Related Experiment Videos
Amino acid substitution matrices from an information theoretic perspective
1National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD 20894.
Journal of Molecular Biology
|June 5, 1991
Summary
Protein sequence alignment uses substitution matrices to score amino acid pairings. Different matrices, like PAM-120 and PAM-200, are better suited for specific tasks such as database searches or homology detection.
Area of Science:
- Bioinformatics
- Molecular Biology
- Computational Biology
Background:
- Protein sequence alignment is crucial for molecular biologists.
- Substitution score matrices are essential for local alignment construction.
- Numerous matrices exist, based on diverse rationales.
Purpose of the Study:
- To analyze the underlying principles of substitution matrices.
- To demonstrate that matrices are log-odds matrices with specific target distributions.
- To provide guidance on selecting appropriate matrices for different protein analysis tasks.
Main Methods:
- Statistical analysis of substitution matrices.
- Application of information theory to quantify scores in bits.
- Comparative evaluation of different PAM matrices (e.g., PAM-120, PAM-200, PAM-250).
Main Results:
- All substitution matrices are implicitly log-odds matrices.
- Matrix scores can be expressed in bits, revealing their information content.
- The choice of matrix impacts the effectiveness of sequence comparison.
Conclusions:
- The PAM-250 matrix, while common, may not be optimal for all applications.
- PAM-120 is recommended for database searches.
- PAM-200 is suggested for comparing proteins with suspected homology.