Applications of Machine Learning in Phylogenetics

Abstract

Machine learning has increasingly been applied to a wide range of questions in phylogeneticinference. Supervised machine learning approaches that rely on simulated training data have beenused to infer tree topologies and branch lengths, to select substitution models, and to performdownstream inferences of introgression and diversification. Here, we review how researchers haveused several promising machine learning approaches to make phylogenetic inferences. Despitethe promise of these methods, several barriers prevent supervised machine learning from reachingits full potential in phylogenetics. We discuss these barriers and potential paths forward. In thefuture, we expect that the application of careful network designs and data encodings will allowsupervised machine learning to accommodate the complex processes that continue to confoundtraditional phylogenetic methods.

Microbial Diversity and Open Questions about the Deep Tree of Life

“…AI is also adept at detecting outliers or anomalies in data. This can help identify genes or sequences that do not follow the expected evolutionary patterns, possibly indicating events like horizontal gene transfer, gene duplication, or convergent evolution ( Mo et al 2023 ; van Hooff and Eme 2023 ). It also opens up the possibility of using protein structures instead of sequences to infer difficult relationships ( Moi et al 2023 ).…”

Challenges and opportunities in applying AI to evolutionary morphology

“…In recent decades, genetic data has dominated phylogenetic studies of extant species (although these data are not available for the vast majority of extinct taxa). It is thus unsurprising that much of the early work on ML in phylogenetics focused on molecular rather than morphological data, and recent reviews of ML approaches for tree building (Mo et al, 2023;Sapoval et al, 2022) have correspondingly focused on molecular phylogenetics instead of morphology-based phylogenetics. This mirrors traditional phylogenetics, where molecular methods, due to sheer quantity of data that can be extracted with relative ease, are often prioritised over morphological data.…”

PhyloJunction: a computational framework for simulating, developing, and teaching evolutionary models

“…Alternatively, it should be straightforward to integrate PhyloJunction's functionalities for summarizing data together with Python machine learning libraries. There is increasing evidence [49] backing machinelearning methods as viable alternatives to frequentist and Bayesian evolutionary inference, especially when the latter is very onerous or impossible [75].…”