Improving Protein Function Prediction Using Evolutionary Trace Annotation
Author Information
Author(s): Ward R. Matthew, Erdin Serkan, Tran Tuan A., Kristensen David M., Lisewski Andreas Martin, Lichtarge Olivier
Primary Institution: Baylor College of Medicine
Hypothesis
Can focusing on key functional features of protein structures improve the accuracy of function prediction?
Conclusion
The study developed a new algorithm that increased annotation specificity to 100% for enzymes while maintaining a sensitivity of approximately 60%.
Supporting Evidence
- The new matching algorithm enabled large-scale searches for reciprocal matches.
- Reciprocal ETA achieved better than 92% accuracy in enzyme predictions.
- ETA was applied to 2935 unannotated structural genomics proteins and predicted enzymatic functions in 320 cases.
Takeaway
This study created a new way to predict what proteins do by looking closely at their important parts, and it worked really well for many proteins.
Methodology
The study used a new algorithm to extract local motifs from protein structures based on evolutionary importance and assessed their similarity to other structures.
Potential Biases
There is a risk of misannotation for non-enzymes when applying the method to unannotated proteins.
Limitations
The method may not accurately predict functions for proteins with low sequence identity.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website