PipeMaster: inferring population divergence and demographic history with approximate Bayesian computation and supervised machine-learning in R

Abstract

Understanding population divergence involves testing diversification scenarios and estimating historical parameters, such as divergence time, population size and migration rate. There is, however, an immense space of possible highly parameterized scenarios that are difsficult or impossible to solve analytically. To overcome this problem researchers have used alternative simulation-based approaches, such as approximate Bayesian computation (ABC) and supervised machine learning (SML), to approximate posterior probabilities of hypotheses. In this study we demonstrate the utility of our newly developed R-package to simulate summary statistics to perform ABC and SML inferences. We compare the power of both ABC and SML methods and the influence of the number of loci in the accuracy of inferences; and we show three empirical examples: (i) the Muller’s termite frog genomic data from Southamerica; (ii) the cottonmouth and (iii) and the copperhead snakes sanger data from Northamerica. We found that SML is more efficient than ABC. It is generally more accurate and needs fewer simulations to perform an inference. We found support for a divergence model without migration, with a recent bottleneck for one of the populations of the southamerican frog. For the cottonmouth we found support for divergence with migration and recent expansion and for the copperhead we found support for a model of divergence with migration and recent bottleneck. Interestingly, by using an SML method it was possible to achieve high accuracy in model selection even when several models were compared in a single inference. We also found a higher accuracy when inferring parameters with SML.

Drivers of unique and asynchronous population dynamics in Malagasy herpetofauna

“…While an exact mutation rate is not available for the majority of Malagasy herpetofauna, this standard deviation encompasses mitochondrial mutation rate estimates for closely related herpetofaunal groups reported in the literature (Daza et al, 2010; Gehring et al, 2013; Kurabayashi et al, 2008; Wollenberg et al, 2011; see Table 1; Table S1 for summary of prior values). We generated 500,000 simulations per model (expansion, constant and bottleneck) for all observed populations, with number of base pairs and number of individuals specific to each observed dataset (see Table S1) under random parameter draws from the aforementioned priors in PipeMaster (Gehara et al, 2017; Gehara et al, 2019). We reduced our observed sequence data into the following summary statistics per population: nucleotide diversity ( π ), number of segregating sites (ss), haplotype diversity ( H ), Tajima's D and the first three bins of the site frequency spectrum (ss1, ss2 and ss3); see Table S5 for full descriptions of summary statistics.…”

Coalescent simulations indicate that the São Francisco River is a biogeographic barrier for six vertebrates in a seasonally dry South American forest

“…We then used ipyrad for de novo assembly (Eaton and Overcast, 2020) and subsequent processing (see Supplementary Table 1 for parameters). Ipyrad outputs a ".alleles" file with phased loci, which we further converted to FASTA format with the function "iPyrad.alleles.loci2fasta" available in the R package PipeMaster (Gehara et al, 2020). 1 Our final dataset consisted of 2,725 unlinked loci (Table 1) from 10 individuals.…”

The demography of extinction in eastern North American birds

“…We explored intraspecific diversity and population structure using principal components analysis (PCA) using the R 3.4.4 [45] package adegenet v. 2.1.2 [46], sNMF [47] and NgsAdmix [48]. Using PipeMaster v. 0.0.9 [49], we estimated the mean and variance of seven population genetic summary statistics. Additional methodological details are available in the electronic supplementary material.…”

“…SML is statistically robust and outperforms traditional methods [50]. To specify model parameters, simulate data, and estimate summary statistics, we used the R package PipeMaster v. 0.0.9 [49], which has been used to test alternative evolutionary histories using both approximate Bayesian computation [51] and SML [52]. For all nine species, we independently tested a constant population size model (CS) plus two single-epoch models: population expansion (EXP) and population bottleneck (BN).…”