Improving Protein Structure Models with Better Alignments
Author Information
Author(s): Sommer Ingolf, Toppo Stefano, Sander Oliver, Lengauer Thomas, Tosatto Silvio CE
Primary Institution: Max-Planck-Institute for Informatics
Hypothesis
Can using a new method to select protein structure models improve their quality?
Conclusion
The study demonstrates that combining systematic alignment variation with modern scoring functions can significantly enhance the quality of protein structure models.
Supporting Evidence
- The method was validated by comparing selected structures to native structures using the TM-score evaluation program.
- An average TM-score improvement of 0.016 was observed, with a maximum improvement of 0.29.
- The study evaluated 1612 target sequences with known structures.
Takeaway
This study shows that by trying different ways to align proteins and using smart scoring methods, we can make better models of protein structures.
Methodology
The study used a new method to generate alternative alignments for protein sequences and selected the best models using a Model Quality Assessment Program (MQAP).
Limitations
The method may struggle to select better models when generating a higher number of low-quality models.
Statistical Information
P-Value
p<0.000000000000001
Statistical Significance
p<0.000000000000001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website