Short branch attraction in phylogenomic inference under the multispecies coalescent

Accurate reconstruction of species trees often relies on the quality input gene estimated from molecular sequences. Previous studies suggested that if sequence length is fixed, maximum likelihood may produce biased which subsequently mislead inference trees. Two key questions need to be answered in this context: what are scenarios result consistently trees? and for those scenarios, there any remedies remove or at least reduce misleading effects In article, we establish a theoretical framework address these questions. Considering scenario where true tree 4-taxon star T∗=(S1,S2,S3,S4) with two short branches leading S 1 2 , demonstrate significantly favors wrong bifurcating [(S ), 3 4 ] grouping branches. We name inconsistent behavior branch attraction, occur real-world data involving internal branch. If no mutation occurs along branch, likely short, equivalent thus will suffer same effect attraction. Theoretical simulation results further attraction arbitrary size. Moreover, primarily caused by lack phylogenetic information data, suggesting converting polytomies can artifacts induced