On average, it takes over one decade to develop a new drug. What if machine learning (ML) and statistical computations could speed up the process? This is what Dr. Hanrui Zhang set out to do as her vision for her professional path shaped up during her studies in the Bioinformatics Graduate Program. Dr. Zhang defended her dissertation titled “Predict Drug Response by Machine Learning” on January 25, 2024, under the mentorship of professor Yuanfang Guan. This dissertation explores the application of ML algorithms in surmounting fundamental challenges in drug development, including stabilizing high-throughput screening outcomes and transforming initial discoveries into clinical practices.
There are many applications for this research with a strong potential for tremendous impact. For example, in cancers, the tumorous cells often mutate in response to a drug, requiring to change drugs often, and therefore to constantly develop treatment options. Machine learning can accelerate this process because it can predict the hot spots for mutations and resistance to a drug. With this individualized knowledge, scientists can develop compounds that can target these areas. Being able to safely develop new drugs more quickly would save the lives of millions of cancer patients.
Dr. Guan’s lab develops ML methods for drug discovery, and this is where Dr. Zhang had the opportunity to participate in a drug-discovery study with Merck, a pharmaceutical company. This scientific collaboration is researching ways to use ML to discover drugs that can target DNA Damage Response (DDR) pathways in tumorous cells. She was very interested in glimpsing at industry R&D processes, and in gaining some understanding of how data meta-learning can be used to develop drugs.
Dr. Zhang also contributed to a study of Parkinson’s disease, using a smartphone to detect and record individual movements. The early onset of the abnormal movements that are the hallmark of this disease can be detected by a smartphone. However, the phone can also be tossed around, for example in a purse, recording noise data that could invalidate the results. To address this issue, Drs. Guan and Zhang developed a data augmentation method that adds action noise to cancel the irrelevant information for a valid prediction. This method earned them first place in the Parkinson’s Disease Digital Biomarker DREAM Challenge out of over one hundred competitors. Working on this project was one of the highlights of Dr. Zhang at U-M. She then realized that there is an intuitive aspect to research, based on a deep understanding of the data and the biology behind it.
“You push 'something' a little bit, and it can make a big difference. There is a great feeling of achievement when it works.”
During her undergraduate studies, Zhang was an intern at Memorial Sloan Kettering Cancer Center (MSK), as an exchange student from Sun Yat-sen University in Guangzhou, China. At MSK, she was encouraged to study at the University of Michigan (U-M) because of its Bioinformatics Graduate Program high ranking reputation and excellence in bioinformatics. “I’ve always wanted to do research, and the U-M Master’s program is very research-oriented,” she said. “And it's cold in Michigan, that’s good for research!”
Asked about her passion for bioinformatics, she said: “I’ve always wanted to do creative work to help human beings and I started with computer science in middle school. This is what I like and what I want to do!” When she first arrived at U-M, although she knew what she liked, she had very little idea about what it would become. After over six years of studies at U-M, she now knows what her career is going to look like, and she is very enthusiastic about it.
“I’m at the beginning of my research career and I hope I can do more to treat many diseases.”
In February 2024, Dr. Zhang is starting a position at the Center for Drug Evaluation and Research at FDA, where she did an internship in summer 2023. As a research scientist and ML/AI reviewer in this division, she will explore using ML to improve the development and regulation process of drugs, such as the treatment for rare diseases.
Outside of work, she likes going hiking and exploring the Bay Area where she relocated. And she has very fond memories of U-M where she also met her fiancé, Adam Beneson, a policy specialist at Google, Youtube.
Dr. Hanrui Zhang and Adam Beneson hiking in the San Bruno Mountains in California.