Espalier: Efficient Tree Reconciliation and Ancestral Recombination Graphs Reconstruction Using Maximum Agreement Forests.
Syst Biol
; 72(5): 1154-1170, 2023 11 01.
Article
em En
| MEDLINE
| ID: mdl-37458753
ABSTRACT
In the presence of recombination individuals may inherit different regions of their genome from different ancestors, resulting in a mosaic of phylogenetic histories across their genome. Ancestral recombination graphs (ARGs) can capture how phylogenetic relationships vary across the genome due to recombination, but reconstructing ARGs from genomic sequence data is notoriously difficult. Here, we present a method for reconciling discordant phylogenetic trees and reconstructing ARGs using maximum agreement forests (MAFs). Given two discordant trees, a MAF identifies the smallest possible set of topologically concordant subtrees present in both trees. We show how discordant trees can be reconciled through their MAF in a way that retains discordances strongly supported by sequence data while eliminating conflicts likely attributable to phylogenetic noise. We further show how MAFs and our reconciliation approach can be combined to select a path of local trees across the genome that maximizes the likelihood of the genomic sequence data, minimizes discordance between neighboring local trees, and identifies the recombination events necessary to explain remaining discordances to obtain a fully connected ARG. While heuristic, our ARG reconstruction approach is often as accurate as more exact methods while being much more computationally efficient. Moreover, important demographic parameters such as recombination rates can be accurately estimated from reconstructed ARGs. Finally, we apply our approach to plant infecting RNA viruses in the genus Potyvirus to demonstrate how true recombination events can be disentangled from phylogenetic noise using our ARG reconstruction methods.
Palavras-chave
Texto completo:
1
Base de dados:
MEDLINE
Assunto principal:
Recombinação Genética
/
Genoma
Tipo de estudo:
Prognostic_studies
Limite:
Humans
Idioma:
En
Ano de publicação:
2023
Tipo de documento:
Article