A classifier for predicting protein secondary structure from amino acid sequences has been proposed and implemented in a previous experiment. NN-GORV-II classifier utilizes the power of Artificial Neural Network and GOR method of protein secondary structure prediction. The Critical Assessment of techniques for Structure Prediction of proteins (CASP) experiments aim at establishing the current state of the art in protein structure prediction. The NN-GORV-II classifier is tested using CASP targets proteins. This test is based on testing a new protein classifier with proteins targets (amino acids) that were never used by the classifier at any prior training or testing stages, hence it's known as blind test. This type of prediction was described as true prediction. The performance of the NN-GORV-II method on the CASP targets: (Q3) is 76.9% with 7.5% standard deviation while the quality of the prediction (SOV3) of the method reached 75.4% with 9.8% standard deviation. The Correlation Coefficients are 0.68, 0.63, and 0.62 for helices, strands, and coils, respectively, indicating strong relationship between predicted and observed secondary structures states.
Bioinformatics, Protein Secondary Structure Prediction, Blind Test, Independent Test CASP
