National Science Foundation grant provides PhD student experience with machine learning under FNL mentor

Image

When Chloe Thangavelu, a Ph.D. student at the University of California, Irvine, received funding from the National Science Foundation for a six-month internship, she eagerly set out to find an internship where she could combine her biological and computational expertise and gain experience with machine learning.  

That’s why she reached out to the Frederick National Laboratory, and specifically, Eric Stahlberg, Ph.D., director of the Cancer Data Science Initiatives, who had exactly the expertise she was looking for. 

“Science is becoming increasingly interdisciplinary, so I view gaining experience in other disciplines as an advantage when tackling challenges in health and medicine." – Chloe Thangavelu, Intern at Frederick National Laboratory.

Stahlberg was quick to respond, welcoming her to join his team.  

“Providing valuable opportunities for the next generation of scientists to gain valuable experience across disciplines is essential to the team science that accelerates progress in cancer,” Stahlberg said. “With the future of cancer research squarely involving data science and machine learning approaches, the time with the Frederick National Laboratory team will deliver benefit and impact long after the individual experience comes to a close.” 

Diving into machine learning 

Thangavelu had some prior experience with coding, but none with machine learning. However, she recognized its value and was eager to leverage it for biological research.   

“With the sheer amount of data in biology, I think machine learning is really well adapted to handle that data load,” she said. “I was curious about how to combine the two.” 

Thus, Thangavelu spent the last six months exploring various machine learning tools, including the ATOM Modeling Pipeline (AMPL), generative molecular design, and CANcer Distributed Learning Environment (CANDLE) under Stahlberg’s mentorship. The goal of her project is to create chemicals using generative molecular design and then leverage AMPL to select the ones with cancer therapeutic properties.  

“There is a learning curve,” Thangavelu noted, “but that’s the point of this internship.” 

The project was also challenging in another way. Her background was in biology, while this project was heavy on chemistry applied to biology. However, it was a challenge she happily accepted. 

“Science is becoming increasingly interdisciplinary, so I view gaining experience in other disciplines as an advantage when tackling challenges in health and medicine,” Thangavelu said. 

Although most of her internship has been remote, she was able to visit the Frederick National Laboratory, where she met with her mentor and other members of the Cancer Data Science Initiatives. She was also introduced to other colleagues across the laboratory, with whom she has remained in correspondence.  

In addition to gaining experience with machine learning and chemistry, she said that this experience has taught her project management skills, and it has helped her decide where she might like to work after she finishes her Ph.D. this summer. She’s now considering future employment at a federal laboratory or government agency.  

“I really enjoy the philosophy at FNL,” she said. “The community consists of collaborative, altruistic scientists and staff striving for the greater good.”  

Stahlberg also benefited from serving as a mentor.  

“Being part of the National Science Foundation Internship program has been a great experience,” he explained. “Working with Chloe these past few months has been absolutely wonderful. She’s been a great addition to the ATOM team, quickly learning the technologies, pushing through challenges, and helping the team reach key goals to impact the community with exciting new AI-based molecular and drug discovery approaches.”  

Training at FNL 

The Frederick National Laboratory offers a unique setting for training the next generation of biomedical leaders. Trainees range from high school students to doctorate degrees: PhD, PharmD, MD, and JD students. Over the past three years, the Cancer Data Science Initiatives has trained more than 65 students on cancer research and the development of data science tools and resources.