Applications of machine learning in phylogenetics.

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Machine learning (ML) is increasingly utilized in phylogenetic inference.
• Supervised ML methods use simulated data for tasks like tree topology inference, branch length estimation, and substitution model selection.
• ML is also applied to downstream phylogenetic analyses, including introgression and diversification.

Purpose of the Study:

• To review the application of supervised machine learning approaches in phylogenetic inference.
• To identify barriers hindering the full potential of ML in phylogenetics.
• To propose future directions for advancing ML in this field.

Main Methods:

• Review of existing literature on supervised machine learning in phylogenetics.
• Analysis of how ML methods are applied to infer tree topologies, branch lengths, and substitution models.
• Discussion of challenges and limitations in current ML applications for phylogenetics.

Main Results:

• Supervised ML has shown promise in various phylogenetic inference tasks.
• Several barriers currently limit the widespread adoption and effectiveness of supervised ML in phylogenetics.
• The review highlights specific areas where ML approaches are being successfully implemented.

Conclusions:

• Despite existing challenges, supervised ML holds significant potential for advancing phylogenetic inference.
• Addressing current barriers through improved network designs and data encodings is crucial.
• Future ML applications are expected to better accommodate complex evolutionary processes that challenge traditional phylogenetic methods.