Rec-DCM-Eigen: reconstructing a less parsimonious but more accurate tree in shorter time

Area of Science:

• Computational Biology
• Phylogenetics
• Genomics

Background:

• Maximum parsimony (MP) methods reconstruct species phylogenies using genome data but are computationally expensive for many species.
• Disk-covering methods (DCMs) offer a divide-and-conquer approach by decomposing species into overlapping subgroups.

Purpose of the Study:

• To develop a novel disk-covering method (DCM) based on spectral methods.
• To create the COGNAC software package for phylogenetic tree reconstruction using the new DCM.
• To enhance the accuracy and efficiency of phylogenetic tree inference.

Main Methods:

• Designed a new DCM utilizing spectral methods.
• Developed the COGNAC software package implementing the new DCM.
• Tested the DCM and COGNAC using gene order data and inversion distance.
• Compared COGNAC's performance against FastME and GRAPPA with Rec-I-DCM3.

Main Results:

• The new DCM effectively reduces the phylogenetic tree search space and eliminates erroneous topologies.
• The DCM enables more accurate initial labeling of internal genomes, improving MP method accuracy.
• COGNAC demonstrated superior accuracy compared to FastME.
• COGNAC reconstructed more accurate phylogenetic trees significantly faster than GRAPPA with Rec-I-DCM3.

Conclusions:

• The novel spectral-based DCM and COGNAC software significantly improve phylogenetic tree reconstruction accuracy and efficiency.
• COGNAC offers a powerful alternative for reconstructing phylogenies, especially for large datasets.
• The method addresses computational limitations of traditional MP approaches and enhances accuracy through improved labeling strategies.