Evaluation of methods for estimating coalescence times using ancestral recombination graphs

Abstract

The ancestral recombination graph (ARG) is a structure that describes the joint genealogies of sampled DNA sequences along the genome. Recent computational methods have made impressive progress towards scalably estimating whole-genome genealogies. In addition to inferring the ARG, some of these methods can also provide ARGs sampled from a defined posterior distribution. Obtaining good samples of ARGs is crucial for quantifying statistical uncertainty and for estimating population genetic parameters such as effective population size, mutation rate, and allele age. Here, we use standard neutral coalescent simulations to benchmark the estimates of pairwise coalescence times from three popular ARG inference programs: ARGweaver, Relate, and tsinfer+tsdate. We compare 1) the true coalescence times to the inferred times at each locus; 2) the distribution of coalescence times across all loci to the expected exponential distribution; 3) whether the sampled coalescence times have the properties expected of a valid posterior distribution. We find that inferred coalescence times at each locus are most accurate in ARGweaver, and often more accurate in Relate than in tsinfer+tsdate. However, all three methods tend to overestimate small coalescence times and underestimate large ones. Lastly, the posterior distribution of ARGweaver is closer to the expected posterior distribution than Relate’s, but this higher accuracy comes at a substantial trade-off in scalability. The best choice of method will depend on the number and length of input sequences and on the goal of downstream analyses, and we provide guidelines for the best practices.

Simultaneous Inference of Past Demography and Selection from the Ancestral Recombination Graph under the Beta Coalescent

“…We first investigate results for both GNNcoal and SM β C when the ARG is reconstructed with ARGweaver [68], which is currently the best ARG inference approach for sample size smaller than 20 [14]. Demographic inference results by both approaches are displayed in Figure S14, and α inference results in Table S4.…”

“…Throughout the study we have assumed haploid data, but the GNN coal and SM β C approach can easily be extended to diploidy [8]. Our results on inferred ARGs stress the need for efforts to be made on improving ARG inference [14]. Thanks to the SMC we are close to model the ARG allowing to infer demographic history, selection and specific reproductive mechanism.…”

Multiple sources of uncertainty confound inference of historical human generation times

“…Large-scale reconstruction of gene genealogies, dating mutations, and T MRCA inference are all exciting developments, but these methods are quite new and will likely advance considerably in the coming years. Currently, existing methods show different accuracies even under overly-idealized conditions (B randt et al ., 2022). We should therefore be cautious of unexpected and subtle biases that can impact downstream analyses.…”

“…The ages of variants dated by at least two of the methods are only moderately correlated ( and : r 2 ≈ 0.28, and : r 2 ≈ 0.33, and : r 2 ≈ 0.64; Figure 3A-C and see Figure S20 from W ohns et al (2022)). Imperfect correlations should be expected, as genealogical reconstruction methods have been shown to vary in the accuracy of estimated T M RCA (B randt et al ., 2022).…”

“…(2022)). Imperfect correlations should be expected, as genealogical reconstruction methods have been shown to vary in the accuracy of estimated T M RCA (Brandt et al, 2022).…”