Phylogenomic analyses of the genus Drosophila reveals genomic signals of climate adaptation

Abstract

Many Drosophila species differ widely in their distributions and climate niches, making them excellent subjects for evolutionary genomic studies. Here, we have developed a database of high-quality assemblies for 46 Drosophila species and one closely related Zaprionus. Fifteen of the genomes were newly sequenced, and 20 were improved with additional sequencing. New or improved annotations were generated for all 47 species, assisted by new transcriptomes for 19. Phylogenomic analyses of these data resolved several previously ambiguous relationships, especially in the melanogaster species group. However, it also revealed significant phylogenetic incongruence among genes, mainly in the form of incomplete lineage sorting in the subgenus Sophophora but also including asymmetric introgression in the subgenus Drosophila. Using the phylogeny as a framework and taking into account these incongruences, we then screened the data for genome-wide signals of adaptation to different climatic niches. First, phylostratigraphy revealed relatively high rates of recent novel gene gain in three temperate pseudoobscura and five desert-adapted cactophilic mulleri subgroup species. Second, we found differing ratios of nonsynonymous to synonymous substitutions in several hundred orthologues between climate generalists and specialists, with trends for significantly higher ratios for those in tropical and lower ratios for those in temperate-continental specialists respectively than those in the climate generalists. Finally, resequencing natural populations of 13 species revealed tropics-restricted species generally had smaller population sizes, lower genome diversity and more deleterious mutations than the more widespread species. We conclude that adaptation to different climates in the genus Drosophila has been associated with large-scale and multifaceted genomic changes.

Keywords: Drosophila; climate adaptation; incomplete lineage sorting; introgression; phylogenomics; phylostratigraphy.

FIGURE 1

Phylogeny and climate niches of the 47 species. (a) Phylogenomic analyses. Species with names in bold are newly sequenced, those in black are improved by additional sequences generated in this study (see Table S4 for details) and those in grey are as previously published. The species for which we added transcriptome data are indicated with stars and those for which we resequenced multiple individuals are indicated with rhombuses. Pie chart areas for each node show the proportions of the three possible topologies for the corresponding branch, with blue denoting the most common topology (i.e., the species tree, as shown) and red and orange the two alternatives (i.e., with each of the two daughter lineages in the species tree as the outgroup instead). The three nodes indicated with red dots are those for which dates had been estimated by Tamura et al. (2004). (b) Climate zones occupied by each species according to the Köppen classification are presented on the right, with grey denoting presence in that environment

FIGURE 2

Causes of gene tree and species tree discordance. (a) Discordance frequency calculated by discovista as a function of branch length in our Drosophila species tree. The grey shaded region indicates the 95% confidence interval from a linear model (“lm” function in R). (b) Species pairwise metrics for the proportions of ILS (above diagonal) and introgression (below diagonal) estimated by quibl. More blue denotes a higher value. (c) Left, ILS is the only factor causing the closer genomic relationship between D. mauritiana and D. sechellia. The grey background indicates the species tree and the red line indicates the ILS genealogy. “T” is the internal branch length under ILS. Middle and right panels show the distribution of the internal branch lengths supporting the closer relationship of D. mauritiana and D. sechellia, which fitted better with the model assuming only ILS (middle, average of Bayesian information criteria (BIC) = –1815.24) than it did with the model assuming a mixture of ILS and introgression (right, average of BIC = –1776.57). The blue line denotes the inferred distribution of introgression, the red line denotes ILS and the black line their combination. The histograms show the distribution of internal branch lengths of individual trees supporting a closer relationship of D. mauritiana and D. sechellia. mau, mauritiana; sim, simulans; sec, sechellia. (d) Left, both ILS and introgression occurred during the evolution of the repleta group. The grey background indicates the species tree, the blue line denotes the genealogy of introgression and the red lines denotes that of by ILS. “T” is the internal branch length for the corresponding genealogies. The middle and right panels show that the distribution of the internal branch lengths supporting the closer relationship of D. mercatorum and the ancestor of the five mulleri subgroup species fitted better with a model assuming a mixture of ILS and introgression (right, average of BIC = –1415.26) than it did with a model assuming ILS only (middle, average of BIC = –1359.66). The histograms show the distribution of internal branch lengths of individual trees supporting a closer relationship of D. mercatorum and the ancestor of the five mulleri subgroup species. hyd, hydei; buz, buzzatii; ald, aldrichi; nav, navojoa; moj, mojavensis; ari, arizonae; mer, mercatorum

FIGURE 3

Temporal dynamics of novel gene gain and loss. (a) Raw numbers of novel genes in D. melanogaster estimated to have arisen in the 34 phylostrata (left) and four eras (top right), and the number of these genes arising in the four eras normalised for the (very different) durations of the eras (lower right). (b) Plots of the absolute numbers of genes estimated to have arisen in each phylostratum for each of the 47 species up to 4290 Ma (left) and zoomed in to 800 Ma (right). (c) Plots of the phylostratum duration‐normalised numbers of genes estimated to have arisen in each phylostratum for each of the 47 species up to 4290 Ma (left) and zoomed in to 800 Ma (right). Each species in (b) and (c) is represented by a different colour (same in each plot). See Figures S6 and S7 for further details. (d) Average lengths of proteins encoded by genes originating in the different evolutionary eras in three species from phylogenetically diverse lineages

FIGURE 4

Distributions of dN/dS values differing between generalists and specialists. The 499 and 457 genes differing in the comparisons of dN/dS values between the generalists and temperate‐continental and tropical specialists respectively are listed in Table S9

FIGURE 5

Population genomic statistics for tropics‐restricted specialists (blue) vs. widespread generalists (red). (a) Top two panels: The average overall nucleotide diversity (Pi) and effective population size (N _e) were generally larger in the generalists than the tropics‐restricted specialists. Bottom two panels: The proportion of deleterious mutations and neutrality index NI_TG were generally smaller in the generalists than the tropics‐restricted specialists. (b) Pi was significantly negatively correlated with niche position. Confidence limits were calculated using the lm function in R. The generalists are shown in red and the specialists in blue in all panels in (a) and (b). (c) Significance of the correlation between effective population size and gene numbers across the 13 species for different phylostrata. Significance was determined using PGLS methods and assuming a Brownian motion model as detailed in the Materials and Methods (see also Figure S16 legend)