ABSTRACT
Schizophrenia is a chronic psychological disorder that is well known for its complexity and difficulty in diagnosis as well as treatment. Various healthcare Informatics data generated through in – vivo, in – vitro and in – silico based approaches for this disease are available through biological databases and scientific search engines. However, it is quite difficult to select the appropriate data from a huge number of available resources. This problem of retrieval and classification of appropriate proteins associated with Schizophrenia from available online molecular data is dealt with in this chapter by using Amino Acid Descriptors and applying a Deep Neural Network-based classifier. This implementation is done in three steps. At the first step, the Protein sequence dataset is prepared by merging molecular level data of Human diseases from eight standard data repositories, viz GWAS, GLAD4U, UniProt, DisGeNET, GenAtlas, PharmGKb, GenCards, and DrugBank. In the second step, Amino Acid Composition-based features and Physicochemical property-based features have been extracted for the conversion of protein sequences into numerical datasheets. Finally, a Deep Neural Network-based classifier is designed to identify the proteins that are associated with Schizophrenia with high, moderate, and low intensity. The effectiveness of the proposed classifier is demonstrated through simulation results.
