To determine which genes are responsible for—or act as biomarkers of—a given disease, researchers must work their way through thousands upon thousands of genes, using previous scientific studies as their guide. After identifying a candidate gene, they knock it out in an experimental model, gather data and then revise their initial hypothesis.
“This is very inefficient,” notes Yang Lu, a computational biologist who imagines a better way forward. “Right now, we are in the right time, in the sense that we have been collecting so much data in the past decades, the resolution of data has become finer and finer, and we have more and more powerful artificial intelligence tools. Can we use AI to accelerate this tedious process? Can we just try to find needles from a haystack in an automatic fashion, purely by looking at data? This is what I pursue.”
That pursuit has brought Lu to the University of Wisconsin-Madison’s Department of Biomedical Engineering, where he’ll continue his work using artificial intelligence, machine learning and other statistical methods to inform biological research. In fall 2025, he joined BME as an assistant professor after spending the past two and a half years at the University of Waterloo in Canada, where his work focused more on data science methodologies.
That’s what Lu originally set out to do as an undergraduate student at Shanghai Jiao Tong University in his native China. He wanted to work for Microsoft as a software engineer and interned at the company. But he discovered computational biology as a master’s student at Shanghai Jiao Tong University, which prompted him to turn down a job offer from Microsoft and pursue a PhD in the United States. He joined the computational biology and bioinformatics program at the University of Southern California in 2013.
“I was exposed to different kinds of training, so I learned the language, talked to different people, like statisticians, biologists and computer scientists,” he says.
A four-and-a-half-year postdoctoral stint in the Department of Genome Sciences at the University of Washington deepened his interest in biological research questions.
Lu has built AI tools to analyze things like the relationship between genotypes (what is encoded) and phenotypes (what is observable), as well as methods to interrogate AI models and assign measures of statistical confidence for each hypothesis they generate. He sees his work as helping biological researchers prioritize their experiments and laying the groundwork for an AI-aided—and greatly accelerated—research workflow in the future.
Now, he’s thrilled to come to UW-Madison as part of the campuswide RISE-AI Initiative and connect with researchers in BME, the School of Medicine and Public Health and across campus.
“These powerful tools cannot be useful without applying to real disease,” he says. “This is the perfect fit, because I can sit with a bunch of people who care about biomedical problems—who really want to solve problems by building something, and then my expertise can greatly help. I know there are a lot of opportunities—not only research opportunities within UW-Madison, but outside the university, for example, companies like GE Healthcare, Epic Systems, and even if we go further away, in Chicago there are many big pharmaceutical companies. They can offer a lot of opportunities to translate my research into real-world impact.”