A Novel Approach to Predict Schizophrenia Disease: Conceptual Framework

Abstract

A multilayer perceptrons (MLP) is proposed to aid in early detection of schizophrenia in this study. Schizophrenia is a disabling illness with unknown pathogenesis. Clinical studies have shown that there is a genetic contribution to the pathogenesis of schizophrenia. The transcription factor 4 (TCF4) gene, vaccinia-related kinase 2 (VRK2) gene as well as other five genes which were newly-identified loci of 1p21.3, 2p32.3, 8pp23.3, 8q21.3 and 10q24.32-p22.1 are among the most prominent susceptibility genes for schizophrenia. It has been hypothesized that TCF4 affects normal brain development and it has been related to different forms of neurodevelopmental disorders. Despite numerous linkage and association analyses (haplotype and meta analyses), the seven gene polymorphisms that contribute to schizophrenia remain to be elucidated. Using a haplotype block-based gene-tagging approach we will conduct an association study of thirty single nucleotide polymorphisms (SNPs) of the above seven gene polymorphisms. Artificial Neural Networks (ANNs) are widely used in prediction application. This research work will demonstrate that the MLP based prediction of schizophrenia which used genetic data (simple combination of thirty genotypic frequencies) of thirty SNPs to improve the detection possibility and accuracy with high consistency. This ANN is expected to exhibit good performance in the prediction of schizophrenia.