Efficient discovery of frequently co-occurring mutations in a sequence database with matrix factorization.

Area of Science:

• Computational Biology
• Virology
• Genomics

Background:

• Viral evolution is driven by the interaction of multiple mutations.
• Identifying co-occurring mutations in large sequence databases is computationally challenging.

Purpose of the Study:

• To develop an efficient computational method for tracking multiple co-occurring mutations.
• To analyze the biological significance of co-mutational positions (CMPs) in viral evolution.

Main Methods:

• Matrix factorization technique to identify subsets of co-mutating positions.
• Validation using a large dataset of SARS-CoV-2 Spike protein sequences.
• Analysis of identified CMPs in relation to viral variants like Delta and Omicron.

Main Results:

• The developed method efficiently identifies co-occurring mutations.
• Demonstrated superior performance compared to brute-force methods.
• Identified key CMPs associated with Delta and Omicron variants, highlighting their role in viral evolution.

Conclusions:

• The method provides valuable insights into viral adaptability by tracking CMPs.
• Understanding CMP dynamics can elucidate mutation persistence and impact across strains.
• Findings may aid in developing improved vaccine design strategies.