Integration of ultra-low coverage whole-genome sequences for reconstructing the evolutionary history of Galapagos...

Genomic data from contemporary and historical samples often need to be coupled for evolutionary reconstructions of multitaxon complexes. However, the genetic data recovered from historical samples may result only in ultra-low coverage whole-genome sequences (ulcWGS; <0.15× depth), leading to inaccurate evolutionary inferences given a preponderance of missing data. Using the Galapagos giant tortoise radiation as a study system (Chelonoidis spp., composed of 13 extant and four extinct lineages), we assembled a novel methodological pipeline that removes potential noise introduced by the missing data and enhances the evolutionary signal from ulcWGS samples. We leveraged existing tools for phylogenomic placement (EPA-ng), population genomic structure (smartsnp) and admixture (Admixfrog, NGSadmix) to demonstrate that the evolutionary history of samples can be uncovered with sequencing depths as low as 0.008-0.139×. Importantly, these approaches do not use genotype imputation of the ulcWGS samples, which would require extensive reference datasets. Our application to two cases of extinct lineages of Galapagos giant tortoises, with and without references from the same lineage, demonstrates the general value of the approach. We confirm where the extinct lineages from San Cristóbal and Santa Fe islands fit into the Galapagos giant tortoise radiation, and that these lineages were evolutionarily distinct entities.