Treeline Provides a Unified Strategy for Optimising Phylogenetic Trees under Alternative Criteria

The choice of optimality criterion is a key consideration in phylogenetic studies. Recent work challenges the notion that more computationally demanding optimisation objectives result in better phylogenetic trees. This finding underscores the importance of comparing trees across optimisation objectives in addition to different models of evolution and data partitions. It is currently cumbersome to optimise trees for alternative objectives because multiple programmes must be used, each with its own optimisation framework. Here, I introduce Treeline for optimising balanced minimum evolution, maximum likelihood and maximum parsimony trees. Treeline explores the optimisation landscape using a new strategy based on perturbing the patristic distance matrix used to initialize candidate trees, which is shown to be particularly effective for the balanced minimum evolution objective. Tests suggest Treeline can be more accurate, memory efficient or faster than existing programmes designed for a single optimisation objective. Treeline's unified nature facilitates comparison of phylogenetic trees across distinct optimisation objectives and models of sequence evolution. Consistent with prior studies, the balanced minimum evolution objective resulted in gene trees that were more consistent with species trees than maximum likelihood or maximum parsimony objectives. With a case study of species from the family Hominidae, including Homo sapiens, I show how Treeline simplifies the process of constructing trees under alternative optimisation objectives and models of evolution. Treeline is part of the DECIPHER package for R and is available from Bioconductor and online (https://DECIPHER.codes/).