Machine Learning and AI have found new use in the hunt for disease targets, and for further estimation if a drug is likely to obtain FDA approval, based on a new approach employed by researchers at the School of Medicine, University of California.
Importantly, the finding could measurably change the way researchers filter big data to obtain meaningful information for benefits of patients, pharmaceutical industry, and public health care systems of nations.
In fact, pharmaceutical and biotech companies and academic organizations have access to unlimited amount of big data and improved tools than ever before to examine such data. However, despite such notable advances in technology, unfortunately, today, the success rate in drug discovery is lower than in 1970s.
The reasons are a few. Firstly, drugs that work perfectly well for preclinical inbred models don’t translate for patients in clinics, where individuals and nature of the disease are unique. This variability in real-life patient outcomes is believed to the vulnerability for any drug discovery initiative.
Meanwhile, for the study, the research team replaced the first and last steps of preclinical drug discovery with two new approaches. These two approaches combines several research disciplines to create new solutions for advancement of life sciences and technology, thus, improve human health.
To establish the hypothesis, the research team used disease model for inflammatory bowel disease. Characteristically, inflammatory bowel disease is complex characterized by inflammation of the intestinal lining. Since inflammatory bowel disease impacts all age groups and reduces the quality of life for patients, it is a priority disease area for drug discovery, and is challenging as no two patients behave similar.