A model for Protein Secondary Structure Prediction meta - classifiers

We present here a mathematical model for the Protein Secondary Structure Prediction (PSSP) problems and research. It also represents un effort to build a uniform foundations for PSSP research. The model, and hence the paper, is designed to facilitate and speed up understanding of the long standing PSSP research and its problems also for people who want to get involved in it. We present an abstract definition of a protein and its structures and discuss the Protein Data Banks, and other Proteomic Data Bases as well as three generations of PSSP algorithms and servers (all of them web-accessible). We also discuss the development of most important results, problems and methods of data preparation for PSSP classifiers. Finally, we describe a model for a Meta-Classifier utilizing all, or a subset of PSSP servers and discuss its relationship with the first ever developed, Bayes Network Meta-Classifiers of [13] based on 4 to 6 servers.