Fast computation of all maximum acyclic agreement forests for two rooted binary phylogenetic trees

TL;DR

This paper introduces two improved algorithms for efficiently computing all maximum acyclic agreement forests between two rooted binary phylogenetic trees, significantly enhancing speed and scalability for biological research.

Contribution

The paper presents two modifications to the allMAAFs algorithm, making it on average 8 times faster for computing agreement forests in phylogenetic trees.

Findings

Algorithm speed increased by a factor of 8 on average.

Enhanced algorithm enables analysis of larger trees.

Facilitates broader biological applications.

Abstract

Evolutionary scenarios displaying reticulation events are often represented by rooted phylogenetic networks. Due to biological reasons, those events occur very rarely, and, thus, networks containing a minimum number of such events, so-called minimum hybridization networks, are of particular interest for research. Moreover, to study reticulate evolution, biologist need not only a subset but all of those networks. To achieve this goal, the less complex concept of rooted phylogenetic trees can be used as building block. Here, as a first important step, the trees are disjoint into common parts, so-called maximum acyclic agreement forests, which can then be turned into minimum hybridization networks by applying further network building algorithms. In this paper, we present two modifications of the first non-naive algorithm --- called allMAAFs --- computing all maximum acyclic agreement forests for two rooted binary phylogenetic trees on the same set of taxa. By a simulation study, we indicate that through these modifications the algorithm is on average 8 times faster than the original algorithm making this algorithm accessible to larger input trees and, thus, to a wider range of biological problems.