Predictive updating methods with application to bayesian classification john hilinski dating
Few recent computational methods of predicting the interface residue have been developed by using different features extracted from known protein interaction sites.
Patch analysis  used a six-parameter function with chemical and physical characteristic features vectors of the known patches, such as flatness and hydrophobicity to predict interface patches.
Independent test results suggested that Naive Bayes Classifier-based method with the protein sequence features as input vectors performed well.
Though NBC was a machine learning algorithm that already existed, different kinds of sequence features and input vector forms give rise to better or worse classification performance.
To obtain the training data of protein-protein complexes with two different chains used to develop a Naive Bayes Classifier, we extracted known biological dimeric protein-protein complexes in the PDB.
To obtain a suitable nonredundant protein sequences dataset from PDB, we applied filtration conditions as follows: We used Uni Prot to filtrate heterodimers in reserved protein database.
Here we creatively used a 181-dimension protein sequence feature vector as input to the Naive Bayes Classifier- (NBC-) based method to predict interaction sites in protein-protein complexes interaction.
The prediction of interaction sites in protein interactions is regarded as an amino acid residue binary classification problem by applying NBC with protein sequence features.