Genetic diversity landscapes in outcrossing and selfingCaenorhabditisnematodes

Abstract

Caenorhabditisnematodes form an excellent model for studying how the mode of reproduction affects genetic diversity, as some species reproduce via outcrossing whereas others can self-fertilize. Currently, chromosome-level patterns of diversity and recombination are only available for self-reproducingCaenorhabditis, making the generality of genomic patterns across the genus unclear given the profound potential influence of reproductive mode. Here we present a whole-genome diversity landscape, coupled with a new genetic map, for the outcrossing nematodeC. remanei. We demonstrate that the genomic distribution of recombination inC. remanei, like the model nematodeC. elegans, shows high recombination rates on chromosome arms and low rates toward the central regions. Patterns of genetic variation across the genome are also similar between these species, but differ dramatically in scale, being 10-fold greater forC. remanei. Historical reconstructions of variation in effective population size over the past million generations echo this difference in polymorphism. Evolutionary simulations demonstrate how selection, recombination, mutation, and selfing shape variation along the genome, and that multiple drivers can produce patterns similar to those observed in natural populations. Convolutional neural networks demonstrate the potential for classifying distinct evolutionary scenarios in simulated populations. Distinguishing these forces confidently with the empirical data, however, will benefit from larger population genomic samples from multiple populations and consideration of an even more extensive training set of simulations.

Biases in ARG-based inference of historical population size in populations experiencing selection

“…Our D. melanogaster results under the population bottleneck scenario (Figure 4D) suggest mis-inference due to BGS effects could partially explain some of their findings, if their methods for masking sites putatively linked to selected loci were too permissive (Pope et al 2023; see Supplementary Note S2 for details). Beyond insects, Relate has recently been used to infer the demographic histories of self-fertilizing (selfing) and outcrossing Caenorhabditis nematodes (Teterina et al 2023), and a grass species (Brachypodium distachyon) with high rates of selfing (Minadakis et al 2023). Selfing increases the effects of background selection (Charlesworth et al 1993;Charlesworth 2003) and selective sweeps (Hedrick 1980;Hartfield and Bataillon 2020), dramatically affects heterozygosity (Charlesworth and Charlesworth 1987) and linkage (Wright et al 2013;Noble et al 2021), and ultimately, changes the genealogical patterns from which Relate derives its demographic estimates (Nordborg and Donnelly 1997;Nordborg 2000;Strütt et al 2023).…”

“…Among recently developed scalable ARG approximation tools, Relate has emerged as the most accessible and widely used for demographic inference as of yet ( e.g. Speidel et al 2019; Almarri et al 2021; Minadakis et al 2023; Moorjani and Hellenthal 2023; Teterina et al 2023; Ishigohoka and Liedvogel 2024). To approximate an ARG, Relate estimates topological features of local genealogies (trees) based on position-specific distance matrices calculated between haplotypes.…”

Hill-Robertson interference may bias the inference of fitness effects of new mutations in highly selfing species

“…elegans is that the true selfing rate in wild C. elegans might be lower than 99%, thus reducing linked effects of selection (Figure 1). Though selfing rate estimates from unique populations range from 80% to ~100% using various methods (Barrière and Félix 2005 elegans estimated a much higher selfing rate of 93%, suggesting that selfing rate can vary across populations (Teterina et al 2023). Importantly, as the samples used in Gilbert et al (2022) were unevenly sampled worldwide (with many samples from Hawaiian populations), variation in selfing rate across populations or through time could additionally influence DFE inference results.…”

“…(3) If possible, it could be extremely helpful to infer the DFE in closely related nonselfing species or in different populations of the same species with known differences in selfing rates. For instance, C. elegans and A. thaliana populations vary in the estimated rates of selfing by ~10%, and the mostly outcrossing A. lyrata has several populations that have evolved partial selfing (Mable et al 2005;Bomblies et al 2010;Teterina et al 2023). Though differences such as population structure and demography need to be carefully accounted for, such comparisons across populations/species could also be used to approximate the extent of HRI in a selfing species.…”

“…Sivasundar and Hey 2005;Teterina et al 2023), in general, the strong LD in C. elegans populations (haplotype blocks of 2.5 Mb on average; Cutter 2006) along with low frequency of males in wild specimens(Barrière and Félix 2005;Richaud et al 2018) suggest the outcrossing rate is very low in most populations. A recent study using whole-genome sequencing data focusing on Hawaiian C.…”