Togkousidis, A. | bioRxiv Science

bioRxiv2024

Much Ado About Nothing: Accelerating Maximum Likelihood Phylogenetic Inference via Early Stopping to evade (Over-)optimization

Maximum Likelihood (ML) based phylogenetic inference constitutes a challenging optimization problem. Given a set of aligned input sequences, phylogenetic inference tools strive to determine the tree topology, the branch-lengths, and the evolutionary parameters that maximize the phylogenetic likelihood function. However, there exist compelling reasons to not push optimization to its limits, by means of early, yet adequate stopping criteria. Since input sequences are typically subject to stochastic and systematic noise, one should exhibit caution regarding (over-)optimization and the inherent risk of overfitting the model to noisy input data. To this end, we propose, implement, and evaluate four statistical early stopping criteria in RAxML-NG that evade excessive and compute-intensive (over-)optimization. These generic criteria can seamlessly be integrated into other phylo-genetic inference tools while not decreasing tree accuracy. The first two criteria quantify input data-specific sampling noise to derive a stopping threshold. The third, employs the Kishino-Hasegawa (KH) test to statistically assess the significance of differences between intermediate trees before, and after major optimization steps in RAxML-NG. The optimization terminates early when improvements are insignificant. The fourth method utilizes multiple testing correction in the KH test. We show that all early stopping criteria infer trees that are statistically equivalent compared to inferences without early stopping. In conjunction with a necessary simplification of the standard RAxML-NG tree search heuristic, the average inference times on empirical and simulated datasets are [~]3.5 and [~]1.8 times faster, respectively, than for standard RAxML-NG v.1.2. The four stopping criteria have been implemented in RAxML-NG and are available as open source code under GNU GPL at https://github.com/togkousa/raxml-ng.

bioinformatics↗

bioRxiv2023

Adaptive RAxML-NG: Accelerating Phylogenetic inference under Maximum Likelihood using dataset difficulty

MotivationPhylogenetic inferences under the Maximum-Likelihood (ML) criterion deploy heuristic tree search strategies to explore the vast search space. Depending on the input dataset, searches from different starting trees might all converge to a single tree topology. Often, though, distinct searches infer multiple topologies with large log-likelihood score differences or yield topologically highly distinct, yet almost equally likely, trees. Recently, Haag et al. introduced an approach to quantify, and implemented machine learning methods to predict, the difficulty of an MSA with respect to phylogenetic inference. Easy MSAs exhibit a single likelihood peak on their likelihood surface, associated with a single tree topology to which most, if not all, independent searches rapidly converge. However, as difficulty increases, multiple locally optimal likelihood peaks emerge, yet from highly distinct topologies. ResultsTo this end, we introduce and implement an adaptive tree search heuristic in RAxML-NG, which modifies the thoroughness of the tree search strategy as a function of the predicted difficulty. Our adaptive strategy is based upon three observations. First, on easy datasets, searches converge rapidly and can hence be terminated at an earlier stage. Second, over-analyzing difficult datasets is hopeless and, thus, it suffices to quickly infer only one of the numerous almost equally likely topologies, to reduce overall execution time. Third, more extensive searches are justified and required on datasets with intermediate difficulty. While the likelihood surface exhibits multiple locally optimal peaks in this case, a small proportion of them is significantly better. Our experimental results for the adaptive heuristic on 9, 515 empirical and 5, 000 simulated datasets with varying difficulty exhibit substantial speedups, especially on easy and difficult datasets (53% of total MSAs), where we observe average speedups of more than 10x. Further, approximately 94% of the inferred trees using the adaptive strategy are statistically indistinguishable from the trees inferred under the standard strategy (RAxML-NG). AvailabilityGNU GPL at https://github.com/togkousa/raxml-ng/tree/adaptive. ContactAnastasis.Togkousidis@h-its.org Supplementary MaterialAvailable

bioinformatics↗