Prediction of Protein-Protein Interactions Using Support Vector Machines

Authors:
Affiliations:
Venue:
BIBE '04 Proceedings of the 4th IEEE Symposium on Bioinformatics and Bioengineering
Year:
2004

Citing 0
Cited 4

Learning a prediction model for protein-protein recognition

Proceedings of the 2nd International Conference on Interaction Sciences: Information Technology, Culture and Human
Prediction of protein protein interactions from primary sequences

International Journal of Data Mining and Bioinformatics
Using support vector machine for improving protein-protein interaction prediction utilizing domain interactions

Proceedings of the First ACM International Conference on Bioinformatics and Computational Biology
Predicting functional protein-protein interactions based on computational methods

LSMS'07 Proceedings of the 2007 international conference on Life System Modeling and Simulation

Quantified Score

Hi-index	0.00

Visualization

Abstract

Protein-protein interactions play a crucial role in thecellular process. Although recent studies haveelucidated a huge amount of protein-proteininteractions within Saccharomyces cerevisiae, manystill remain to be identified. This paper presents a newinteraction prediction method that associates domainsand other protein features by using Support VectorMachines (SVMs), and it reports the results ofinvestigating the effect of those protein features on theprediction accuracy. Cross-validation tests revealedthat the highest F-measure of 79%, was obtained bycombining the features "domain," "amino acidcomposition," and "subcellular localization." Theseprediction results were more accurate than thepredictions reported previously. Furthermore,predicting the interaction of unknown protein pairsrevealed that high-scoring protein pairs tend to sharesimilar GO annotations in the biological processhierarchy. This method can be applied across species.