|
The discovery and understanding of functional relationships amongst genes and their gene products are fundamental to
our understanding of human disease. Data regarding protein-protein interactions in humans
are not complete and computational techniques along with data provided from other organisms
can be used to better inform relationships in humans. We demonstrate the utility of using genome scale data from
Drosophila melanogaster to predict protein-protein relationships for human proteins specific to disease. To find the most
likely candidates for protein-protein interaction, the predicted relationships are tested against
human protein interaction and known disease data, then ranked through a Support Vector Machine. An illustrative example
related to hBrm and hSNF5 shows the validity of our approach. This work was accepted to the 2007 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology
(CIBCB). To search the results of our predicted disease-related human protein interactions click here.
|
|
