Researchers at Worcester Polytechnic Institute (WPI) are combining artificial intelligence with traditional laboratory biology to untangle a fundamental genetic process that could hold keys to better disease detection and treatment. The work, led by computer science professor Dmitry Korkin and biology professor Natalie Farny, focuses on how alternative splicing — a natural mechanism that shuffles genetic instructions — influences the production of critical proteins in the human body.

According to WPI, the research centers on a class of molecules called RNA binding proteins (RBPs). These molecules act like construction barricades, blocking off sections of genetic code while leaving others free to issue instructions. RBPs are both the product of alternative splicing and active participants in the ongoing process, creating a complex feedback loop.

"These proteins are part of remarkable processes in molecular biology, and it is important to understand the processes better because these proteins are implicated in specific diseases," Korkin said in a university announcement. "With greater knowledge of how these proteins form and function, researchers may be able to identify better markers for disease or even identify new targets for medical treatments."

Alternative splicing is a natural process that allows a single gene to produce multiple different proteins. The WPI team describes it as akin to taking a word like "steam" and snipping out letters to create "seam," "stem," or "team." This flexibility is what allows humans, with a relatively small number of genes, to generate the vast complexity required for life. However, when the splicing process goes awry, it can produce nonfunctional or even harmful RBPs. Some RBPs have been linked to serious conditions including cancer and schizophrenia.

The four-year project is supported by a $1,611,153 grant from the National Institutes of Health. Korkin, the principal investigator, will develop machine learning strategies to analyze large datasets containing information about RBPs, focusing on predicting the stability and function of alternatively spliced versions. Farny, the co-principal investigator, will then test those predictions in laboratory experiments. Her results will feed back into Korkin's models, creating what the researchers describe as a "symbiotic cycle" between computational prediction and biological validation.

"The unique aspect of this project is that we have this symbiotic cycle between two labs," Korkin said. "The cycle, from AI predictions to laboratory validation, can be repeated over and over, which will open up opportunities for future discoveries."

For students interested in careers at the intersection of computer science and biology, the project illustrates a growing trend in biomedical research. Both Korkin and Farny are affiliated with WPI's Bioinformatics and Computational Biology program, a field that blends data science with molecular biology. The work highlights how skills in machine learning and data analysis are becoming increasingly valuable in understanding complex diseases.

"Humans are complex organisms, and we generate that complexity with a relatively small number of genes," Farny said. "It's almost mind-boggling how many potential combinations can be created by splicing human genes in different ways. With this project, we're trying to understand that complexity."