PhmTox researchers see an “exciting future” with the expanding impact of Artificial Intelligence
Artificial Intelligence (AI) is proving to be a vital tool in Michigan State University’s Department of Pharmacology and Toxicology.
Several department researchers – including Dr. Bin Chen, Dr. Sudin Bhattacharya, Dr. Brian Johnson, and Dr. Rance Nault – are actively using AI to supplement and advance their research.
AI creates systems that can perform tasks that typically require human intelligence. This can involve automating repetitive tasks, analyzing data for insights, and enhancing decision-making processes. ChatGPT is perhaps the most well-known form of AI, designed to have conversations with users, responding to questions, prompts, and instructions in a natural, human-like way.
And while AI has caused concern in some quarters for its ability to create, for example, videos that look all too real but aren’t, and student writing assignments that are tormenting and angering teachers, in the science realm, AI is proving valuable and important.
“There are so many different types of AI,” Dr. Nault said. “I find it hard to put it into a single bucket. It’s really just a way to take in a lot of data and have a computer figure out things that we would normally miss or which we might take a long time to do otherwise. It’s a way to process a lot of information and try to make heads or tails of it.”
Dr. Nault recently published a paper dealing directly with that premise.
With the assistance of AI, Dr. Nault and Keji Yuan, a data scientist working in the Nault lab, are seeking to understand how using a graph neural network can help identify disruptions in metabolism resulting from chemical exposures.
The study’s goal was to apply a network-based graph neural network (GNN) to uncover novel or hidden metabolic changes in response to a toxicant -- in this case, dioxin.
“We applied AI to try and figure out how toxicants change metabolism,” Dr. Nault said. “Metabolism is a big, connected network of reactions, and when we look at the effects of these chemicals, we look at them as individual pathways, which are treated as silos. What’s really nice with AI is it can capture the connections between the pathways in the form of a graph on which AI can be trained on.”
Together with a few investigators, Dr. Chen created the Center for AI-Enabled Drug Discovery in the College of Human Medicine. Along with the goal of aiding in the discovery of potential new drugs, the center brings together the local drug discovery community with experts in AI and big data to help facilitate collaborations and train the next generation of scientists with specialized expertise.
His embrace of AI began several years ago, and he sees it as a valuable research tool, especially in his study of such diseases as Idiopathic Pulmonary Fibrosis (IPF), a chronic lung disease, and several cancers.
“We had a very interesting case study for drug discovery when it came to IPF,” he said. “For IPF, my collaborator Dr. Xiaopeng Li’s lab is capable of culturing tissues from patient transplants. The beauty of this model is it can capture the crucial tissue environment, but it can only accommodate testing a small number of drugs. With our AI platform, we can prioritize the most promising candidates from millions of compounds.”
He believes AI and its nearly limitless potential will be a true game changer.
“I think it will become an essential tool in research,” he said. “For researchers, as more and more models achieve unprecedented precision, we need to be open to the opportunity of AI. The Center for AI-Enabled Drug Discovery participates in a number of research proposals, and adding the component of AI has proven to be highly beneficial. Our center has supported multiple investigators in securing competitive grants and has helped postdoctoral scholars win K99 awards.
“In the future, based on the usage of AI models, it will be more like the electricity revolution. It changed the entire industry. It changed the status quo. Think about how to incorporate AI into research. And as more tools are developed, the next few years are going to be very, very exciting.” Dr. Bhattacharya, whose research lies in the area of computational toxicology in the Institute for Quantitative Health Science & Engineering, is using generative AI to simulate adverse biological responses.
“You learn to predict these responses from existing data,” he said. “A big part of that comes from single-cell data, and we can simulate millions of different cells with AI. We use existing data to predict how ‘artificial cells’ might behave. And there’s tremendous promise in using these tools to understand how genes work together to regulate cell function. It just provides a better understanding, especially when it comes to drug treatments.”
Dr. Bhattacharya also spoke of how AI can help develop an intriguing and possibly groundbreaking development of “virtual twins” in the treatment of disease.
The idea is for AI to represent a human body on a computer to locate disease and create more advanced and in-depth personalized treatment. “It can simulate cells and how they talk to each other,” he said. “You can create this copy on a computer for a particular person with their genes and cells in that model. Then you can simulate that model to see how it responds to various drugs.”
He said that would reduce time-consuming and intrusive testing on individuals in the search for a disease and how to treat it. It would also reduce testing with animal models.
“That would be a pretty powerful development,” he said, adding that AI is an interesting next step in the field of pharmacology and toxicology.
“Both toxicology and pharmacology have an AI thread running through them,” he said. “Toxicology is what might happen to healthy cells and tissue when exposed to undesirable or excessive doses of drugs. In pharmacology, you already have a disease and you’re trying to help the body recover.”
He added that AI can do the same type of testing in simulation that would be cost-prohibitive if researchers had to conduct thousands of experiments. “AI can generate new molecules that can be used as drugs,” Dr. Bhattacharya said. “It can simulate situations that would be cost-prohibitive to test in the lab.”
All three researchers see artificial intelligence as a valuable and necessary tool moving forward and have found it useful in gaining answers quickly, accurately, and affordably.
“I think AI is a phenomenal tool,” Dr. Nault said. “It can help guide us in the right direction for discovery and decision making as the models become better and better.”