Read Me

The web-server iPPBS-Opt used to predict the Protein-Protein Binding Sites(PPBS). In this study,a new model is proposed for PPBS prediction. Based on the physicochemical descriptors, a protein could be converted into several digital signals and then stationary wavelet transform and sliding window method were used to analyze ensemble learning framework was proposed. the KNNC (K-nearest neighbors cleaning) and IHTS (inserting hypothetical training samples) techiques were used to balance the training dataset. The random forests algorithm was adopted to conduct prediction using each descriptor features and the final result was gotten by integrating all the random forests results via voting.

To obtain the predicted result with the anticipated success rate, the entire sequence of the query protein rather than its fragment should be used as an input.