novel environments? (2) What is the role of selection and chance during adaptation? (3) What is the tempo and mode of evolution? (4) How constrained is evolution? With the recent advent of high throughput techniques in genome-wide analysis, also available for non-model organisms (Ellegren, 2014 ), the field of evolutionary biology is now addressing essential questions more thoroughly: (1) What is the genetic basis of adaptation? (2) Are there many genes of small effect or few genes of major effect involved? (3) What is the role of genetic drift vs . selection on candidate genes during local adaptation? (4) What is the mutation rate and how does it change during evolution? (5) Does genomic evolution mimic phenotypic evolution in timing and pattern? (Orr, 2005 ; Stapley et al., 2010 ). Until recently, only population genetics modeling and comparative analyses across populations addressed these questions. Both present strengths and limitations (Magalh?es and Matos, 2012 ). The combination of experimental evolution and genomic techniques allows unprecedented resolution to the evolutionary mechanisms underlying phenotypic and genomic change (Burke, 2012 ; Burke and Long, 2012 ; Dettman et al., 2012 ; Lobkovsky and Koonin, 2012 ; Barrick and Lenski, 2013 ; Baldwin-Brown et al., 2014 ; Schl?tterer et al., 2014 ). In particular, these approaches may help us disentangle the role of historical contingencies, the effect of chance events, and the power of selection during adaptation to novel environments. Selection, history and chance are not mutually exclusive. It is, thus, of utmost importance to define their relative roles in shaping evolution in general, and adaptation to novel environments in particular (Bedhomme et al., 2013 ). While selection is seen as a deterministic process leading to adaptation, both previous history and chance events are evolutionary contingencies that may lead to disparate, unpredictable results (Lenormand et al., 2009 ). The classic question of whether evolution is repeatable if we “replay” the tape of life (Gould, 1989 ) can now be more thoroughly addressed from phenotypes to genomes (Lobkovsky and Koonin, 2012 ). The relative role of chance and selection can be tackled by analyzing differences between populations that start from the same ancestral population, while evolving in a novel environment. If selection plays the most important role, it is expected that populations will evolve in parallel, both phenotypically and genotypically, with chance events (e.g., founder effects, genetic drift, and random mutations) having relatively reduced impact. Experimental evolution studies have shown abundant examples of parallel genomic evolution (Lobkovsky and Koonin, 2012 ; Stern, 2013 ), although this is not always the case (Arendt and Reznick, 2008 ; Elmer and Meyer, 2011 ). Conspicuous in their abundance are studies of parallel genomic evolution in asexual microorganisms (Tenaillon et al., 2012 ; Barrick and Lenski, 2013 ). In experiments starting from the same clone evolution depends on de novo mutations (due to a lack of standing genetic varia