Turkish Universities Accelerate Drug Discovery with 'DrugGEN' AI Project
A collaborative project by scientists from Hacettepe, Gazi, and Middle East Technical University (METU) is using artificial intelligence to accelerate new drug development processes. The project, named ‘DrugGEN’, not only aims to shorten the years-long drug development timeline but also to reduce its high costs. The research was published in Nature, one of the world’s most prestigious scientific journals.
The study, which began in 2021 and involved 11 faculty members, initially focused on developing treatments for liver cancer. Prof. Dr. Tunca Doğan of Hacettepe University’s Department of Computer Engineering explained the work.
“Our AI model can rapidly develop a molecule, provided that experts first tell us which protein, naturally found in our body, needs to be interacted with to treat the disease,” said Prof. Dr. Doğan. “We were inspired by large language models like ChatGPT. Just as language models produce a correct and meaningful answer to a question, our model generates molecules for a user-defined protein target.”
Successful Results in Cancer Study
In their research, the team focused on liver cancer, based on expertise from METU’s Cancer System Biology Laboratory. “They told us which protein we should develop a molecule for based on their research data,” Doğan stated. “The AI then designed many molecules.”
After successful computational tests, the team moved to the experimental phase. “Out of the 5 molecules produced by the AI, two managed to bind to the protein,” Doğan reported. These molecules were then tested on laboratory cell lines, where they showed the same effect. Doğan noted that human trials are the next necessary, complex step.
Goal: Reducing 15 Years to 2
Prof. Dr. Doğan highlighted the inefficiencies of current methods, where drug companies test massive libraries of molecules over long periods. “The time it takes for a single drug from its starting point to being available in a pharmacy is, on average, 15 years. Its current cost is around $2 billion.”
“With AI,” he continued, “we aim to reduce the pre-human trial period to 1-2 years and lower the cost, for example, to $1 million. We also want this to be a study that not only a few large pharmaceutical companies in the world can do, but for Turkey to be able to develop its own drugs.”