PI2PE consists of five predictors, developed by the Zhou group at FSU. Three of the predictors are based on position-specific sequence profiles, which are generated by running PSI-blast (blast release 2.2.13) on the non-redundant protein sequence database (nr; 3,625,149 entries; May 2006 release). Click a flag below to access the predictor. Please cite:
Tjong, H.; Qin, S.B.; and Zhou, H.-X. (2007) PI2PE: Protein Interface/Interior Prediction Engine. Nucl. Acids Res. 35, W357-W362,
for using the web servers and cite papers listed under these servers for methods.
WESA
is a meta-predictor, based on a Weighted Ensemble of five methods, for Solvent Accessibility of residues, using the protein sequence as input. It has an expected accuracy of 80%. The prediction can be used for structure prediction.
cons-PPISP
is a consensus neural-network Protein-Protein Interaction Site Predictor. The input is the unbound structure of a protein, which is known to bind another protein. The prediction can be used to drive docking of the protein-protein complex or to assist the scoring of docked structures.
meta-PPISP
is built on three individual web servers: cons-PPISP, PINUP, and Promate. Cross validation showed that meta-PPISP outperforms all the three individual servers. At coverages indentical to those of the individual methods, the accuracy of meta-PPISP is higher by 4.8 to 18.2 percentage points.
DISPLAR
is a DNA-Interaction Site Predictor, with data such as sequence profiles of a List of Adjacent Residues as input. The predictions of cons-PPISP and DISPLAR can be used to build structural models for multi-component protein-DNA complexes.
TransComp
implements the Transient-Complex theory for predicting protein-protein and protein-RNA association rate constants. This predictor provides critical missing kinetic information for quantitative modeling of protein interaction networks.