Reconstructing protein and gene phylogenies using reconciliation and soft-clustering

Area of Science:

• Bioinformatics
• Computational Biology
• Evolutionary Genetics

Background:

• Eukaryotic genes can produce multiple protein isoforms, but traditional gene tree reconstruction methods use only one protein product, leading to inaccuracies.
• Existing phylogenetic methods overlook the complexity of alternative splicing and protein isoforms, necessitating improved gene tree reconstruction.

Purpose of the Study:

• To develop new computational methods for reconstructing gene and protein trees that explicitly account for alternative protein isoforms.
• To address the limitations of current phylogenetic reconstruction methods by incorporating the full information from alternative protein products.

Main Methods:

• Extended the concept of reconciliation to protein trees, defining MinDRGT and MinDRPGT problems for minimizing double reconciliation costs.
• Proposed a shift from hard-clustering to soft-clustering for protein ortholog groups to better model gene family evolution.
• Developed exact and heuristic algorithmic solutions for the defined reconciliation problems.

Main Results:

• Demonstrated the application of the new methods on protein and gene trees from the Ensembl database.
• Provided algorithmic solutions for gene and protein tree reconstruction considering alternative isoforms.
• Successfully applied novel computational approaches to real biological data, validating their utility.

Conclusions:

• The proposed methods offer a more accurate approach to gene and protein tree reconstruction by handling alternative protein isoforms.
• This work advances phylogenetic analysis by providing tools to correct for inaccuracies caused by ignoring protein diversity within genes.
• The developed algorithms and software (available on GitHub) can enhance evolutionary studies of gene and protein families.