How and how much does RAD-seq bias genetic diversity estimates?

Abstract

BackgroundRAD-seq is a powerful tool, increasingly used in population genomics. However, earlier studies have raised red flags regarding possible biases associated with this technique. In particular, polymorphism on restriction sites results in preferential sampling of closely related haplotypes, so that RAD data tends to underestimate genetic diversity.ResultsHere we (1) clarify the theoretical basis of this bias, highlighting the potential confounding effects of population structure and selection, (2) confront predictions to real data from in silico digestion of full genomes and (3) provide a proof of concept toward an ABC-based correction of the RAD-seq bias. Under a neutral and panmictic model, we confirm the previously established relationship between the true polymorphism and its RAD-based estimation, showing a more pronounced bias when polymorphism is high. Using more elaborate models, we show that selection, resulting in heterogeneous levels of polymorphism along the genome, exacerbates the bias and leads to a more pronounced underestimation. On the contrary, spatial genetic structure tends to reduce the bias. We confront the neutral and panmictic model to “ideal” empirical data (in silico RAD-sequencing) using full genomes from natural populations of the fruit fly Drosophila melanogaster and the fungus Shizophyllum commune, harbouring respectively moderate and high genetic diversity. In D. melanogaster, predictions fit the model, but the small difference between the true and RAD polymorphism makes this comparison insensitive to deviations from the model. In the highly polymorphic fungus, the model captures a large part of the bias but makes inaccurate predictions. Accordingly, ABC corrections based on this model improve the estimations, albeit with some imprecisions.ConclusionThe RAD-seq underestimation of genetic diversity associated with polymorphism in restriction sites becomes more pronounced when polymorphism is high. In practice, this means that in many systems where polymorphism does not exceed 2 %, the bias is of minor importance in the face of other sources of uncertainty, such as heterogeneous bases composition or technical artefacts. The neutral panmictic model provides a practical mean to correct the bias through ABC, albeit with some imprecisions. More elaborate ABC methods might integrate additional parameters, such as population structure and selection, but their opposite effects could hinder accurate corrections.Electronic supplementary materialThe online version of this article (doi:10.1186/s12862-016-0791-0) contains supplementary material, which is available to authorized users.

Expansion history and environmental suitability shape effective population size in a plant invasion

“…Moreover, rare alleles contribute important equilibrium genetic variation (Luikart, Allendorf, Cornuet, & Sherwin, ) and native C. solstitialis populations have been previously shown to harbour more rare alleles than invading populations in North America (Barker et al, ). There is also a tendency for RAD‐seq to underestimate π in more diverse genomes (Arnold, Corbett‐Detig, Hartl, & Bomblies, ; Cariou, Duret, & Charlat, ), although given the loss of rare alleles from invading populations, we might expect this to affect native populations more strongly than invading populations.…”

Genomic signatures of population bottleneck and recovery in Northwest Atlantic pinnipeds

“…These studies, using both simulated and empirical data, have quantified the number of retained homologous loci among species with varying divergence dates. Simulations suggest that hundreds of orthologous, phylogenetically informative loci can be identified in species with divergence times of up to 60 million years (Cariou, Duret, & Charlat, 2016a,2016b; Rubin, Ree, & Moreau, 2012), but empirical studies have successfully applied RADseq to phylogenetic studies of species only from more recent radiations (<100,000 years; Keller et al., 2013; Wagner et al., 2013) and find that the number of shared SNPs drops, sometimes quickly, as divergence dates increase (Lexer et al., 2013; Pante et al., 2015). Among our two studied seal species, harbor and gray, that diverged approximately 5 mya (Arnason et al., 2006), we found over 100,000 homologous RAD loci.…”

“…While both datasets included similar numbers of loci and sequences per individual, many more polymorphic loci and SNPs were identified among gray seal loci than harbor seal loci (Figure 4). We acknowledge that there exist several sources of bias in estimates of genetic diversity from RAD sequencing data (Arnold, Corbett‐Detig, Hartl, & Bomblies, 2013; Cariou, Duret, & Charlat, 2016a,2016b; DaCosta & Sorenson, 2014; Davey et al., 2013; Gautier et al., 2013), but contend that relative comparisons of diversity between species for which data have been collected and processed in a similar manner should be robust (Lozier, 2014). …”

Comparing Pool‐seq, Rapture, and GBS genotyping for inferring weak population structure: The American lobster (Homarus americanus) as a case study

“…Using simulated data, the authors compared the accuracy and robustness of their F ST estimator under several sources of bias that commonly affect sequencing datasets.They demonstrated that their estimator was robust regarding the variance of coverage across loci and observed that sequencing error implicated a negligible bias for Pool-seq F ST estimates. While simulation studies of Pool-seq and GBS data have already been performed to report sources of bias (e.g.,Arnold, Corbett-Detig, Hartl, & Bomblies, 2013;Cariou, Duret, & Charlat, 2016;Gautier et al, 2013;Guo et al, 2013;Hivert et al, 2018), no study pertaining testing the effect of different levels of pooling, different number of individuals and different number of SNPs on the accuracy of F ST estimates for these three libraries protocols. They also noted that the smaller is the pool size, the higher is the effect of experimental bias.…”