Innovative tool harnesses power of AI to discover drug combinations at unprecedented speeds
In the fight against COVID-19, researchers around the globe are scrambling to find the most effective treatment for the rapidly changing, poorly understood virus behind the disease.
Typically, when new, deadly bacterial or viral infections emerge, researchers develop a treatment that combines several different drugs – a labor intensive, time consuming, trial-and-error process to identify drug candidates and appropriate dosages.
MRIGlobal scientist Gene Olinger, Ph.D., is among a multidisciplinary team of scientists who developed a pioneering artificial intelligence (AI) platform to dramatically increase the efficiency of treatment development.
Known as ‘IDentif.AI’ (Identifying Infectious Disease Combination Therapy with Artificial Intelligence), research was led by Professor Dean Ho from the National University of Singapore (NUS).
The team used 12 carefully selected drug candidates for treating infection in lung cells, then used AI to significantly reduce the number of experiments needed to interrogate the range of combinations and optimal dosages of the 12 drugs. Using this process, it took the team three days to come up with multiple drug regimens out of billions of combinations, an unprecedented level of speed and accuracy.
The effort was based on the University’s long standing efforts to use AI to provide personalized cancer treatments to the patients. Building on this work, the new research was step forward to using similar methods to treat infectious diseases.
“It was great to be part of a multidisciplinary team that focused converging approaches and technology with the goal to rapidly develop potential antivirals,” said Olinger. “While a first step, future efforts planned by the team will include additional approaches that will prepare us for pandemics like we are experiencing with COVID-19. There is great hope in the future of science ability to prevent and stop pandemics.”
Results were published in Advanced Therapeutics on 16 April 16, 2020. For the full paper click here (Open Access): https://lnkd.in/ggRATKf