Improving Profile Comparison for Protein Similarity Detection
Author Information
Author(s): Sadreyev Ruslan I, Wang Yong, Grishin Nick V
Primary Institution: University of Texas Southwestern Medical Center
Hypothesis
Does considering scores between unrelated proteins improve the statistical significance of profile comparisons in protein sequence analysis?
Conclusion
Using background distributions based on all known non-homologs significantly enhances the quality of remote homology detection.
Supporting Evidence
- Inclusion of background distributions for subjects improves homology detection quality.
- Using all known non-homolog distributions makes a significant contribution to performance.
- The study suggests that query calibration may be unnecessary when subject relationships are known.
Takeaway
This study shows that looking at scores from unrelated proteins helps scientists find similarities between proteins better, which is important for understanding how they work.
Methodology
The study used a novel approach to estimate statistical significance by combining background score distributions for both query and subject proteins.
Limitations
The study may not account for all variables affecting protein similarity detection, and the results are based on specific databases.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website