From whole bodies to single cells: A guide to transcriptomic approaches for ecology and evolutionary biology.

RNA sequencing (RNAseq) methodology has experienced a burst of technological developments in the last decade, which has opened up opportunities for studying the mechanisms of adaptation to environmental factors at both the organismal and cellular level. Selecting the most suitable experimental approach for specific research questions and model systems can, however, be a challenge and researchers in ecology and evolution are commonly faced with the choice of whether to study gene expression variation in whole bodies, specific tissues, and/or single cells. A wide range of sometimes polarised opinions exists over which approach is best. Here, we highlight the advantages and disadvantages of each of these approaches to provide a guide to help researchers make informed decisions and maximise the power of their study. Using illustrative examples of various ecological and evolutionary research questions, we guide the readers through the different RNAseq approaches and help them identify the most suitable design for their own projects.

Reproductive isolation arises during laboratory adaptation to a novel hot environment.

BACKGROUND: Reproductive isolation can result from adaptive processes (e.g., ecological speciation and mutation-order speciation) or stochastic processes such as "system drift" model. Ecological speciation predicts barriers to gene flow between populations from different environments, but not among replicate populations from the same environment. In contrast, reproductive isolation among populations independently adapted to the same/similar environment can arise from both mutation-order speciation or system drift. RESULTS: In experimentally evolved populations adapting to a hot environment for over 100 generations, we find evidence for pre- and postmating reproductive isolation. On one hand, an altered lipid metabolism and cuticular hydrocarbon composition pointed to possible premating barriers between the ancestral and replicate evolved populations. On the other hand, the pronounced gene expression differences in male reproductive genes may underlie the postmating isolation among replicate evolved populations adapting to the same environment with the same standing genetic variation. CONCLUSION: Our study confirms that replicated evolution experiments provide valuable insights into the mechanisms of speciation. The rapid emergence of the premating reproductive isolation during temperature adaptation showcases incipient ecological speciation. The potential evidence of postmating reproductive isolation among replicates gave rise to two hypotheses: (1) mutation-order speciation through a common selection on early fecundity leading to an inherent inter-locus sexual conflict; (2) system drift with genetic drift along the neutral ridges.

Evolution of Phenotypic Variance Provides Insights into the Genetic Basis of Adaptation.

Most traits are polygenic, and the contributing loci can be identified by genome-wide association studies. The genetic basis of adaptation (adaptive architecture) is, however, difficult to characterize. Here, we propose to study the adaptive architecture of traits by monitoring the evolution of their phenotypic variance during adaptation to a new environment in well-defined laboratory conditions. Extensive computer simulations show that the evolution of phenotypic variance in a replicated experimental evolution setting can distinguish between oligogenic and polygenic adaptive architectures. We compared gene expression variance in male Drosophila simulans before and after 100 generations of adaptation to a novel hot environment. The variance change in gene expression was indistinguishable for genes with and without a significant change in mean expression after 100 generations of evolution. We suggest that the majority of adaptive gene expression evolution can be explained by a polygenic architecture. We propose that tracking the evolution of phenotypic variance across generations can provide an approach to characterize the adaptive architecture.

Unraveling the Molecular Basis of Stabilizing Selection by Experimental Evolution.

Stabilizing selection provides a challenge to molecular population genetics. Although stabilizing selection is ubiquitous, its genomic signature is difficult to distinguish from demographic signals. Experimental evolution provides a promising approach to characterize genomic regions exposed to stabilizing selection. A recent experimental evolution study of Aedes aegypti populations evolving either with or without sexual selection found a pattern of genetic differentiation suggestive of relaxed stabilizing selection. I argue that this study could not have detected the signal of relaxed stabilizing selection. I highlight why incorrect statistical methods resulted in a high number of false positive candidate single nucleotide polymorphism (SNPs) and discuss the fallacy of functional validation of candidate SNPs for polygenic traits by RNA-mediated knockdown.

Non-additive effects between genotypes: Implications for competitive fitness assays.

Competitive fitness assays are widely used in evolutionary biology and typically rely on a reference strain to compare different focal genotypes. This approach implicitly relies on the absence of interaction between the competing genotypes. In other words, the performance of the reference strain must not depend on the competitor. This report scrutinized this assumption by competing diverged Drosophila simulans populations against a common reference strain. We detected strong evidence for interaction between the competing genotypes: (1) Frequency-dependent selection was common with opposite effects in genetically diverged populations. (2) Temporal heterogeneity of fitness estimates, which can be partially attributed to a competitor-specific delay in the eclosion of the reference strain. We propose that this inconsistent behavior of the reference strain can be considered a specific case of a genotype × environment interaction. Focal populations could modify the environment of the reference strain, either indirectly by altering the microbiome composition and food availability or directly by genotype-specific cannibalism. Our results provide new insights into the interaction of diverged genotypes and have important implications for the interpretation of competitive fitness assays.

Effects of larval crowding on the transcriptome of Drosophila simulans.

Larval crowding is one common ecological stressor for many insect species. In Drosophila, high larval density alters multiple widely-studied phenotypes including life-history traits, morphology and behavior. Nevertheless, we still miss a holistic view of the full range of phenotypic changes and the underlying molecular mechanisms. In this study, we analyzed the adult transcriptomes of high and low larval density fly cohorts, and highlighted the molecular basis of the plastic traits. Increased cellular energy metabolism and locomotion, along with reduced reproductive investment, are key responses to high larval density. Moreover, we compared the expression changes among cohorts with different developmental delays caused by larval crowding. The majority of genes induced by larval crowding showed the strongest expression alterations in cohorts with intermediate delay. Furthermore, linear expression changes were observed in genes related to nutrition and detoxification. Comparing different high-density cohorts could provide insights into the varied responses to distinct larval crowding-induced stresses such as space competition, food degradation and waste accumulation.

The genomic distribution of transposable elements is driven by spatially variable purifying selection.

It is widely accepted that the genomic distribution of transposable elements (TEs) mainly reflects the outcome of purifying selection and insertion bias (1). Nevertheless, the relative importance of these two evolutionary forces could not be tested thoroughly. Here, we introduce an experimental system, which allows separating purifying selection from TE insertion bias. We used experimental evolution to study the TE insertion patterns in Drosophila simulans founder populations harboring 1040 insertions of an active P-element. After 10 generations at a large population size, we detected strong selection against P-element insertions. The exception were P-element insertions in genomic regions for which a strong insertion bias has been proposed (2-4). Because recurrent P-element insertions cannot explain this pattern, we conclude that purifying selection, with variable strength along the chromosomes, is the major determinant of the genomic distribution of P-elements. Genomic regions with relaxed purifying selection against P-element insertions exhibit normal levels of purifying selection against base substitutions. This suggests that different types of purifying selection operate on base substitutions and P-element insertions. Our results highlight the power of experimental evolution to understand basic evolutionary processes, which are difficult to infer from patterns of natural variation alone.

Temperature-driven gene expression evolution in natural and laboratory populations highlights the crucial role of correlated fitness effects for polygenic adaptation.

The influence of pleiotropy on adaptive responses is a highly controversial topic, with limited empirical evidence available. Recognizing the pivotal role of the correlation of fitness effects, we designed an experiment to compare the adaptive gene expression evolution of natural and experimental populations. To test this, we studied the evolution of gene expression in response to temperature in two Drosophila species on a natural temperature cline in North America and replicated populations evolving in hot- and cold-temperature regimes. If fitness effects of affected traits are independent, pleiotropy is expected to constrain the adaptive response in both settings, laboratory and natural populations. However, when fitness effects are more correlated in natural populations, adaptation in the wild will be facilitated by pleiotropy. Remarkably, we find evidence for both predicted effects. In both settings, genes with strong pleiotropic effects contribute less to adaptation, indicating that the majority of fitness effects are not correlated. In addition, we discovered that genes involved in adaptation exhibited more pleiotropic effects in natural populations. We propose that this pattern can be explained by a stronger correlation of fitness effects in nature. More insights into the dual role of pleiotropy will be crucial for the understanding of polygenic adaptation.

Evolution of Metabolome and Transcriptome Supports a Hierarchical Organization of Adaptive Traits.

Most organismal phenotypes have a polygenic basis, which enables adaptive phenotypic responses on ecological time scales. While adaptive phenotypic changes are highly parallel in replicate populations, this does not apply to the contributing loci. In particular for small populations, the same phenotypic shift can be fueled by different sets of alleles at alternative loci (genetic redundancy). Although this phenomenon is empirically well supported, the molecular basis of the genetic redundancy is not yet understood. To fill this gap, we compared the heterogeneity of the evolutionary transcriptomic and metabolomic response in ten Drosophila simulans populations which evolved parallel high-level phenotypic changes in a novel temperature environment but used different allelic combinations of alternative loci. We showed that the metabolome evolved more parallel than the transcriptome, confirming a hierarchical organization of molecular phenotypes. Different sets of genes responded in each evolved population but led to the enrichment of similar biological functions and a consistent metabolic profile. Since even the metabolomic response was still highly heterogeneous across evolved populations, we propose that selection may operate on pathways/networks.

How predictable is adaptation from standing genetic variation? Experimental evolution in Drosophila highlights the central role of redundancy and linkage disequilibrium.

Experimental evolution is well-suited to test the predictability of evolution without the confounding effects of inaccurate forecasts about future environments. Most of the literature about parallel (and thus predictable) evolution has been carried out in asexual microorganisms, which adapt by de novo mutations. Nevertheless, parallel evolution has also been studied in sexual species at the genomic level. Here, I review the evidence for parallel evolution in Drosophila, the best-studied obligatory outcrossing model for adaptation from standing genetic variation in the laboratory. Similar to asexual microorganisms, evidence for parallel evolution varies between the focal hierarchical levels. Selected phenotypes consistently respond in a very predicable way, but the underlying allele frequency changes are much less predictable. The most important insight is that the predictability of the genomic selection response for polygenic traits depends highly on the founder population and to a much lesser extent on the selection regime. This implies that predicting adaptive genomic response is challenging and requires a good understanding of the adaptive architecture (including linkage disequilibrium) in the ancestral populations. This article is part of the theme issue 'Interdisciplinary approaches to predicting evolutionary biology'.

Genome-wide selection signatures reveal widespread synergistic effects of two different stressors in Drosophila melanogaster.

Experimental evolution combined with whole-genome sequencing (evolve and resequence (E&R)) is a powerful approach to study the adaptive architecture of selected traits. Nevertheless, so far the focus has been on the selective response triggered by a single stressor. Building on the highly parallel selection response of founder populations with reduced variation, we evaluated how the presence of a second stressor affects the genomic selection response. After 20 generations of adaptation to laboratory conditions at either 18°C or 29°C, strong genome-wide selection signatures were observed. Only 38% of the selection signatures can be attributed to laboratory adaptation (no difference between temperature regimes). The remaining selection responses are either caused by temperature-specific effects, or reflect the joint effects of temperature and laboratory adaptation (same direction, but the magnitude differs between temperatures). The allele frequency changes resulting from the combined effects of temperature and laboratory adaptation were more extreme in the hot environment for 83% of the affected genomic regions-indicating widespread synergistic effects of the two stressors. We conclude that E&R with reduced genetic variation is a powerful approach to study genome-wide fitness consequences driven by the combined effects of multiple environmental factors.

The Transposition Rate Has Little Influence on the Plateauing Level of the P-element.

The popular trap model assumes that the invasions of transposable elements (TEs) in mammals and invertebrates are stopped by piRNAs that emerge after insertion of the TE into a piRNA cluster. It remains, however, still unclear which factors influence the dynamics of TE invasions. The activity of the TE (i.e., transposition rate) is one frequently discussed key factor. Here we take advantage of the temperature-dependent activity of the P-element, a widely studied eukaryotic TE, to test how TE activity affects the dynamics of a TE invasion. We monitored P-element invasion dynamics in experimental Drosophila simulans populations at hot and cold culture conditions. Despite marked differences in transposition rates, the P-element reached very similar copy numbers at both temperatures. The reduction of the insertion rate upon approaching the copy number plateau was accompanied by similar amounts of piRNAs against the P-element at both temperatures. Nevertheless, we also observed fewer P-element insertions in piRNA clusters than expected, which is not compatible with a simple trap model. The ping-pong cycle, which degrades TE transcripts, becomes typically active after the copy number plateaued. We generated a model, with few parameters, that largely captures the observed invasion dynamics. We conclude that the transposition rate has at the most only a minor influence on TE abundance, but other factors, such as paramutations or selection against TE insertions are shaping the TE composition.

Natural variation in Drosophila shows weak pleiotropic effects.

BACKGROUND: Pleiotropy describes the phenomenon in which a gene affects multiple phenotypes. The extent of pleiotropy is still disputed, mainly because of issues of inadequate power of analyses. A further challenge is that empirical tests of pleiotropy are restricted to a small subset of all possible phenotypes. To overcome these limitations, we propose a new measurement of pleiotropy that integrates across many phenotypes and multiple generations to improve power. RESULTS: We infer pleiotropy from the fitness cost imposed by frequency changes of pleiotropic loci. Mixing Drosophila simulans populations, which adapted independently to the same new environment using different sets of genes, we show that the adaptive frequency changes have been accompanied by measurable fitness costs. CONCLUSIONS: Unlike previous studies characterizing the molecular basis of pleiotropy, we show that many loci, each of weak effect, contribute to genome-wide pleiotropy. We propose that the costs of pleiotropy are reduced by the modular architecture of gene expression, which facilitates adaptive gene expression changes with low impact on other functions.

Pool-GWAS on reproductive dormancy in Drosophila simulans suggests a polygenic architecture.

The genetic basis of adaptation to different environments has been of long-standing interest to evolutionary biologists. Dormancy is a well-studied adaptation to facilitate overwintering. In Drosophila melanogaster, a moderate number of genes with large effects have been described, which suggests a simple genetic basis of dormancy. On the other hand, genome-wide scans for dormancy suggest a polygenic architecture in insects. In D. melanogaster, the analysis of the genetic architecture of dormancy is complicated by the presence of cosmopolitan inversions. Here, we performed a genome-wide scan to characterize the genetic basis of this ecologically extremely important trait in the sibling species of D. melanogaster, D. simulans that lacks cosmopolitan inversions. We performed Pool-GWAS in a South African D. simulans population for dormancy incidence at 2 temperature regimes (10 and 12°C, LD 10:14). We identified several genes with SNPs that showed a significant association with dormancy (P-value < 1e-13), but the overall modest response suggests that dormancy is a polygenic trait with many loci of small effect. Our results shed light on controversies on reproductive dormancy in Drosophila and have important implications for the characterization of the genetic basis of this trait.

Evolution of phenotypic variance in response to a novel hot environment.

Shifts in trait means are widely considered as evidence for adaptive responses, but the impact on phenotypic variance remains largely unexplored. Classic quantitative genetics provides a theoretical framework to predict how selection on phenotypic mean affects the variance. In addition to this indirect effect, it is also possible that the variance of the trait is the direct target of selection, but experimentally characterized cases are rare. Here, we studied gene expression variance of Drosophila simulans males before and after 100 generations of adaptation to a novel hot laboratory environment. In each of the two independently evolved populations, the variance of 125 and 97 genes was significantly reduced. We propose that the drastic loss in environmental complexity from nature to the laboratory may have triggered selection for reduced variance. Our observation that selection could drive changes in the variance of gene expression could have important implications for studies of adaptation processes in natural and experimental populations.

Long-term gut microbiome dynamics in Drosophila melanogaster reveal environment-specific associations between bacterial taxa at the family level.

The influence of the microbiome on its host is well-documented, but the interplay of its members is not yet well-understood. Even for simple microbiomes, the interaction among members of the microbiome is difficult to study. Longitudinal studies provide a promising approach to studying such interactions through the temporal covariation of different taxonomic units. By contrast to most longitudinal studies, which span only a single host generation, we here present a post hoc analysis of a whole-genome dataset of 81 samples that follows microbiome composition for up to 180 host generations, which cover nearly 10 years. The microbiome diversity remained rather stable in replicated Drosophila melanogaster populations exposed to two different temperature regimes. The composition changed, however, systematically across replicates of the two temperature regimes. Significant associations between families, mostly specific to one temperature regime, indicate functional interdependence of different microbiome components. These associations also involve moderately abundant families, which emphasizes their functional importance, and highlights the importance of looking beyond the common constituents of the Drosophila microbiome.

Highly Parallel Genomic Selection Response in Replicated Drosophila melanogaster Populations with Reduced Genetic Variation.

Many adaptive traits are polygenic and frequently more loci contributing to the phenotype are segregating than needed to express the phenotypic optimum. Experimental evolution with replicated populations adapting to a new controlled environment provides a powerful approach to study polygenic adaptation. Because genetic redundancy often results in nonparallel selection responses among replicates, we propose a modified evolve and resequence (E&R) design that maximizes the similarity among replicates. Rather than starting from many founders, we only use two inbred Drosophila melanogaster strains and expose them to a very extreme, hot temperature environment (29 °C). After 20 generations, we detect many genomic regions with a strong, highly parallel selection response in 10 evolved replicates. The X chromosome has a more pronounced selection response than the autosomes, which may be attributed to dominance effects. Furthermore, we find that the median selection coefficient for all chromosomes is higher in our two-genotype experiment than in classic E&R studies. Because two random genomes harbor sufficient variation for adaptive responses, we propose that this approach is particularly well-suited for the analysis of polygenic adaptation.

The genetic architecture of temperature adaptation is shaped by population ancestry and not by selection regime.

BACKGROUND: Understanding the genetic architecture of temperature adaptation is key for characterizing and predicting the effect of climate change on natural populations. One particularly promising approach is Evolve and Resequence, which combines advantages of experimental evolution such as time series, replicate populations, and controlled environmental conditions, with whole genome sequencing. Recent analysis of replicate populations from two different Drosophila simulans founder populations, which were adapting to the same novel hot environment, uncovered very different architectures-either many selection targets with large heterogeneity among replicates or fewer selection targets with a consistent response among replicates. RESULTS: Here, we expose the founder population from Portugal to a cold temperature regime. Although almost no selection targets are shared between the hot and cold selection regime, the adaptive architecture was similar. We identify a moderate number of targets under strong selection (19 selection targets, mean selection coefficient = 0.072) and parallel responses in the cold evolved replicates. This similarity across different environments indicates that the adaptive architecture depends more on the ancestry of the founder population than the specific selection regime. CONCLUSIONS: These observations will have broad implications for the correct interpretation of the genomic responses to a changing climate in natural populations.

Fine Mapping without Phenotyping: Identification of Selection Targets in Secondary Evolve and Resequence Experiments.

Evolve and Resequence (E&R) studies investigate the genomic selection response of populations in an Experimental Evolution setup. Despite the popularity of E&R, empirical studies in sexually reproducing organisms typically suffer from an excess of candidate loci due to linkage disequilibrium, and single gene or SNP resolution is the exception rather than the rule. Recently, so-called "secondary E&R" has been suggested as promising experimental follow-up procedure to confirm putatively selected regions from a primary E&R study. Secondary E&R provides also the opportunity to increase mapping resolution by allowing for additional recombination events, which separate the selection target from neutral hitchhikers. Here, we use computer simulations to assess the effect of different crossing schemes, population size, experimental duration, and number of replicates on the power and resolution of secondary E&R. We find that the crossing scheme and population size are crucial factors determining power and resolution of secondary E&R: A simple crossing scheme with few founder lines consistently outcompetes crossing schemes where evolved populations from a primary E&R experiment are mixed with a complex ancestral founder population. Regardless of the experimental design tested, a population size of at least 4,800 individuals, which is roughly five times larger than population sizes in typical E&R studies, is required to achieve a power of at least 75%. Our study provides an important step toward improved experimental designs aiming to characterize causative SNPs in Experimental Evolution studies.

Parallel gene expression evolution in natural and laboratory evolved populations.

Ecological adaptation is frequently inferred by the comparison of natural populations from different environments. Nevertheless, inference of the selective forces suffers the challenge that many environmental factors covary. With well-controlled environmental conditions, experimental evolution provides a powerful approach to complement the analysis of natural populations. On the other hand, it is apparent that laboratory conditions differ in many ways from natural environments, which raises the question as to what extent selection responses in experimental evolution studies can inform us about adaptation processes in the wild. In this study, we compared the expression profiles of replicated Drosophila melanogaster populations which have been exposed to two distinct temperature regimes (18/28 and 10/20°C) in the laboratory for more than 80 generations. Using gene-wise differential expression analysis and co-expression network analysis, we identified 541 genes and three coregulated gene modules that evolved in the same direction in both temperature regimes, and most of these changes probably reflect an adaptation to the space constraint or diurnal temperature fluctuation that is common in both selection regimes. In total, 203 genes and seven modules evolved temperature-specific expression changes. Remarkably, we detected a significant overlap of these temperature-adaptive genes/modules from experimental evolution with temperature-adaptive genes inferred from natural Drosophila populations covering two different temperature clines. We conclude that well-designed experimental evolution studies are a powerful tool to dissect evolutionary responses.