WFABC: a Wright–Fisher ABC‐based approach for inferring effective population sizes and selection coefficients from time‐sampled data

Abstract

With novel developments in sequencing technologies, time-sampled data are becoming more available and accessible. Naturally, there have been efforts in parallel to infer population genetic parameters from these data sets. Here, we compare and analyse four recent approaches based on the Wright-Fisher model for inferring selection coefficients (s) given effective population size (N(e)), with simulated temporal data sets. Furthermore, we demonstrate the advantage of a recently proposed approximate Bayesian computation (ABC)-based method that is able to correctly infer genomewide average N(e) from time-serial data, which is then set as a prior for inferring per-site selection coefficients accurately and precisely. We implement this ABC method in a new software and apply it to a classical time-serial data set of the medionigra genotype in the moth Panaxia dominula. We show that a recessive lethal model is the best explanation for the observed variation in allele frequency by implementing an estimator of the dominance ratio (h).

On the Analysis of Intrahost and Interhost Viral Populations: Human Cytomegalovirus as a Case Study of Pitfalls and Expectations

“…In a complementary approach, we also estimated the fitness effects of all segregating alleles in longitudinally sampled populations using the Wright-Fisher ABC (WFABC) approach (17). Here, N e and the locus-specific s for each segregating site are jointly estimated in an approximate Bayesian framework.…”

“…We utilized an approximate Bayesian framework to infer the viral effective population size and the DFE, using the program WFABC (17,20,21). In brief, allele frequencies were inferred from multiple time point samples, and the variance in allele frequencies was used to estimate viral population size and per-site selection coefficients in a two-step process, assuming independence between sites.…”

Thinking too positive? Revisiting current methods of population genetic selection inference

“…Owing to advances in sequencing technologies, there is now an increased availability of multitime point data of not only an experimental, but also a nonexperimental nature (ranging from ancient samples to those from longitudinal medical studies or field work). This temporal dimension spurred the development of multiple time point based methods over the last few years -all seeking to estimate some combination of selection coefficient and (i) effective population size [32,33], (ii) migration [34], (iii) the age of the selected mutation [31], or (iv) the recombination rate [35]. These methods differ both in the underlying models and in their respective performance, and hence in the conditions to which they are best suited (Table 1; Table S1 in the supplementary material online).…”

“…Subsequently, Malaspinas et al [31] use an approximate transition density to compute the likelihood, whereas Mathieson and McVean [34] use an expectation-maximization algorithm to maximize the likelihood, which speeds up the estimation procedure. By contrast, Foll et al [32,33] do not use a Hidden Markov Model (HMM), but an ABC approach to obtain the posterior distribution of the selection coefficient after simulating allele frequency trajectories using the Wright-Fisher model under a range of selection coefficients. This approach results in lower accuracy for very small selection coefficients, but also represents the fastest and least biased method to date.…”

“…Secondly, such methods have, thus far, been unable to effectively tease apart the effects of direct versus linked selection. By contrast, given the computational efficiency of recently proposed ABC-based approaches [32,33], it has indeed become possible to effectively utilize such approaches in order to identify candidate sites from whole genome polymorphism data, thereby avoiding the first limitation. In addition, because of its efficiency and ability to handle multiple summary statistics, ABC appears as the most promising framework in which to consider more complex selection models under nonequilibrium demographic scenarios.…”