Phylogenetic double placement of mixed samples.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Identifying organisms in mixed samples is crucial for biodiversity, food safety, and evolutionary studies.
• Genome skimming generates low-coverage sequencing data, necessitating new methods for analyzing mixed samples without full genome assembly.

Purpose of the Study:

• To address the problem of phylogenetic double-placement: identifying organisms in a mixed sample and their phylogenetic positions.
• To develop a computational tool for analyzing mixed samples, particularly in the context of genome skimming.

Main Methods:

• Developed a model relating distances between mixed and reference samples using Jaccard indices of k-mer sets.
• Formalized phylogenetic double-placement as a non-convex optimization problem solvable numerically.
• Introduced the MIxed Sample Analysis tool (MISA) to implement the developed model.

Main Results:

• The MISA tool successfully identifies constituent organisms in mixed samples.
• The method performs well on simulated and biological datasets, despite underlying assumptions.
• Phylogenetic placement is achieved simultaneously with mixture decomposition.

Conclusions:

• The developed model and MISA tool offer an effective solution for the phylogenetic double-placement problem.
• This approach has broad applications in biodiversity research, food provenance, and evolutionary reconstruction.
• The method demonstrates robust performance in practical scenarios using genome skimming data.