Instructions
1. Enter the email address to which notification of prediction completion will be sent.
2. Select from one of two prediction modes:
a. "Protein": Protein-protein binding site prediction.
b. "DNA": DNA binding site prediction.
3. Specify chain identifier (column 22 of ATOM records in the PDB file) of protein.
4. Submit a protein structure for analysis in one of four ways:
a. Enter four letter PDB code (public domain PDB files only).
b. Upload PDB file (public domain or user-defined PDB file).
c. Copy and paste contents of any PDB file into a text box.
d. Batch processing.
Note on NMR structures: PPI-Pred can only process SINGLE models, please edit the PDB file accordingly.
Updates
09/09/2005: Previously, in cases where a query protein has no known homologs, the error message "Pir File Could Not Be Opened" appeared. PPI-Pred can now process such proteins successfully by assigning a conservation score of 0.5 to each residue thus negating the influence of the conservation attributes on the SVM prediction.
12/09/2005: Batch processing is now possible.
Troubleshooting
If any problems are encountered whilst using PPI-Pred, do not hesitate to email James Bradford (J.R.Bradford@leeds.ac.uk) with details on the nature of the problem and the PDB file submitted.
