By Kimberley Clementi
An Zheng, a newly minted Ph.D. in UC San Diego’s Department of Computer Science and Engineering (CSE), is among seven university students selected to receive the 2022 UC San Diego Chancellor’s Dissertation Medal. During commencement ceremonies on June 11, he was recognized for outstanding doctoral research.
Zheng’s dissertation, Deep learning predicts the impact of non-coding genetic variants in human traits and diseases, was the Jacobs School of Engineering winner, chosen for the quality, methodology and potential impact of his work. Zheng’s research applies deep learning techniques to understand the biological processes in gene expressions and identify genetic variants that contribute to human diseases.
“In my dissertation, I presented AgentBind, a novel deep learning framework that leverages neural networks to interpret the impact of non-coding variants on genetic diseases,” said Zheng.
Based on three deep learning cases, Zheng’s dissertation demonstrates that AgentBind’s two-step transfer learning scheme enables this framework to accommodate small datasets. It can also quantify the importance of each nucleotide in the input DNA sequences and characterize sequence features predictive of biological functions. Zheng concludes that these capabilities make AgentBind a practical, reliable and versatile tool for non-coding DNA analyses.
The dissertation’s findings point to other potential benefits as well. AgentBind could allow medical experts to identify treatment targets at a lower cost and, more relevant to patient outcomes, could accelerate diagnosis speed. Zheng foresees applications for a host of genetic diseases including various types of cancer and brain-related diseases like Alzheimer’s and bipolar disorder.
“I feel very honored to receive this award. My Ph.D. journey was not easy. Behind each chapter of my dissertation are the efforts of mine and my colleagues’,” said Zheng.
Zheng considers himself lucky to have studied at UC San Diego and in the CSE department. While all the CSE professors were tremendously helpful, Zheng credits his advisor Melissa Gymrek, an an assistant professor in CSE and UC San Diego School of Medicine, and co-advisor Assistant Professor Hao Su for expanding his knowledge and keeping him at the forefront of his field.
“Professor Gymrek not only worked with me to shape my research ideas but also taught me how to conduct rigorous research step by step. Her guidance, trust and openness to new ideas have been indispensable to me during my training and have pushed me to become a better researcher and independent thinker,” said Zheng.
In May, Zheng joined the health AI team at Google Research as a software engineer. This position follows three consecutive internships with Google working in the areas of brain genomics, geo machine intelligence and cloud Borg accelerators.
“In this team, I will be able to continue to create new machine learning tools, together with experts in different domains, and conduct interdisciplinary research into improving healthcare,” said Zheng.