Written in the online issue of July 12, 2021 Communications on Nature, researchers at the University of California San Diego School of Medicine describe a new approach that uses machine learning to look for disease targets, and then predicts whether a drug is likely to receive FDA approval.
The results of the study could measurably change the way researchers go through big data to find meaningful information with significant benefits for the nation’s patients, pharmaceutical industry, and health systems.
“Academic laboratories and pharmaceutical and biotechnology companies have access to an unlimited amount of ‘big data’ and better tools than ever to analyze this data. However, despite these incredible technological advances, success rates in drugs the findings are lower today than in the 1970s, ”said Pradipta Ghosh, MD, lead author of the study and professor in the departments of medicine and cellular and molecular medicine at UC San Diego School of Medicine.
“This is mainly due to the fact that drugs that work perfectly in preclinical inbreeding models, such as laboratory mice, which are identical or genetic to each other, are not translated into patients in the clinic, where each individual and their disease it is unique. This variability in the clinic is believed to be the Achilles heel for any drug discovery program. “
In the new study, Ghosh and colleagues replaced the first and final step of preclinical drug discovery with two new approaches developed at the University of San Diego Institute of Network Medicine (iNetMed), which brings together various research disciplines to develop new solutions to advance in the life sciences and technology and improve human health.
The researchers used the disease model for inflammatory bowel disease (IBD), which is a recurrent, complex and multifaceted autoimmune disorder characterized by inflammation of the intestinal lining. Because it affects all ages and reduces the quality of life of patients, IBD is a priority area for the disease. drug discovery and it is a difficult condition to treat because no two patients behave similarly.
The first step, called target identification, used an artificial intelligence (AI) methodology developed by The Center for Precision Computational System Network (PreCSN), the computational arm of iNetMed. The AI approach helps to model a disease by mapping successive changes in gene expression at the onset and during disease progression. What differentiates this mapping from other existing models is the use of mathematical precision to recognize and extract all possible fundamental rules from gene expression patterns, many of which are ignored by current methodologies.
The underlying algorithms ensure that the identified gene expression patterns are “invariant” regardless of the different disease cohorts. In other words, PreCSN creates a map that extracts information that is applied to all patients with IBD.
The final step, called target validation in preclinical models, was performed in a first phase 0 clinical trial of its species using a living organoid biobank created from patients with IBD at the Center for Excellence in Research HUMANOID (CoRE), the translation arm of iNetMed.
The “0” phase approach is to test the efficacy of drugs identified using the AI model in organoid models of human disease: human cells grown in a 3D environment that mimic disease outside the body. In this case, a gut affected by IBD.
“The concept of a ‘phase 0 trial’ was developed so that most drugs fail between phases I and III. Before proceeding to patients in the clinic, the ‘phase 0’ test is effective in disease models. where ineffective compounds can be rejected early in the process, saving millions of dollars, ”said Soumita Das, Ph.D., senior co-author of the study, director of the HUMANOID Center and associate professor in the Department of Pathology UC San Diego School of Medicine.
Biopsy tissues for study were taken during colonoscopy procedures with IBD patients. These biopsies were used as a source of stem cells to grow organoids.
“There were two major surprises. First, we saw that, despite being far removed from the immune cells in the intestinal wall and the billions of microbes found in the intestinal lining, these organoids from patients with IBD showed the revealing features of a leaky gut with broken cell edges, ”Das said.
“Second, the drug identified by the AI model not only repaired the broken barriers, but also protected them from the attack of pathogenic bacteria that we added to the intestinal model. These findings imply that the drug could work in both outbreaks. acute as in for maintenance therapy to prevent these outbreaks “.
The researchers found that the computational approach had a surprisingly high level of accuracy in several cohorts of patients with IBD and, along with the “0” phase approach, developed first-class therapy to restore and protect the barrier. intestinal with leaks in IBD.
“Our study shows how the probability of success in phase III clinical trials, for any purpose, can be determined with mathematical accuracy,” said Debashis Sahoo, Ph.D., co-author of the study he leads. PreCSN and is an associate professor in the departments of Pediatrics and Computer Science at UC San Diego School of Medicine and UC San Diego.
“Our approach could provide the predictive power that will help us understand how diseases progress, evaluate the potential benefits of a drug, and make strategies on how to use a combination of therapies when current treatment fails,” Sahoo said.
The authors said the following steps include testing whether the drug that passed the human phase “0” trial on a plate can pass the phase III trials at the clinic; and whether the same methodologies can be used with other diseases, ranging from various types of cancer and Alzheimer’s disease to non-alcoholic fatty liver diseases.
“Our plan has the potential to break the status quo and deliver better drugs for chronic diseases that do not yet have good therapeutic solutions,” Ghosh said.
Debashis Sahoo et al, Discovery guided by artificial intelligence of a barrier protective therapy in inflammatory bowel diseases, Communications on Nature (2021). DOI: 10.1038 / s41467-021-24470-5
University of California – San Diego
Citation: Artificial Intelligence Could Be New Plan for Precision Drug Discovery (2021, July 12) Retrieved July 12, 2021 from https://medicalxpress.com/news/2021-07-artificial-intelligence -blueprint-precision-drug.html
This document is subject to copyright. Apart from any fair treatment for the purposes of private study or research, no part may be reproduced without written permission. Content is provided for informational purposes only.