Demographic model selection using random forests and the site frequency spectrum

Abstract

Phylogeographic data sets have grown from tens to thousands of loci in recent years, but extant statistical methods do not take full advantage of these large data sets. For example, approximate Bayesian computation (ABC) is a commonly used method for the explicit comparison of alternate demographic histories, but it is limited by the "curse of dimensionality" and issues related to the simulation and summarization of data when applied to next-generation sequencing (NGS) data sets. We implement here several improvements to overcome these difficulties. We use a Random Forest (RF) classifier for model selection to circumvent the curse of dimensionality and apply a binned representation of the multidimensional site frequency spectrum (mSFS) to address issues related to the simulation and summarization of large SNP data sets. We evaluate the performance of these improvements using simulation and find low overall error rates (~7%). We then apply the approach to data from Haplotrema vancouverense, a land snail endemic to the Pacific Northwest of North America. Fifteen demographic models were compared, and our results support a model of recent dispersal from coastal to inland rainforests. Our results demonstrate that binning is an effective strategy for the construction of a mSFS and imply that the statistical power of RF when applied to demographic model selection is at least comparable to traditional ABC algorithms. Importantly, by combining these strategies, large sets of models with differing numbers of populations can be evaluated.

Process‐based species delimitation leads to identification of more biologically relevant species*

“…delimitR modifies and expands an approach introduced by Smith et al. () that uses the SFS and a random forest (RF) classifier to perform demographic model selection. Briefly, under the algorithm described by Smith et al.…”

“…Briefly, under the algorithm described by Smith et al. (), the user defines a set of models and specifies these models by hand in the coalescent simulator fastsimcoal26 (fsc26; Excoffier et al. ).…”

“…Details about generating the model set are given below, along with a brief description of Steps 2 and 3, but see Smith et al. () for additional details.…”

Vanishing refuge? Testing the forest refuge hypothesis in coastal East Africa using genome‐wide sequence data for seven amphibians

“…Moreover, newly published model‐based methods for genome‐wide SNP data (Smith et al. ; Xue & Hickerson, ) as well as those already existing for mtDNA (Huang, Takebayashi, Qi, & Hickerson, ; Oaks, ) will enable further work for assessing concordant patterns of co‐diversification amongst co‐distributed taxa in tropical hotspots. These methods allow empirical investigation of the synchronicity of divergences within and between species and populations, detecting if co‐distributed taxa exhibit shared responses to major historical events.…”

Integrating life history traits into predictive phylogeography

“…The RF classifier they developed had very high accuracy (98.78% overall accuracy) in predicting the presence or absence of cryptic diversity across this disjunction when environmental variables, along with taxonomy as a proxy for life history traits, were used as predictor variables. We include here the same seven taxa and add five other mesic forest disjuncts that have been recently examined: robust lancetooth snails ( Haplotrema vancouverense ) (Smith et al, ), three species/complexes of taildropper slugs ( Prophysaon andersoni , P. dubium , and P. vanattae/P. humile ; Wilke & Duncan, ; Smith et al, ) and red alders ( Alnus rubra ) (Ruffley et al, ).…”