Researchers at The Ohio State University are using artificial intelligence to accelerate development of prescription drugs which could save lives, time and money.
Xia Ning is an OSU computer science and engineering professor and lead author of a study that researched how generative AI can speed up the drug development process and produce results quicker than research can do by hand. Ning said her team is developing this technology in a safe way as AI is in the news more.
"I always believe that we really need to use AI responsibly. It's probably even more serious for medicine because in the end it's about human lives," Ning said.
The AI framework is called G2Retro.
The framework uses a process called retrosynthesis, a procedure where a target molecule is transformed into potential reactants and the steps to create the chemical recipe, or the synthesis route, needed for a drug can be identified. There is not always one synthesis route and the AI is used to identify all the possibilities scientists and drug makers can use.
The study was published in the Communications Chemistry journal. Her co-authors included Ziqi Chen, Oluwatosin Ayinde, James Fuchs and Huan Sun.
Ning has been with OSU for nearly six years and has split her academic and professional life studying and researching medicine and artificial intelligence.
Ning's team found the AI could work through 40,000 different chemical reactions to create prescription drugs in minutes and narrow that to hundreds that work best to create a new drug. This work was previously done manually, taking much more time.
Ning said the AI supplies multiple different synthesis routes and options and ranks different options for each molecule for what is best. But she said the AI sometimes creates options that do not exist in nature, so they're able to factor those out to narrow down to options that do exist.
To show how effective this AI could be, the team tested it on existing drugs to see if it could predict, accurately, their chemical sequences and the reactions needed to produce them. The AI accurately predicted the molecular makeup of existing drugs that treat blood and skin diseases, heart failure, COVID-19 and fungal infections.
It also provided alternate chemical reactions that could be used to create the same medicine.
Ning said she has high confidence that this breakthrough in AI technology is the next big development in the field and more researchers should focus on its applications in medicine. "If someone wants to synthesize a molecule, then just dump the molecule into our tool and we can predict like which potential reaction type can be used to synthesize that molecule."
Ning said as the efficacy, ethics and AI are debated more due to the popularity of Chat GPT, she wants people to realize AI can also be helpful in these functions. She said there aren't many risks with using AI in this way because this isn't taking up the entire process of creating medicines.
Prescriptions drugs still go through trials, tests and face approval by the FDA before they are deemed safe and put out to the public.
