Inferring duplication episodes from unrooted gene trees.

Area of Science:

• Evolutionary molecular biology
• Bioinformatics
• Phylogenetics

Background:

• A fundamental challenge in evolutionary molecular biology is identifying genomic duplication events and their evolutionary history.
• Reconstructing these events involves inferring single gene duplications and clustering them, often by reconciling gene trees with a species tree.

Purpose of the Study:

• To propose novel computational solutions for the genomic duplication problem.
• To develop methods for clustering gene duplications that allow for minimal reconciliation costs.
• To address the genomic duplication problem assuming unrooted gene trees.

Main Methods:

• Developed algorithms based on new theoretical properties of unrooted gene tree reconciliation.
• Designed exact and heuristic algorithms for clustering gene duplications.
• Utilized a clustering method termed 'minimum episodes'.

Main Results:

• Presented the first solutions to the genomic duplication problem allowing minimal single gene duplications.
• The proposed clustering method, minimum episodes, is effective for unrooted gene trees.
• Algorithms were evaluated on empirical datasets.

Conclusions:

• New theoretical insights into unrooted reconciliation for duplication cost were established.
• The developed algorithms successfully confirmed known genomic duplication events.
• The algorithms demonstrate practical applicability in evolutionary studies.