ProLoc-GO: A Method for Predicting Protein Subcellular Localization Using Gene Ontology
Author Information
Author(s): Huang Wen-Lin, Tung Chun-Wei, Ho Shih-Wen, Hwang Shiow-Fen, Ho Shinn-Ying
Primary Institution: Institute of Information Engineering and Computer Science, Feng Chia University, Taichung, Taiwan
Hypothesis
An accurate prediction method for predicting subcellular localization of novel proteins without known accession numbers is worth developing.
Conclusion
ProLoc-GO effectively predicts protein subcellular localization using input sequences, achieving test accuracies of 88.1% and 83.3% for two datasets.
Supporting Evidence
- ProLoc-GO achieved test accuracies of 88.1% and 83.3% for SCL12 and SCL16 datasets, respectively.
- The method outperformed existing SVM-based methods which had accuracies below 35%.
- Using known accession numbers, ProLoc-GO reached accuracies of 90.6% and 85.7% for SCL12 and SCL16 datasets.
Takeaway
This study created a tool that helps figure out where proteins are located in a cell just by looking at their sequences, even if we don't know much about them.
Methodology
The study used a genetic algorithm combined with a support vector machine (SVM) to select informative Gene Ontology terms for predicting protein localization.
Limitations
The method may not perform as well for proteins with very low sequence similarity to known proteins.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website