Skip to Main Content
A method of protein secondary structure pattern discovery is presented. The TEIRESIAS algorithm has been improved to discover protein secondary structure patterns. Four protein secondary structure pattern dictionaries have been built for four organisms. The distribution of patterns and common patterns' structure in different dictionaries is different. Different organism's proteins represent different biological language. Based on the organism-specific dictionary, a hidden Markov model is built to predict proteins secondary structure. Dictionary-based prediction has been tested on four organisms and compared with the profile network from HeiDelberg (PHD) method. The experimental results show that our predict method is better than the PHD method for modified segment overlap (SOV) assessment.