COPCOP: A Novel Algorithm and Parallel Optimization Framework for Co-Evolutionary Domain Detection.

Area of Science:

• Molecular Biology
• Bioinformatics
• Evolutionary Biology

Background:

• Co-evolution is fundamental in biological systems, driving the structural and functional maintenance of interacting proteins and their domains.
• Existing methods primarily detect co-variation, overlooking co-conservation signals crucial for understanding molecular interactions.
• Proteins with multiple domains interacting with various partners exhibit complex co-evolutionary patterns.

Purpose of the Study:

• To develop a novel algorithm, COPCOP, for detecting both co-positive selection (co-variation) and co-purifying selection (co-conservation) in interacting protein domains.
• To enhance the accuracy and scope of co-evolutionary analysis in molecular systems.

Main Methods:

• Development of the COPCOP algorithm to identify dual signals of co-variation and co-conservation.
• Implementation of a multi-level parallel acceleration strategy for COPCOP on the Tianhe-2 supercomputer.
• Comparative analysis against existing co-variation detection programs.

Main Results:

• COPCOP effectively detects both co-variation and co-conservation signals in protein domains.
• The algorithm demonstrates superior performance compared to the popular CAPS program.
• The parallel acceleration strategy enables efficient large-scale co-evolutionary domain detection.

Conclusions:

• COPCOP offers a more comprehensive approach to co-evolutionary analysis by integrating co-variation and co-conservation.
• The algorithm advances the understanding of molecular interactions and protein domain evolution.
• The computational framework supports large-scale biological data analysis for evolutionary insights.