Application of Machine Learning to Immune Disease Prediction

J.-H. Lin, Yuh-Jyh Hu*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The intrusion of viruses, germs or parasites can trigger the immune system to protect our body from the harms done by so-called immunogens. However, these protein antigens can sometimes disable our immune activities and cause immune diseases. Common immune diseases include allergies, autoimmune diseases, and infectious diseases. Recently due to environmental changes, the number of cases of immune diseases has been increasing dramatically. They sometimes take months for the patients to fully recover, or even take lives when the situation gets worse. Therefore, an early accurate prediction of immune diseases can provide valuable information for preventive medicine. Previous studies for the most part focused on the diseases caused by allergens, and thus lacked the analysis of other immune diseases such as autoimmune diseases and infectious diseases. To fill the gap, we applied machine learning techniques to construct accurate classification models for three types of immune diseases, allergy, autoimmune disease and infectious disease, caused by different protein antigens. This study consists of three stages: (a) collected and processed antigen data related to immune diseases, including allergy, autoimmune disease, and infectious disease, (b) analyzed the properties of these protein antigens at the sequence level and the structural level to select and develop new features for classification modeling, and (c) demonstrated the application of machine learning to build classification models for immune disease prediction.
Original languageAmerican English
Pages (from-to)38-42
Number of pages5
JournalInternational Journal of Engineering and Innovative Technology
Volume7
Issue number11
DOIs
StatePublished - May 2018

Fingerprint Dive into the research topics of 'Application of Machine Learning to Immune Disease Prediction'. Together they form a unique fingerprint.

Cite this