As a postdoctoral fellow at the Broad Institute in Cambridge, Massachusetts, Dr. Gavin Ha developed a computational tool that can sift through data generated from low-cost DNA sequencing to estimate how much of the DNA fragments — cell-free DNA — circulating in a patient’s blood originally came from tumor cells. Though he’s now a faculty member at Fred Hutchinson Cancer Research Center, Ha still makes sure this tool — and many others that he has developed in his career, which are all open-source and freely available to other researchers — are updated with new features and continue to function.
“I still respond when people alert me to a bug or have technical questions [about my tools],” said Ha, a computational biologist who develops machine learning and statistical tools. These tools help scientists and Ha himself analyze and understand how changes to DNA can drive cancer development, as well as detect DNA hallmarks that distinguish tumors from normal DNA.
“It’s very rewarding that people are still using these tools and that it’s making an impact in their research,” Ha said.
This interest in personal connection is reflected in Ha’s goals for his own work, which melds computer science and biology.
“I started out as a computer nerd, but I always thought there had to some good applications for this, beyond video games,” said Ha, who majored in computer science and microbiology and immunology at the University of British Columbia. “My major wasn’t in bioinformatics, it was a mix of disciplines. There were some early bioinformatics pioneers [at UBC] and they taught courses that unified [computer science and biology].”
By the time Ha was ready to go to graduate school, UBC had put together a graduate program in bioinformatics. During his Ph.D., conducted at UBC and the BC Cancer Agency, Ha was able to begin applying his skills to develop computational methods with real-world implications for human health, creating tools that could reveal DNA changes characteristic to ovarian and breast cancer cells. A minor component in one of his projects in his graduate work first introduced Ha to cell-free DNA, or DNA that’s released from dead or dying cells and found its way into the bloodstream. Though the cell-free DNA sequencing techniques at the time weren’t as well established as they are now, scientists like Ha saw that it might one day be possible to transform clinical care of cancer patients by using use information from tumor DNA in the blood.
Though not the focus of his Ph.D., cell-free DNA stuck in Ha’s mind as he transitioned into a postdoctoral fellowship at the Broad and the Dana-Farber Cancer Institute. About a year into his postdoc, Ha was able to finally expand his studies into cell-free DNA and liquid biopsies, which continue to be a major area of research.
If successful, liquid biopsies will allow doctors to detect cancer, like breast and prostate tumors, through a simple blood draw. Ha is working to develop algorithms that will be able to comb through the genetic analysis of the blood drawn to reveal the presence of DNA released from cancer cells somewhere in the body. Scientists hope to someday use liquid biopsies to easily detect and diagnose cancer, monitor how patients’ tumors are responding to treatment, whether they’re developing drug resistance, and discover when a tumor has recurred.
“What excites me is finding new problems to solve,” Ha said. The potential to improve human health is also a big draw, he said: “There’s a real urgency to translate these methods to the clinic.”
Liquid biopsies are designed to be simple — for the patient. The science behind them is quite complex. Much of the DNA from a cancer cell is indistinguishable from a normal, healthy cell. The molecules of cell-free DNA floating through the blood are just small chunks, not even full chromosomes, so scientists like Ha must figure out how to distinguish normal and tumor DNA using small snapshots of a bigger picture. Some of Ha’s projects focus on characterizing hallmark changes that set tumor DNA apart from normal DNA and drive a tumor’s development and progression and shape its behavior.
“[Analyzing cell-free DNA in the blood] is very, very challenging to do, because we're looking at potentially one in a million or one in 10 million molecules.”
The other major challenge for scientists working on liquid biopsies is sifting through enormous amounts of DNA molecules from normal cells (mostly white blood cells) to find the rare few molecules of DNA that leaked out of a tumor cell.
“It’s very, very challenging to do, because we're looking at potentially one in a million or one in 10 million molecules,” Ha said.
His goal is to develop the sophisticated computation methods that are needed to find the needle of tumor DNA in the haystack of normal DNA. Once accomplished, the next step is to move the methods to the clinic and test their efficacy for patient care.
“Once you come up with a good assay, with a good analytical, computational approach — which is what we focus on — then can we test it in a practical setting?” Ha said. “For example, in clinical trials, can you stratify people into different treatment arms based on decisions from cell-free DNA?”
His work sits squarely in the continuum between pure biology and pure computation. It gives him the opportunity to collaborate with scientists from many different research fields, as well as act to connect computational approaches with biological problems that need solving, Ha said.
“I find that the best way to think creatively and find new ideas is through collaboration, because I learn so much from others,” he said. “I am confident that being a part of the scientific mission of the Steam Plant will present new opportunities to collaborate and begin another chapter in my research.”
― By Sabrina Richards, Nov. 9, 2020