Skip to navigation Skip to content

News

#IAmScience: Dong Xu

Data connects all: ‘Champion collaborator’ Xu bridges research disciplines with bioinformatics

DongXuGraphic

By Cara Penquite | Bond LSC

Dong Xu extracts wonder from numbers with a keyboard and eager teams of scientists at his fingertips.

With his salt-and-pepper hair visible above the cubicle walls and his voice softly but steadily articulated, the beauty of bioinformatics takes shape in his mind although it might not be inherently evident in the rows of computers tucked into a small first floor lab.

Xu weaves a multifaceted masterpiece of research methodologies and makes sense of a sea of data from cell biologists, plant scientists, engineers and many more with hundreds of publications to show for it.

“For research, collaborating is key because the nature of research requires many views, skills [and] knowledge,” Xu said.

A Bond Life Sciences Center researcher and a newly endowed Curators professor of bioinformatics in the School of Engineering, Xu harnesses data to better understand biological systems such as revealing a better picture of protein dynamics or analyzing genetic codes for individual cells. With advanced data interpretation and machine learning, the Xu lab opens opportunities for more in-depth research across Mizzou and the world.

With so much existing on a microscopic level in living creatures and new technologies to collect data continually evolving, massive amounts of information are extracted in biological research. Xu’s expertise is invaluable to interpret massive amounts of information, but he takes it a step further predicting how pieces of biological systems interact.

His recent efforts include using deep learning — a subset of artificial intelligence and machine learning — to understand how the shape of protein binding sites change when molecules bind to other areas of the protein. This research, recently published in Nature Communications, and represents one of many collaborations the Xu lab is known for — in this case with Jilin University in Changchun, China.

The Xu lab’s help is crucial to researchers whose specialties lie in biological systems rather than computer science. Bond LSC Director Walter Gassmann, recognizes Xu’s niche.

“Biology has become so complex that you can’t be expert in everything,” Gassmann said. “The way to move forward is to collaborate with people, and that’s what Bond LSC is all about — bringing people with different disciplines under one roof, letting them rub shoulders and figuring out how best to solve a problem.”

With computational analysis being key to many research studies, the Xu lab collaborates with various labs at the research center.

“Dong is the champion collaborator [with] so many connections in the center,” Gassmann said.

Xu makes note of the particular importance of working together when mentoring young researchers. Students in the Xu lab work in teams and alongside biologists, allowing the researchers to learn from one another.

“There is actually an African saying, ‘it takes a village to raise a child,’” Xu said. “The same is true for mentoring students. It really takes many people to mentor a student. That includes not only community members and collaborators, but also peers.”

Xu fosters that shared work ethic by giving credit where it is due when it comes to research. While other institutions sometimes only give the lead investigator full credit for a project, Mizzou recognizes the work put in by each collaborator for any given project.

“MU is a very collaborative environment,” Xu said. “It really supports interdisciplinary research, [and] I don’t take it for granted. Interdisciplinary research usually relies on administrative support.”

Bioinformatics was more of an afterthought to Xu despite its prominence in his life now. His studies started with bachelor’s and master’s degrees in physics, and it was not until he started Ph.D. studies that his focus shifted.

“I was working on the biophysics of saturated proteins, and that’s a computational analysis,” Xu said. “I became very interested in computational work, so since then I’ve been working on this computational biology, or bioinformatics, for about 30 years.”

Xu’s drive to know all things data is revealed in his excitement to talk science.

With a small chuckle he noted that research does not usually bring money and fame. While he could use his computer skills to gain such rewards working for tech companies, he chooses to pursue science instead.

“To be a scientist requires passion,” Xu said. “You really need passion and to believe in the impact, the value of research. I do feel a reward in that regard.”

The Xu lab develops machine learning, a subset of artificial intelligence that involves teaching computers to mimic human intelligence. That starts with developing technology to collect data and facilitate its analysis.

This stretches into deep learning, more advanced machine learning that involves layers of neural networks. The input information is organized through these networks as the computer makes sense of the information.

People encounter deep learning in their day-to-day lives. The algorithm that sorts photos by faces in your phone’s photo app — that’s deep learning.

“It can start from these pixels and then can retrieve information like eyes [and] nose, and then can [find] a match,” Xu said. “It’s not only deep layers, but also the complex architecture of the network, so that’s why it’s called deep learning.”

A similar process is used in the Xu lab’s work. Rather than sort pixels to recognize faces, Xu’s work sorts data — like amino acid sequences — to predict protein interactions.

By finding trends in large sets of data, deep learning can speed up existing processes.

Mark Hannink, a Bond LSC principle investigator, recently saw this in action. Xu developed techniques to read scientific articles and generate diagrams of protein pathways described in the database. While people can read and summarize these articles, the process takes time and leaves the latest of the diagrammed pathways outdated.

With deep learning, each time a new article is published, the diagram could automatically update.

Working closely with Xu, Hannink saw his mentoring approach in action.

“What I’ve been impressed with is how good of a mentor he is to the students working on this project,” Hannink said.

While high-impact research is one goal, Xu also takes care to develop scientific minds.

“Not only [do] we produce papers [and] software, we also really need to produce high-quality, next-generation researchers,” Xu said. “So, our goal is to mentor people to be successful, and many of our lab members are really on the right track to be successful.”

Article originally published on Decoding Science.