A unified genealogy of modern and ancient genomes

Abstract

The sequencing of modern and ancient genomes from around the world has revolutionized our understanding of human history and evolution. However, the problem of how best to characterize ancestral relationships from the totality of human genomic variation remains unsolved. Here, we address this challenge with nonparametric methods that enable us to infer a unified genealogy of modern and ancient humans. This compact representation of multiple datasets explores the challenges of missing and erroneous data and uses ancient samples to constrain and date relationships. We demonstrate the power of the method to recover relationships between individuals and populations as well as to identify descendants of ancient samples. Finally, we introduce a simple nonparametric estimator of the geographical location of ancestors that recapitulates key events in human history.

Evaluation of methods for estimating coalescence times using ancestral recombination graphs

“…Relate does not use time discretization, and tsdate uses a discretization scheme where the time points are the quantiles of the lognormal prior distribution on node ages (Wohns et al . 2022).…”

“…For this reason, inferring ARGs and sampling from their posterior distribution seemed like a nearly impossible endeavour some years ago, but important methodological developments now allow us to do so. Today, there are several methods available to estimate the full ARG or approximations of it, including ARGweaver (Rasmussen et al 2014), Relate (Speidel et al 2019) and tsinfer+tsdate (Kelleher et al 2019; Wohns et al 2021).…”

“…The formulation of the coalescent with recombination approximated as a Markovian process generating tree sequences in the SMC (McVean and Cardin 2005) and SMC’ (Marjoram and Wall 2006) and as a copying process of individual sequences by Li and Stephens (2003), paved the way for more scalable ARG inference methods. Notably, ARGweaver (Rasmussen et al 2014) based on the SMC or SMC’ model, and Relate (Speidel et al 2019) and tsinfer+tsdate (Kelleher et al 2019; Wohns et al 2021) based on the model by Li and Stephens (2003).…”

“…Tsdate (Wohns et al 2021) is a method that estimates coalescence times of tree sequences. Here, we used this method to date tree sequences inferred by tsinfer (Kelleher et al 2019).…”

“…We extracted coalescence times for every 4th vs. every 4+1th sample in the case of 80 samples, and 10th vs. every 10+1th sample in the case of n=200.tsinfer and tsdateVCF files generated by msprime were provided as input to the python API using cyvcf2.VCF and converted to tsinfer samples input object using the add_diploid_sites function described in the tsinfer tutorial (https://tsinfer.readthedocs.io/en/latest/tutorial.html# reading-a-vcf). Genealogies were inferred with tsinfer (version 0.2.0(Kelleher et al 2019)) with default settings and dated with tsdate (version 0.1.3(Wohns et al 2021)) using the same parameter values as in the simulations (Ne=10000, mutation_rate=2e-8),…”