PhylteR: Efficient Identification of Outlier Sequences in Phylogenomic Datasets

Area of Science:

• Phylogenomics
• Computational Biology
• Evolutionary Biology

Background:

• Gene tree incongruences in phylogenomics complicate species tree inference.
• Manual detection of outliers in large datasets is infeasible.
• Automated methods for identifying outlier sequences are needed.

Purpose of the Study:

• To introduce PhylteR, a novel method for rapid and accurate detection of outlier sequences in phylogenomic datasets.
• To provide a tool that automates the identification of species from individual gene trees that do not follow the general evolutionary trend.

Main Methods:

• PhylteR utilizes DISTATIS, a 3D extension of multidimensional scaling, to compare multiple distance matrices.
• Distance matrices are extracted from individual gene phylogenies, representing evolutionary distances between species for each gene.

Main Results:

• PhylteR demonstrated higher sensitivity and precision in outlier detection compared to existing methods on simulated data.
• The method is robust to incongruences caused by incomplete lineage sorting (ILS).
• Application to a large Carnivora dataset confirmed PhylteR's ability to identify known outliers and improve gene tree-species tree concordance after removal.

Conclusions:

• PhylteR offers an efficient and accurate solution for automated outlier detection in phylogenomics.
• The method aids in dataset characterization and error identification through graphical outputs.
• PhylteR is available as an R package and containerized versions for broad accessibility.