The goal
The study, published in March, utilized an AI tool developed by UT post-doctoral researcher Danny Diaz to create the active ingredient in the common Alzheimer’s medication galantamine.
Galantamine is typically manufactured using daffodils. However, the extraction process is time-consuming and costly due to unpredictable crop yields and weather, Diaz said.
Just one prescription of galantamine requires hundreds of daffodils and costs around $70-$116, according to health care technology company Oracle Health.
Instead of relying on daffodils, Diaz said the goal of the study is to eventually use bacteria to produce the active ingredient.
"Basically the goal is to turn sugar, which is like food for the bacteria, into a pharmaceutical drug for Alzheimer's," said Diaz, who now leads the Deep Proteins group at the Institute for Foundations of Machine Learning. "Eventually, the hope is to make the drug much more affordable."
How it works
By genetically modifying bacteria, the researchers can essentially "program" the microbes to create the active ingredient in galantamine as a byproduct of its normal metabolic function.
Diaz said the challenge has been creating the right instruction manual for the bacteria, which involves changing their protein structures.
"There are almost infinite combinations [of protein sequences] to try. ... So the AI can narrow that research space down and save time by significantly lowering the risk of not finding something that works," Diaz said.
What's next
So far, Diaz said the team is about one-fourth of the way to completing the bacteria's proper mutation, and he expects the study could wrap up within the next five years.
The bigger picture
With 2024 being named UT's "Year of AI," the university has become involved in dozens of AI research projects and programs in recent years.
In 2023, the school introduced an AI master's program, and in January, it announced the creation of the Center for Generative AI, which will host one of the largest computer systems in academia.
Diaz said he believes AI research will have the largest impact in biotechnology and health care.
"It's really hard for humans to wrap their head around a lot of these problems because they're just so big and expensive to iterate. ... AI is going to be revolutionary in how we manufacture drugs," Diaz said.
