Investigation of Amyotrophic Lateral Sclerosis Using Bioinformatics Tools

Amyotrophic lateral sclerosis (ALS) is a lethal neurological disease that causes death after a setback of the respiratory system. The biological pathways of ALS are more convoluted and challenging; thus, diagnosis and treatment remain mysterious. One of the valuable techniques in the diagnosis of ALS is gene expression profiling. It is concerned with finding biomarkers involved in ALS occurrence and propagation. Moreover, machine learning techniques (ML) introduce a powerful tool for analysing and classifying high-dimensional genetic datasets. In this paper, (ML) techniques are employed to identify discriminative genes and differentiate between ALS and non-ALS patients. Fisher Score, statistical Student Ttest, Principal Component Analysis (PCA), and Ant Colony Optimization (ACO) are employed to select the best feature sets. Three classification models are introduced to assess the output of biomarker gene sets: k-nearest neighbor (KNN), Ensemble learning (ENS), and support vector machine (SVM) classifiers. The classification accuracy attained 77.54%, 89.86%, and 91.30%.