Evidence Based Adaptive Neuro-Fuzzy Inference System for ASD Detection

Autism spectrum disorder (ASD) is an intricate neuro disorder. It affects people's brain development and social behavior. Numerous researches have been evolved based on the etiology of the disorder. However, the transcriptomic study of blood-based gene expression can provide complete insight into autism and eventually it can be used as a diagnostic biomarker. Depth knowledge associated with the gene expression can be utilized in the prediction of the genes that causes autism namely the candidate genes. Most of the statistical approaches and regression analysis are employed for autism candidate gene detection, which in turn involves complexity in processing highly dense data. The classification of the candidate gene is not yet performed efficiently. In this paper, a novel framework that has been proposed comprises of the Adaptive neuro-fuzzy system (ANFIS) to classify the mutants that can cause autism. The evidence database incorporates identified autism genes from various databases. The gene expression omnibus database GSE26415 is utilized for implementation of the proposed work.