Washington State University researcher Sascha Duttke has been awarded a prestigious Microsoft Research Fellowship with a stipend to further his work into decoding DNA’s “spatial grammar” — the hidden rules that determine how genes are switched on and off.
Fellowships with Microsoft Research are awarded under numerous AI-related “challenges” aimed at advancing scientific understanding and societal benefit. Duttke’s fellowship is tied to the “Generative models for regulatory genomics” challenge, which focuses on developing new AI systems capable of interpreting the rules embedded in regulatory DNA.
Duttke, an assistant professor in the WSU College of Veterinary Medicine’s School of Molecular Biosciences, hopes the collaboration with Microsoft Research will help in the creation of a new AI model that can decode DNA’s “spatial grammar” to give researchers a better understanding of how genes are controlled and how small inherited or acquired genetic changes drive health and disease.
“This is an incredible opportunity to fuel biological discovery with advanced AI,” Duttke said. “We now understand that the genome functions like a language, and modern machine learning principles can help us learn not only the words, but also their grammar, which could ultimately help scientists better understand gene regulation and improve how we diagnose and treat disease.”
This is an incredible opportunity to fuel biological discovery with advanced AI.
Sascha Duttke, assistant professor
Washington State University
Duttke’s research has shown that the meaning of a genetic sequence depends not only on the presence of certain regulator binding sites, but on their precise arrangement. Small shifts in spacing can cause dramatic changes in gene activity. In collaboration with AI expert Ananth Kalyanaraman in WSU’s School of Electrical Engineering and Computer Science, the team has been dissecting these complex positional rules that govern the numerous permutations of patterns within a DNA sequence.
“With this fellowship and the additional expertise and resources from Microsoft Research, it may become possible to decode the underlying regulatory grammar at scale — the sky is the limit,” Kalyanaraman said.
If successful, the technology could have broad scientific and medical applications.
The fellowship does more than highlight Duttke’s work applying AI to human and animal health. It is also an example of the many WSU researchers at the forefront of applying AI to develop real-world solutions across fields as diverse as energy management, agriculture, K–12 education, engineering, and health systems.
The fellowship will also support Saiman Dahal, a computer science graduate student in the School of Electrical Engineering and Computer Science. Dahal has been working with Kalyanaraman lab member Krishnu Thapa and Duttke lab member Bayley McDonald to collaboratively develop and test machine‑learning approaches to detecting spatial grammar in regulatory DNA.
