A Novel Protein Secondary Structure Intelligent Prediction System

Protein secondary structure prediction is one of major challenges in bioinformatics, data ming. In this paper, we propose a novel intelligent prediction system model-Compound Pyramid System Model, which may become the classic model for predicting protein secondary structure. It consists of four components by intelligent interfaces and synthesizing several methods such as SAC (Structural association classifier), AAC (Attribute association classifier) and KDTICM. The model is applied to the domain knowledge, and the effective attributes are chosen by Causal Cellular Automata. Assessments using RS126 and CB513 datasets indicate that the CPSM method can achieve average Q3 accuracy approaching 84.31% and 86.78%. The prediction result of in CASP8 dataset shows that its performance is better than previously reported methods and accessible prediction servers. The result shows that our method has strong universality ability. The fully automated prediction server of CPSM is available at http://kdd.ustb.edu.cn/protein_ Web/. which has a significant international impact.