EXPLORING MACHINE LEARNING APPROACHES IN PREDICTION OF AUTOIMMUNE DISORDERS USING EPIGENETIC MODIFICATIONS

Abstract

Epigenetic Modifications occur due to an interplay between the genetic and environmental factors and are an important cause for the occurrence of autoimmune disorders where there is disruption of immune tolerance and the body starts to attack its own cells. DNA methylation, histone modifications, and ncRNAs are the major modifications observed in autoimmune disorders such as Systemic Lupus Erythematosus, Rheumatoid Arthritis, Multiple Sclerosis etc. Due to the recent advancements in Artificial intelligence and Machine learning, training of epigenetic data can be used to obtain better predictive and analytical responses. In this thesis we assess the application of various Machine learning techniques such as Supervised, Unsupervised and Deep learning in models after preprocessing and evaluation, which are implied to find the predictive capabilities of the epigenetic data in the diagnosis of autoimmune conditions. We used a DNA methylation profile dataset for Systemic Lupus Erythematosus and applied machine learning algorithms such as Random Forest, Support vector machines, Logistic Regression, XGBoost, Naive Bayes and Artificial neural networks and their evaluation metrics were obtained. It was concluded that Support vector machines worked the best for model development and gave an AUROC of 0.97. Gaps in the current knowledge along with future implications are also highlighted.