Related Experiment Videos
PSIC: profile extraction from sequence alignments with position-specific counts of independent observations
S R Sunyaev1, F Eisenhaber, I V Rodchenkov
1European Molecular Biology Laboratory, Meyerhofstrasse1, Postfach 10. 2209, D-69012 Heidelberg, Germany.
Protein Engineering
|June 9, 1999
Summary
A new method, position-specific independent counts (PSIC), offers improved sequence weighting for phylogenetic analysis. This technique enhances the identification of distantly related protein sequences and aids in fold family assignment.
Area of Science:
- Bioinformatics
- Computational Biology
- Structural Bioinformatics
Background:
- Sequence weighting in multiple sequence alignments aims to manage redundant data from similar sequences.
- Conventional methods assume uniform phylogenetic change efficiency across all sequence positions, a limitation addressed by novel approaches.
Purpose of the Study:
- To introduce and evaluate the Position-Specific Independent Counts (PSIC) method for sequence weighting.
- To demonstrate PSIC's effectiveness in profile extraction and protein fold family assignment.
Main Methods:
- PSIC calculates position-specific sequence weights using statistical concepts to determine independent observations based on sequence similarity at each alignment position.
- The method enables rapid computation of weights, even for large alignments.
Main Results:
- PSIC was applied to profile extraction and fold family assignment for proteins with known structures.
- The method proved highly effective in discovering distantly related sequences, outperforming Hidden Markov Models and profile methods like WiseTools and PSI-BLAST in several instances.
Conclusions:
- PSIC offers a powerful and efficient approach to sequence weighting in bioinformatics.
- The method enhances the discovery of remote homology and improves protein structure-based classification.