Populus not only has significant economic and ecological values, but also serves as a model tree that is widely used in the basic research of tree growth, physiology, and genetics. However, high levels of morphological variation and extensive interspecific hybridization of Populus pose an obstacle for taxonomy, and also to the understanding of phylogenetic interspecific relationships and biogeographical history. In this study, a total of 103 accessions representing almost all the wild species of Populus were collected and whole-genome re-sequenced to examine the phylogenetic relationships and biogeography history. On the basis of 12,916,788 nuclear single nucleotide polymorphisms (SNPs), we reconstructed backbone phylogenies using concatenate and coalescent methods, we highly disentangled the species relationships of Populus, and several problematic taxa were treated as species complexes. Furthermore, the phylogeny of the chloroplast genome showed extensive discordance with the trees from the nuclear genome data, and due to extensive chloroplast capture and hybridization of Populus species, plastomes could not accurately evaluate interspecies relationships. Ancient gene flow between clades and some hybridization events were also identified by ABBA-BABA analysis. The reconstruction of chronogram and ancestral distributions suggested that North America was the original region of this genus, and subsequent long dispersal and migration across land bridges were contributed to the modern range of Populus. The diversification of Populus mainly occurred in East Asia in recent 15 Ma, possibly promoted by the uplift of the Tibetan Plateau. This study provided comprehensive evidence on the phylogeny of Populus and proposed a four-subgeneric classification and a new status, subgenus Abaso. Meanwhile, ancestral distribution reconstruction with nuclear data advanced the understanding of the biogeographic history of Populus.
