Welcome to CirPred



We have previously developed CPred and (PS)2, a viable CP site predictor and a protein structure prediction system, respectively. Here we integrate the artificial intelligence methods applied in CPred and the modelling algorithm of (PS)2, along with MODELLER and Gromacs, and present the CirPred.

Circular permutation (CP) has been an important technique for protein bioengineering and for researches of protein folding, stability and function. However, not all positions in a protein structure are permissive for creating viable, i.e. foldable and adequately stable, permutants. CirPred predict viable CP sites by integating four machine learning methods, inclusive of artificial neural network (ANN), support vector machine (SVM), random forest and hiearchical feature integration. It features giving a probability estimate for each residue of the query protein. Residues possessing higher estimates are more feasibile for creating viable circular permutants.