Role of Statistical Moments and Various Sequence-based Features in Predicting Protein Functions

Proteomics is concerned with the study and analysis of protein-related data. Protein is the elementary unit of the body which is responsible for various functions such as enzyme controlling and activation, signal transduction throughout organs, cell growth, and different cell-related activities. Protein molecules are made up of amino acids, which are linked with a chain to form a protein molecule. The availability of protein data in the form of protein sequences has made the possibility of applying computational tools and techniques. The current study discussed a novel feature extraction mechanism in detail which is used to obtain useful, yet obscure information from the sequence data. The mechanism involves the use of position-based and composition-based characteristics of amino acids within a protein sequence along with the incorporation of central, raw, and Hahn moments. The output of this method is a single input feature vector (IFV) for each sequence that contains all the relevant statistical information about the sequence. These feature vectors serve as an input to the prediction algorithms which are used in the identification of membrane-protein sites, post-translational modification (PTM) sites, and various other protein-related sites which are responsible for many disorders like cancer, neurological diseases, septic shock, and many related diseases.