Home

In order to make sense of gene/protein sets researchers compare them to known functional sets contained in biological pathways. The curation effort for pathway databases requires plenty of time and resources, therefore pathways cover only percentages of known reference proteins (SwissProt).

This project is another effort to extend and construct pathways in an automatic manner which speeds up the annotation process. It uses machine learning techniques to classify experimental protein interactions as functional based on multiple attributes of the interactions. The results are compared using different methods of classification.

Without the certainty provided by a manual curation process of the interactions inside the molecular reactions that make up pathways, the resulting extended interaction networks can be taken as candidates for further pathway curation or as tools to perform functional analysis in larger, more comprehensive networks with the hope of finding pointers for the next steps in the analysis process.

This is a follow up study, which explores alternative methods to the ones presented by Wu, G., et al. (2010). "A human functional protein interaction network and its application to cancer data analysis." Genome Biol 11(5): R53..

• Data sources
• Toronto Dataset
• Summary