Classification of structural protein domain based on hidden Markov model

PDZ domains are a standout amongst the most ordinarily discovered protein-protein connection domains in organisms from bacteria to people. These domains are classified into classes I, II and III, contingent upon their binding accomplices and the type of bonds that have been framed. PDZ domains comprise of around 80 to 90 amino acids and distinguished as a locale of taxonomy homology among a varied list of signaling proteins. They have been presented to act as key players that are required in various diseases intervened with the PDZ domain connections; they contain ranging from cystic fibrosis to cancer. In this investigation, we concentrated on the taxonomy of PDZ domains as Class I or II given the primary sequence of the PDZ domains. We utilized the Hidden Markov Model and in view of the domain's essential amino acid successions in helping PDZ domain taxonomy. We assemble our model utilizing a data set which contains 115 interesting human and mice PDZ domains. Three models of emission matrices are investigated utilizing a unigram, bigram, and trigram of amino acid. Our model achieves the high forecast precision of PDZ domain classes with an accuracy of (83.25%).