Supplementary Material for Predicting Protein-Protein Interactions from Protein Domains Using a Set Cover Approach. IEEE/ACM Transactions on Computational Biology and Bioinformatics, Vol 4 pp. 78-87. Jan-March 2007 can be found here. The portal for PPI predictions has been replaced by a Cytoscape plugin that can do protein and domain interaction predictions using the Maximum Specificity Set Cover Approach (MSSC) as well as other algorithms like Maximum Likelihood Estimation (MLE) and the Sum-Product Algorithm (SPA) for PPI/DDI prediction.

The source code for PPI is available for download. It is has been compiled using g++ 3.2.2 and 3.2.3 on Linux and 3.3.3 on Cygwin.

The data which you upload, and which PPI generates, will be available for 12 hours after the time you first schedule the run. If you do not have cookies enabled, or you delete them before fetching your data, the information from your run will become unavailable.

Input data sets:

Yeast (Bader) Training Set

Yeast (Bader) Test Set

Complex Training Set

Complex Test Set

Non complex Training Set

Non Complex Test Set

Yeast Interactions Training Set

26S Proteasome Complex Test Set

RNA Polymerase II Complex Test Set

Yeast (Bader) Protein Prediction Set

Yeast (Bader) Specificity and Sensitivity Data

Complex Prediction Set

Complex Specificity and Sensitivity Data

Non-complex Prediction Set

Non-complex Specificity and Sensitivity Data

26S Proteasome Prediction Set

RNA Polymerase II Prediction Set

You may download any of the various data sets used by the Protein Protein Interaction program. These sets are in plaintext format.

• SwissPFam Protein Structure Database
• SwissPfam domain architecture for Yeast predictions
• DIP_80
• DIP_20
• DIP5_80
• DIP5_20