Evan Cofer, 26, doctoral student
Cofer is a doctoral student in quantitative and computational biology and is pursuing the Graduate Certificate from the Center for Statistics and Machine Learning (CSML). He works with Professor Olga Troyanskaya, jointly appointed in Computer Science and the Lewis-Sigler Institute for Integrative Genomics. Before assuming his Ph.D. studies, he earned a bachelor's degree in computer science at Trinity University in San Antonio, Texas, and his master's degree in quantitative and computational biology at Princeton.
Cofer works at the intersection of computer science and biology by applying cutting-edge machine learning and data mining techniques to sizeable biological data compendia.
"I try to find things that maybe aren't obvious by looking at data," Cofer said. "In particular, my work focuses on applying deep learning to genomics, a subfield in biology that's concerned with the mapping, evolution, structure, and function of genomes."
Cofer's research focuses on a subset in genomics called regulatory genomics. This domain encompasses 80 to 90 percent of the human genome. While this amount is large, these subsequences do not encode protein sequences and are controversially known as "junk DNA" because they do not seem to have a biological function.
"We don't know what it does, and there has been much focus on trying to figure out what it does," said Cofer. "In the past ten years, we have realized that they have regulatory potential. Near this junk DNA, there might be an important genetic sequence, and this junk DNA may regulate when that genetic sequence gets transcribed, and eventually translated into a protein."
Cofer and other biologists are using various technical methods to detect regulatory activity throughout a genome. However, Cofer said there is still difficulty decoding what these non-coding sequences do just by looking at them. That's where the work that Cofer and his fellow lab mates are doing comes in: developing deep learning models to map DNA sequences to regulatory activity. These experiments and others can lead to insights into how cancer evolves or how autism, congenital heart defects, and other conditions occur within humans.
A related project in that vein is DeepArk, a deep learning model capable of predicting epigenetic features from DNA sequences for model organisms such as the fruit fly. Other work he has undertaken includes developing machine learning tools for biologists who don't have the same level of expertise in computer science.
In pursuit of his research, the graduate CSML certificate program has been beneficial, Cofer said.
"SML 510, the research seminar class, has been great," he said. "I like the class because it has made me better at speaking about my research to an audience that may not be working in biology. It's also awesome to see what other people are working on, and it has helped me feel more connected to the machine learning community in Princeton."
Cofer is unsure what he will do after earning his Ph.D. degree; a postdoc position is one option.
"I've enjoyed my research experience at Princeton very much. It's been a lot of fun working in the lab," he said.
Cofer is a member of the Genetics Society of America and the Association for Computing Machinery.
Cofer enjoys hiking, bird watching, cooking, and playing retro video games in his free time.