The distribution of fitness effects of nonsynonymous mutations varies phylogenetically across animals Authors: Meixi Lin, Sneha Chakraborty, Carlos Eduardo G. Amorim, Sergio F. Nigenda-Morales, Annabel C. Beichman, Paulina G. Nuñez-Valencia, Jonathan C. Mah, Jacqueline A. Robinson, Christopher C. Kyriazis, Christian D. Huber, Andrew E. Webb, Sarah D. Kocher, Frederick I. Archer, Andrés Moreno-Estrada, Robert K. Wayne, Kirk E. Lohmueller Published: 2026 ABSTRACT The distribution of fitness effects (DFE) describes the selection coefficients of newly arising mutations and fundamentally influences population genetic processes. However, the extent and mechanisms of differences in the DFE for non-synonymous mutations have not been systematically investigated across species with divergent phylogenetic histories and ecologies. Here, we inferred the DFE in natural populations of 11 animal (sub)species, including humans, mice, fin whales, vaquitas, wolves, collared flycatchers, pied flycatchers, halictid bees, Drosophila, and mosquitoes. We found that mammals have a higher proportion of strongly deleterious mutations (defined as s≤−0.01; 22% to 47% in mammals; 0.0% to 5.4% in insects and birds) and a lower proportion of weakly deleterious mutations than insects and birds. Further, the DFE co-varies with phylogeny, such that the mean mutation effects are more similar in closely related species (Pagel’s λ = 0.84, P = 0.01). Next, we investigated whether various summary statistics of the DFE were related to variation in life-history traits across these organisms. We found some support for genome size, body mass, and long-term effective population size being correlated with the DFE. Overall, our findings are consistent with predictions derived independently from the Fisher’s Geometric Model (FGM), which defines organismal complexity as the number of phenotypes under selection. FGM predicts that mutations are more deleterious in complex organisms, while strongly deleterious mutations occur more frequently in smaller populations. Our study demonstrates strong phylogenetic signal in the evolution of a fundamental population genetics parameter, and proposes that, through mechanisms of epistasis, long-term population size and organismal complexity could be underlying variation in the DFE across animals. Introduction Mutations are the raw material on which other evolutionary forces act. The distribution of fitness effects (DFE) represents the relative proportion of new mutations with different selection coefficients (s) affecting the fitness of the individuals carrying them. In addition to its intrinsic value for understanding how mutations affect fitness, the DFE is a fundamental quantity in evolutionary genetics [1]. The DFE is essential for studying the maintenance of genetic variation, identifying causal variants underlying complex human disease, detecting selective sweeps while considering negative selection, and understanding the long-term survival of small populations [1–6]. Consequently, considerable effort has been put into characterizing the biological factors underlying DFE evolution, and estimating the DFE for a given species of interest [1,7]. However, the patterns of and mechanisms behind variation in the DFE have not been systematically surveyed across many species on a phylogenetic timescale [3]. Several theoretical models have proposed that organismal complexity and long-term effective population size (Na) are pivotal biological factors influencing how the DFE might differ across species. While these two factors are common to many studies, predictions for how they affect the DFE vary across models [8–10]. While intuitive, a precise definition of complexity remains elusive. For example, complexity can refer to the number of unique cell types and protein-protein interactions [11,12]. Theoretically, complexity can refer to the number of genetically uncorrelated phenotypic traits under selection, denoting the dimensionality of phenotypic space in Fisher’s Geometric Model (FGM; [13,14]). Nevertheless, how complexity influences the DFE depends on the underlying model and associated assumptions. On the one hand, in complex organisms, highly connected networks confer robustness against the deleterious effects of mutations, therefore, more complex organisms are predicted to have fewer deleterious mutations [9,15]. Alternatively, assuming that the average effect of a mutation on each individual phenotype does not decrease with increasing complexity, the increased dimensionality of the phenotypic space results in larger effect sizes for random mutations, predicting a higher proportion of deleterious mutations in complex organisms [7,10,16]. Similarly, the effect of long-term population size on the DFE is multifaceted. In species with larger effective population sizes, more efficient natural selection favors stable proteins that are then less susceptible to disruption of function by subsequent mutations, thereby diminishing the average deleterious effects of new mutations [8,17]. In species with smaller effective population sizes, the reduced selective strength leads to the accumulation of deleterious mutations (termed “drift load"), moving the population from the fitness optimum. Consequently, a higher proportion of beneficial mutations is expected in species with smaller Na to compensate for this drift load [10]. Taken together, it has been hypothesized that the population-scaled selection coefficient (2Nas) could be conserved across species, as the interplay between selective strength and protein stability may offset variation in long-term population sizes [8]. Despite the growing number of published DFEs in animal species, a lack of standardized DFE comparisons across lineages has hindered testing hypotheses regarding the evolutionary factors underlying potential variation in the DFE across species [3]. In natural populations, the DFE is usually estimated by contrasting allele frequencies, summarized by the site frequency spectrum (SFS), of neutrally evolving and selected variants in genomic data sets [18,19]. While DFE estimates for model organisms like humans [19,20], Drosophila [18,21], mice [7,22], and nematodes [23] have been available, the DFEs for non-model organisms, such as flycatchers [24], great apes [25], Populus [26], Hawaiian monk seals [27], and vaquitas [28], are emerging either species by species or within groups of closely related taxa. These DFE estimates agree that most amino-acid changing mutations are nearly neutral (−10−5 0) and strongly deleterious (s≤−0.01) mutations are less common [1]. Additionally, the DFE appears to vary across divergent species [24,25,27,28]. For example, humans were found to carry a higher proportion of strongly deleterious mutations (s≤−0.01) than Drosophila [7,22]. Interestingly, closely related species demonstrated similar inferred DFE parameters [25,29]. Unfortunately, the scarcity of high-quality genomic data and genetic parameter estimates (e.g., mutation rates) in non-model organisms [30], coupled with methodological differences and computational inefficiencies in DFE inference [20,31,32], have prevented a more comprehensive comparison. The recent increase in available genomic resources for non-model organisms now offers unprecedented opportunities for comparative population genomics [30]. Using such data, here we infer the DFE in 11 animal (sub)species from 10 species (Canis lupus lupus and Canis lupus arctos are subspecies of wolves) with diverse phylogenetic relationships and life history traits using varDFE (this study). We find signals of phylogenetic dependence in the evolution of DFE across species, with mutations being more deleterious in mammals compared with insects and birds. To evaluate the biological processes behind the variation of this fundamental distribution across species, we correlate proxies for organismal complexity and long-term population size with the DFE in different species. Overall, we find that mutations are more deleterious in more complex species and species with smaller long-term population sizes. These findings are consistent with Fisher’s Geometric Model and imply that epistasis could have played a role in DFE evolution. Results Population genetic inference for 11 animal (sub)species We retrieved high-quality population-level polymorphism datasets for 11 animal (sub)species. These data include three insect species (mosquitoes [33], Drosophila [7], and halictid bees [34]), two bird species (pied flycatchers and collared flycatchers [35]), and six mammal (sub)species from five species (arctic wolves [36,37], gray wolves [38], vaquitas [28], fin whales [39], mice [22], and humans [7]; arctic wolves and gray wolves are subspecies of wolves; Table 1, Text A in S1 Appendix). To further understand the mechanisms of DFE evolution, we also compiled several life history traits and genomic attributes for each species from literature, including mutation rate (Text B in S1 Appendix), long-term population size, body mass, generation time, age at maturity, maximum longevity, genome size and recombination rate (sources and citations in S1 Data). The 11 (sub)species we evaluated are distinctive in phylogenetic position (Fig 1A) and the compiled biological features that could affect the DFE. For example, the literature-derived long-term population sizes (NaLIT) varied from 4,485 (vaquitas) to 1.7×106 (Drosophila) individuals. All datasets were filtered with the same stringency to retain only high-confidence genotype calls and at least eight diploid high-quality individuals in coding regions. We tallied the number of variants at different minor allele frequencies and generated folded site frequency spectra (SFS) for synonymous and nonsynonymous/missense variants (SYN-SFS and MIS-SFS, respectively) for each species (Fig A in S1 Appendix, S2 Data). To implement a robust but flexible workflow to compare the DFE across species, we developed varDFE (this study), a Python package based on the ∂a∂i and Fit∂a∂i packages (Fig B in S1 Appendix; [20,40]). 10.1371/journal.pbio.3003976.t001Table 1Summary of DFE inference in 11 (sub)species’ datasets.Common nameScientific nameClassDemographySFS mask NaSYN |s|median α β′ mosquito Anopheles coluzzii Insectathree-epochTRUE2,286,8750.000130.300.0019fruit fly Drosophila melanogaster Insectatwo-epochFALSE2,766,461 4.3×10−5 0.360.00038halictid bee Lasioglossum calceatum Insectatwo-epochTRUE280,4670.000120.300.0017pied flycatcher Ficedula hypoleuca Avestwo-epochFALSE246,9180.000120.290.0018collared flycatcher Ficedula albicollis Avesthree-epochFALSE138,0270.000260.220.0097gray wolf Canis lupus lupus Mammaliathree-epochFALSE45,5170.00170.120.72arctic wolf Canis lupus arctos Mammaliatwo-epochFALSE74,1390.00620.116.30vaquita Phocoena sinus Mammaliatwo-epochFALSE6,2640.000340.130.096fin whale Balaenoptera physalus Mammaliatwo-epochFALSE12,4110.000910.140.19house mouse Mus musculus castaneus Mammaliatwo-epochFALSE207,6820.00110.210.048human Homo sapiens Mammaliatwo-epochFALSE7,0410.00120.190.073Summary of DFE inference in 11 (sub)species’ datasets, assuming a gamma-distributed DFE for new neutral or deleterious mutations (s≤0). “Demography” is the demographic model of choice used for long-term population size inference. “SFS mask” reports if singletons are masked in the site-frequency spectrum during the inference. “NaSYN” is the synonymous-site-based long-term population size in each dataset. “|s|median” is the estimated median mutation effects (s≤0). “α” and “β′” are the inferred shape and scale parameters, respectively, for the gamma DFE. Summary of DFE inference in 11 (sub)species’ datasets. Summary of DFE inference in 11 (sub)species’ datasets, assuming a gamma-distributed DFE for new neutral or deleterious mutations (s≤0). “Demography” is the demographic model of choice used for long-term population size inference. “SFS mask” reports if singletons are masked in the site-frequency spectrum during the inference. “NaSYN” is the synonymous-site-based long-term population size in each dataset. “|s|median” is the estimated median mutation effects (s≤0). “α” and “β′” are the inferred shape and scale parameters, respectively, for the gamma DFE. Mutations are more deleterious in mammals compared to insects and birds. (A) The phylogenetic relationship among the (sub)species included in this study. (B) The mean deleterious mutation effects (|s|mean) for 11 (sub)species, assuming a gamma-distributed DFE of neutral and deleterious mutations. From top to bottom, species are increasingly divergent from humans: mice, fin whales, vaquitas, arctic wolves, gray wolves, collared flycatchers, pied flycatchers, halictid bees, Drosophila and mosquitoes. Numbers show the ranks of |s|mean. (C) The shape parameter (α) of the gamma-distributed DFE for 11 (sub)species. Numbers show the ranks of α. (D) Proportions of mutations in various categories of |s|. From left to right, mutations range from (nearly) neutral (−10−50) to the fitness optimum. The formula for this FGM-derived DFE, which models the full DFE (−0.5≤s≤0.5), is a function of three key parameters in FGM (eq 3): long-term population size (NaFGM), mutation pleiotropy (m), and scale of mutation effects (σ). We inferred the parameters for this FGM-derived DFE for 11 animal (sub)species (eq 3, derived in [10]). The FGM-derived DFE provided an equally good or better fit to the MIS-SFS compared to the gamma DFE (Figs AI and AJ in S1 Appendix, S17 Data). In addition, the inferred long-term population sizes from the FGM-derived DFE (NaFGM) largely agree with the estimates of NaSYN derived from demographic inference using genetic variation data (Radj2 = 0.36, linear regression; P = 0.03, PGLSλ; Fig 4A), providing additional validation of this model. However, NaFGM estimates are often higher than NaSYN, which could be due to the different biological interpretations in these two Na estimates. NaSYN is inferred from demographic modeling using putatively neutral polymorphisms, whereas NaFGM is inferred from deleterious variation at nonsynonymous sites. NaFGM reflects the equilibrium distance from the fitness optimum, and likely captures a much longer time period than the neutral coalescent time that NaSYN represents [45]. The vaquita population’s Na estimate deviated the most between the two methods, likely reflective of its long-term small population size and recent bottleneck [28]. Fitting a Fisher’s Geometric Model (FGM)-derived DFE. (A) The inferred long-term effective population size (NaFGM) from the FGM-derived DFE (y-axis) is correlated with the effective size estimates from our demographic inference (x-axis). Axes are on a log10 scale. (B) The estimated scale of mutation effects (σ) and mutation pleiotropy (m) parameters for 11 (sub)species, grouped by the Class of each species. Note that σ is higher in mammals compared to insects. The y-axis is on a log10 scale. (C) The estimated proportion of beneficial mutations (s > 0) is negatively correlated with estimates of long-term population size from the literature. The x-axis is on a log10 scale. (D) Proportions of mutations in various categories of s estimated from the FGM-derived DFE for each species. From left to right, mutations range from beneficial (s > 0), (nearly) neutral (−10−50; shape) and β (β>0; scale) are parameters to be inferred. For each species, 100 replicates were run from a permuted starting parameter of α = 0.2 and β = 4,000 using the DFE1D_inferenceFIM module in varDFE [7]. The best-fit parameters with the maximum Poisson log-likelihood were chosen, and the corresponding Akaike information criterion (AIC) was calculated. To quantify uncertainty of the parameter estimates, we Poisson resampled the MIS-SFS 200 times, and repeated the DFE inference to obtain the 95% confidence interval [19,20]. Recall that we optimized the gamma distribution parameters for the population-scaled selection coefficient γ. To obtain the distribution of s, we unscaled the gamma distribution by 1/(2NaSYN), therefore, s~Gamma(α,β′), where β′=β/2NaSYN and NaSYN had been inferred from demography estimation. The mean selection coefficient |s|mean for each species was calculated given the gamma distribution’s property that |s|mean=αβ′. Similarly, |2Nas|mean=αβ. The median selection coefficient |s|median was found as the value of s where the cumulative distribution function equals 0.5 using the pgamma function in R (v.4.4.2 [76]). To compute the proportion of mutations with different values of s or 2Nas, we found the cumulative probability in the given range using the pgamma function. To estimate the proportion of effectively neutral mutations, we added a point mass at neutrality (pneu≥0) to the original gamma distribution. The DFE function can be written as γ~Neugamma(α,β,pneu). The distribution of s is s~Neugamma(α,β′,pneu), where β′=β/(2NaSYN). The mean selection coefficient |s|mean for each species is |s|mean=pneu2*2NaSYN*10−5+(1−pneu)αβ′. The median selection coefficient |s|median was found where the cumulative distribution function equals 0.5. To examine the robustness of the DFE inference, we conducted additional analyses assessing the effects of the assumed functional form of the DFE, demographic model misspecification, singleton-masking treatments, mutation rate variation, differences in nonsynonymous to synonymous mutation rate ratio (rL), and GC-biased gene conversion (detailed in Text D in S1 Appendix). To account for uncertainty in the functional form of the DFE, we performed DFE inference assuming three different functional forms (gamma, neugamma, and lognormal) under the best-fit demographic models and singleton masking treatments. For the lognormal distribution, we parameterized the DFE for each species as γ~Lognormal(μs,σ2). The distribution of s was unscaled as s~Lognormal(μs ′,σ2), where μs ′=μs−ln(2NaSYN). The median selective effects |s|median for each species was calculated using the lognormal distribution’s property that |s|median=exp(μs ′). Phylogenetic signal detection To evaluate whether the DFE co-varies with phylogeny, we calculated Pagel’s λ for the log10 transformed mean selection coefficient (log10(|s|mean)), the log10 transformed median selection coefficient (log10(|s|median)), and the shape parameter (α) in the gamma-distributed DFE, life history traits and genomic attributes using the phytools package (v.2.4.4 [77]). The underlying phylogenetic structure and divergence times for 11 (sub)species were compiled from the TimeTree database (http://timetree.org/). To test if the candidate life history traits and genomic attributes (X) are correlated with the observed DFE parameters, and correct for phylogenetic dependency when appropriate, we utilized the PGLSλ method in phytools. The regression was specified as log10(|s|mean)~log10(X), log10(|s|median)~log10(X), or α~log10(X) to normalize the data, except when X was genome size, in which case X was not log-transformed. The λ estimated in PGLSλ does not equate to the Pagel’s λ estimated individually for X or Y, but represents whether the correlation of X and Y can be accounted for by phylogenetic dependence in residual error. Candidate explanatory variables (X) include synonymous-site-based and literature-based long-term population size (NaSYN and NaLIT) in diploids, body mass in grams, generation time in years per generation, age at maturity in days, maximum longevity in years, genome size in Mb, recombination rate in recombination events per bp per generation, and mutation rate in mutations per bp per generation. P-values were adjusted using a Bonferroni correction across the nine explanatory variables considered. To control for outliers or account for the long branches between mammals and insects, we repeated the Pagel’s λ and PGLSλ analysis excluding the arctic wolves, or only within mammals (S15 Data, S16 Data). To formally test for variation in the DFE across species, we used a grid-search approach (DFE1D_gridsearch in varDFE) to compare each species’ DFE estimates to a null model where the DFE was constrained to be the same across species by likelihood-ratio tests (LRT). 2000 grid points were evenly spaced in biologically meaningful ranges for shape (α = 0.1 to 0.5) and scale (β′ = 0.0001 to 0.5) parameters. Using integrate in ∂a∂i we obtained the expected SFS for each α−β′ pair and calculated the Poisson log-likelihood relative to the empirical MIS-SFS in each species. For the null model, we summed log-likelihoods across species to find the MLE. To test whether the shape (α) and scale (β′) are different in any x number of species, the LRT was constructed as Λ=−2*ln(L0/L1)=−2*(LL0−LL1), where LL0 is the log-likelihood for the null model, with two DFE parameters inferred (α1^=α2^=...=αx^,β1′^=β2′^=...=βx′^; params = 2) given the inferred demographic parameters for each species (ΘD,1,ΘD,2,...,ΘD,x), and LL1 is the log-likelihood for the alternate model, where each species is allowed to have its own DFE parameters (α1^,α2^,αx^,...,β1′^,β2′^,βx′^; params = 2x). Across the 11 (sub)species tested, asymptotically, Λ should follow a χ2 distribution with df=2x−2=20. However, as Λ is a composite likelihood, its empirical distribution under the null model of constrained DFE parameters is broader than the χ2 distribution [7]. Although small P-values (e.g., P < 10−16) should not be interpreted literally, simulations showed that Λ did not exceed 34 under the null model in a two-species comparison, suggesting that values of Λ like those seen here would be unlikely under the null hypothesis [7]. We also calculated the pairwise LRT statistics in all possible pairs of species (x = 2). For example, the LRT statistics comparing humans (population 1) and Drosophila (population 2) DFE can be written as follows: Hierarchical clustering for log10(Λ) was performed by the hclust function in R. To test the significance of the variation of γ=2Nas across species, we repeated the LRT by evenly spacing 2000 grid points in shape (α = 0.1 to 0.5) and population-scaled scale (β = 100 to 2.5×104) parameters. We also calculated Pagel’s λ for the log10 transformed population-level mean selection coefficient (log10(|2Nas|mean)). We repeated PGLSλ analysis with log10(|2Nas|mean) as the response variable, following methods described above. Fitting the FGM-derived DFE To investigate the genetic mechanisms underlying DFE variation, we implemented a functional form of the DFE directly derived from the Fisher’s geometric model [10]. The FGM proposes that fitness (w) can be described as a Gaussian function of the distance from the optimum (z), w(z)=exp(−z2), in an n-dimension phenotypic space. Random mutations affect a subset of m phenotypes (m≤n), with fitness effect size r following a zero-mean Gaussian distribution with scale σ. The DFE for a population at mutation-selection-drift equilibrium can be described using the mutation pleiotropy (m), scale of mutation effects (σ) and long-term population size (NaFGM). As this DFE was derived originally by Lourenco and colleagues [10], we refer to it as s~Lourenco(m,σ,NaFGM). Here Γ(.) is the gamma function and K(.) is the modified Bessel function of the second kind [10]. While complex in its expression, this FGM-derived distribution is functionally equivalent to the previously described gamma and lognormal distributions for generating mutational effects. Therefore, we estimated its parameters (m, σ, and NaFGM) following methods described in the DFE inference section, through maximum-likelihood optimization by fitting the observed differences between MIS-SFS and SYN-SFS across minor allele frequency bins. To compute the proportion of mutations with different values of s, we found the cumulative probability for s in given ranges using the scipy.integrate.quad function (v.1.8.0) in Python. The proportion of beneficial mutations was obtained by calculating the cumulative probability for s > 0. We performed two PGLSλ regressions using the methods described above: first, log10(NaFGM) inferred in this FGM-derived DFE and the log10(NaSYN) obtained in demographic inference; second, the proportion of beneficial mutations and log10(NaLIT). To test for phylogenetic signal in the inferred m and σ parameters, we calculated Pagel’s λ (S15 Data). Simulating a human DFE with the Drosophila demography To investigate whether Fit∂a∂i can accurately detect strongly deleterious mutations in species with large population sizes, we performed a simulation study using Drosophila demography and the human DFE. We used the stdpopsim library to obtain coding sequence (CDS) annotations for Drosophila melanogaster chromosomes 2L, 2R, 3L, and 3R from the FlyBase_BDGP6.32.51_CDS [78]. CDS regions were labeled as “exon", and non-coding regions as “bkgd". To simulate realistic genomic regions, we randomly sampled 108 1-Mb segments from these chromosomes with probabilities proportional to chromosome lengths. Annotations for each segment were subset and converted to SLiM’s coordinate format. Next, we implemented a two-epoch Drosophila demographic model in which the ancestral population size was set to 2,766,461, and the population expanded to 7,482,090 over 507,689 generations. We used a uniform recombination rate of 2.03806×10−8 which is an average of recombination rates for Drosophila chromosomes from the stdpopsim catalog. The mutation rate was set to 1.5×10−9 per bp per generation. Deleterious mutations followed a human-like gamma DFE (mean = –0.0276, shape = 0.189) with a nonsynonymous-to-synonymous ratio of 2.85. Background (noncoding) regions were assigned neutral mutations only. All mutations were additive (semi-dominant). From each simulation run on the randomly sampled 1-Mb Drosophila chromosome segments, we output SFS using a sample size of 100 or a sample size of 10. To reduce computational cost, we applied a scaling factor Q = 100, adjusting population size, mutation rate, recombination rate, and selection coefficients accordingly. One out of 108 simulations did not complete due to time constraints on the cluster and was discarded. To match the total length of chromosomes 2L, 2R, 3L, and 3R in Drosophila, we sampled the SFS outputs from 107 successful simulations with replacement, and summed 107 SFS outputs to generate one bootstrapped SFS that mimics a resampled version of the simulated Drosophila genome (~ 107 Mb). To assess inference uncertainty, we repeated the sampling 50 times, producing 50 bootstrapped SFSs for each sample size (2n = 100 or 2n = 10). We then inferred the DFE on the 50 bootstrapped SFSs as done for the empirical data. We assumed either a gamma-distributed or a lognormal-distributed DFE during inference. To evaluate the influence of selection at linked sites on synonymous mutations, we also inferred demographic parameters from the “bkgd" (non-coding) regions to compare with the demographic parameters inferred from synonymous mutations. Supporting information Supporting information Figs A–AM and Text A–F. The data and code needed to generate Figs A–AM can be found in https://zenodo.org/records/21802302. (PDF) Polymorphism dataset and life history traits summary. (XLSX) Synonymous and missense site-frequency spectra built from each subspecies, and from SSWW-only sites in humans and mice. (TXT) Demographic inference summary. (TXT) Demographic inference model selection results. (TXT) Demographic inference comparisons with previous studies. (XLSX) DFE inference summary assuming that the DFE follows a gamma distribution. (TXT) DFE inference summary assuming that the DFE follows a neugamma distribution. (TXT) DFE inference summary assuming that the DFE follows a gammalet distribution. (TXT) DFE inference summary assuming that the DFE follows a lognormal distribution. (TXT) DFE inference summary assuming that the DFE follows a gamma distribution under different demographic model assumptions. (TXT) DFE inference summary assuming that the DFE follows a gamma distribution under different singleton-masking treatments. (TXT) DFE inference summary assuming that the DFE follows a gamma distribution with alternative mutation rates. (TXT) DFE inference summary assuming that the DFE follows a gamma distribution with alternative nonsynonymous-to-synonymous mutation rate ratios (rL). (TXT) DFE inference summary assuming that the DFE follows a gamma distribution in SSWW-only sites in humans and mice. (TXT) Pagel’s λ analysis summary. (TXT) Phylogenetic generalized least squares model summary. (TXT) DFE inference summary assuming that the DFE follows a Fisher’s geometric model-derived distribution (Lourenco DFE). (TXT) DFE inference comparisons with previous studies. (XLSX)