Analysis of Differentially Expressed Genes in Alzheimer's, Parkinson's, and Frontotemporal Dementia Using Machine Learning Algorithm

The brain diseases such as Fronto-Temporal Dementia (FTD), Parkinson&s;s (PD), and Alzheimer&s;s (AD) are neurodegenerative disease release the neuroinflammation and protein misfolding. However the diagnosis of such diseases ruins challenging due to complexity of the dat and variability of the data. The present study identifies molecular signatures by applying machine learning (ML) techniques by using curated microarray datasets. The datasets are collected from National Center for Biotechnology Information (NCBI ) Gene Expression Omnibus (GEO).Based on R based analysis on GEO (GEO2R) as well as Limma package, the expressed genes (|logFC| ≥ 0.5; adj.P.Val ≤ 0.05) were obtained. ML model with Support Vector Machine (SVM), was developed by training and validating these genes. SVM showed very high precision as well as recall. The heatmaps are ensured the data quality. The work underlines the capability of statistical biomarkers and collective learning to distinguish between disease-specific gene expression profiles, henceforth enabling the development of gene-based diagnostics. In future, to enhance the interpretability of the model and improvement of the model, genomic structural data can be incorporated. This helps to identify the abnormal genes for the further conclusion.