Related Experiment Videos
Multiple model approach--dealing with alignment ambiguities in protein modeling
K Pawłowski1, L Jaroszewski, A Bierzyñski
1Institute of Biochemistry & Biophysics, Polish Academy of Sciences, Warszawa, Poland.
Summary
Ambiguous protein sequence alignments hinder homology-based structure prediction. This study uses multiple alignments and a threading algorithm to identify the best alignment and generate accurate protein models, improving structure prediction reliability.
Area of Science:
- Computational biology
- Structural bioinformatics
- Protein structure prediction
Background:
- Homology-based protein structure prediction relies on accurate sequence alignments.
- Distantly homologous proteins present ambiguous alignments, limiting prediction accuracy.
Purpose of the Study:
- To develop a robust method for protein structure prediction despite ambiguous sequence alignments.
- To improve the reliability of homology-based modeling for distantly related proteins.
Main Methods:
- Employing multiple plausible sequence alignments in a protein modeling procedure.
- Generating numerous models for each target protein.
- Evaluating all generated models using a threading algorithm.
Main Results:
- The approach successfully identifies optimal alignments.
- Reasonable protein models are produced, with quality dependent on structural similarity.
- Demonstrated successful structure prediction for the S100A1 protein dimer.
Conclusions:
- Utilizing multiple alignments and threading enhances protein structure prediction accuracy.
- This method overcomes limitations posed by ambiguous alignments in homology modeling.
- The strategy is effective for predicting structures of proteins with unknown structures.