New feature coding methods in disorder region estimate of protein

The proteins which are the main constituent of bios (living creatures), take part in almost every biochemical activities. For instance, changes of biochemical reaction, derived from protein dysfunction, causes a set of serious illnesses. Therefore determining the structure of proteins is of vital importance. Firstly it must be understood the structure of proteins to understand the fuctions of them. To determine these structures, it is used such in vitro (laboratory environment) methods like Nuclear Magnetic Resonance Spectroscopy (NMR), Nuclear Overhauser Effect, X-Ray Crystallography, DNA microarray technology. Application of in silico (computer, calculating environment) methods are more favored over in vitro methods in terms of cost/benefit and time. This is accelerated the studies like determination of the physicochemical relations of amino acids, modeling of proteins's structures, prediction of pathways. In this study it has been developed two pieces new feature codification methods by using different biochemical and physical features, of protein builder amino acids (resudiue) to predict the irregular zones of proteins. First method has been developed on the basis of physicochemical features of proteins and amino acids, position-score matris of proteins (PSM) and SPINE-X protein features. It has been used the theory of ripple in developing of second method. This feature codification methods has been tested on CASP 10 dataset which is issued by machine learning algorithms and Protein Structure Prediction Center. Acquired experimental results have been compared with the methods in the literature.