Computational biologist Sohini Ramachandran and her team are mining human genetic data to understand the impact of our ancestry on our health.
Not long after Sohini Ramachandran, PhD, joined the Department of Ecology and Evolutionary Biology, in 2010, Brown’s undergraduates were buzzing about the new assistant professor.
“Even though I never took a class with her, I knew who she was,” says Sahar Shahamatdar ’13 MD’22 PhD’23, who was then an engineering concentrator. “She was well known … for being an amazing professor and mentor.”
Aaron Behr ’15, who was studying biology and computer science, heard similar talk. “Dr. Ramachandran was consistently recommended as a really great person to work with, a really great adviser doing really interesting research,” he says.
It was the kind of recognition Ramachandran had hoped for when, as a postdoc at Harvard, she was mulling offers for her first faculty appointment. “I really wanted to be in a campus like Brown where I had access to undergraduates and could mentor them,” she says.
That’s because mentors played a big role in the shaping of her academic career, beginning when she was just 15 years old. In the lab of Stanford biology professor Marcus Feldman, PhD, the high school student found what seemed to her an uncommon degree of acceptance and encouragement. “Not only did he give me my own project and treat me like a researcher, but his whole lab did. That was the culture that he created,” she says.
Last year Ramachandran was granted tenure as well as the directorship of Brown’s Center for Computational Molecular Biology (CCMB). Now an associate professor with joint appointments in biology and computer science, she is fulfilling her goal to “pay it forward,” regularly welcoming budding young scientists into her lab with equal measures of support and autonomy.
For Behr, now a software developer at Oracle, completing his undergraduate thesis with Ramachandran literally changed his life. “It ended up being such a cool project that I really got to feel like I owned, and she worked really hard to make that feasible,” he says. “No doubt I was a candidate of interest to [Oracle] because of my experience.”
Ramachandran’s path to science began with role models at home. She grew up near Sacramento, where her parents were statistics professors in the California State University system; her sister, Raga, who is seven years older, told one of her biology professors at Stanford about Sohini.
“Raga was very smart herself, and she said, oh, my sister is much smarter,” recalls Feldman, a mathematician, theoretical biologist, and pioneer in the field of cultural evolution. Intrigued, he invited Sohini’s parents to bring the teenager to his lab. “We started working on our genetic evolution problems, which she picked up how to do very, very quickly,” he says. “It staggered me because, you know, you don’t often see high school students that smart.” He adds: “When people came to my group and they met her, they didn’t know she was a high school student. They thought she was a PhD student.”
Feldman made a huge impression on Ramachandran, too. “He had students who did their degrees in math and physics and biology, and everyone came together in a really collegial way,” she says. “That group showed me at a very young age how wonderful it can be to do collaborative work … and to have this perspective that you can learn from anyone, no matter what their background is.”
In 1998 Ramachandran enlisted Feldman to advise her research project for the Westinghouse Science Talent Search, a science competition for high school seniors. She placed fourth in the country, and was the top female finisher. “Of the 40 of us [in the talent search that year], more than 30 have PhDs now, and many of us are professors,” she says. “That’s really a testament to the fact that mentoring … has a huge influence on young people.”
Ramachandran stayed at Stanford, but decided to major in mathematical and computational sciences as an undergrad. By then, her sister was in the MD/PhD program at the University of Pennsylvania. “I didn’t really want to compete with her,” Ramachandran says. “I thought, I have to do something different from her. And I was still drawn back to biology.” As she considered PhD programs, she consulted with Feldman, who was thrilled to welcome her back to the department. And her timing couldn’t have been more perfect.
The draft sequence of the human genome was released in 2001; evolutionary and medical geneticists thought they’d struck gold. “The idea was, this is going to unlock all the answers to human health,” Ramachandran says. She began working with Feldman to characterize the distribution of human genetic diversity throughout the world. “One genomic sequence doesn’t fully capture how human genetic variation, which is quite low compared to other species, is distributed and may contribute to disease,” she says. But by using the human genome as a “historical text” for her dissertation, she proposed the serial founder effect model that showed that our genetic diversity today decreases with geographic distance from Africa, where humans originally evolved.
“Those data I analyzed in graduate school pale in comparison to the data we have now, in terms of numbers of individuals and number of genomic sites sequenced,” Ramachandran says. But “even with new samples, it’s still pretty clear that when humans crossed large continental barriers, they didn’t go back, and that is one of the things that’s hugely limiting genetic diversity or constraining it in different parts of the world. So the robustness of that pattern is a contribution I feel really excited about.”
Ramachandran and Feldman have continued to collaborate since she graduated, in 2007. They just published a paper together in the Philosophical Transactions of the Royal Society B; at his 75th birthday celebration last year, she delivered a talk on novel statistical approaches for identifying adaptive mutations. “I don’t regard her as a student anymore; I regard her as a friend and colleague,” Feldman says. “And she’s a very valued colleague.”
A Common Goal
After Stanford, Ramachandran was elected to the Harvard Society of Fellows, where for three years she studied coalescent theory, a statistical approach modeling genetic divergence from a common ancestor, with John Wakeley, PhD, a professor of biology.
Around that same time, Brown was building on the success of its undergraduate computational biology concentration to found an interdisciplinary center that would bring together expertise in biology, math, and computer science to mine genetic data for answers to questions about evolution, ecology, and disease. The new Center for Computational Molecular Biology needed faculty, and Ramachandran took note.
“The center is really one of the things that drew me to Brown,” she says. “The fact that I was going to … have close colleagues in computer science and applied mathematics and biostatistics was a huge decider for me.”
Ramachandran chose Ecology and Evolutionary Biology as her home department even though, she jokes, “I never took a biology class.” But this was a “huge bonus for EEB,” says department chair David Rand, PhD, the Stephen T. Olney Professor of Natural History, who helped establish CCMB and hire its founding faculty.
“A lot of the best people doing genomics, population genetics, and computational biology were trained in mathematics, statistics, or physics, because the models and data are really statistically complicated,” Rand
says. Furthermore, her skills met important needs in EEB and University wide. “She’s certainly brought expertise in human population genetics and evolution that was missing at Brown,” he says. “And her course on biostatistics using R as a programming language has been a huge contribution to the biology curriculum.”
Location played a role in Ramachandran’s success, too. EEB had no available office or lab space in its buildings, so she wound up in the Center for Information Technology instead. “That’s been a real blessing, because being physically so close to [computer science]and to applied math has made collaborating a lot easier,” she says.
Rand says “some very exciting papers” have resulted from that proximity. “That’s exactly the kind of synergy you look for,” he says.
“Often in academia it’s thought that interdisciplinary research is the luxury of the senior professor, and as a junior faculty member you really have to build depth and strength in one recognizable discipline,” Ramachandran says. “But I think Brown offers an infrastructure for young researchers to do interdisciplinary research that doesn’t make those two things mutually exclusive.”
Now, as director of CCMB, she says she’s excited to grow the center’s collaborations further, and “strengthen the interdisciplinary community that’s focused on molecular genetics and genomics.” And again, it’s all about location: next spring the center will move a few blocks north to the newly renovated Brown Office Building, which will also house the Data Science Initiative and the Carney Institute for Brain Science. CCMB is hiring a new computational biologist, as well as a geneticist who specializes in ancient DNA with whom Ramachandran will advise an archaeology postdoc.
“Now with [the Carney Institute]and the Data Science Initiative,” she says, “I think that our center is getting rejuvenated … and also is an example of a successful way to bridge so many different disciplines in working on a common goal together.”
Before she came to Brown, Ramachandran’s research mined human genomes for clues about our ancestry. But over the past several years her focus has changed “in ways I hadn’t anticipated,” she says. “We’re shifting from studying the causes of human genetic variation to the consequences of it.”
As Ramachandran developed tools for detecting selection and improving genome-wide association studies (known as GWAS), her work became more relevant to clinical researchers who were gathering huge amounts of patient data and needed to interpret it. “Medical genomics and genomic evolution are fields where we’re just awash in a sea of data,” she says. “But because gathering that data and sequencing it properly and doing quality control is so complicated, the focus of many papers is not usually on new methods, and methodologically extracting the most that they can from that data. And I think right now there’s a real space for people who do methods development.”
A recent project with St. Jude’s Children’s Research Hospital exemplifies this. Researchers there wanted to understand why Hispanic-American children with acute lymphoblastic leukemia seemed more likely than kids of European descent to relapse following the same treatment protocol. To analyze the genetic data, Ramachandran; Ben Raphael, PhD, a former associate professor of computer science at Brown who’s now at Princeton; and her graduate student Priyanka Nakka PhD’18 developed PEGASUS, an open-source software package that analyzes gene associations by phenotype. They found multiple pathways to leukemogenesis.
“There are different interacting genes that are producing the same disease in different ancestries,” Ramachandran says. The researchers are gathering more samples now, from affected children and their parents, to further test their findings. “We really want to push on it more quantitatively,” she says.
Another potential collaboration arose after she gave a talk about PEGASUS to the lab of Abrar Qureshi, MD, MPH, the chair of dermatology. He mentioned that he’d observed an association in his patients between severe psoriasis and type 2 diabetes, both autoimmune diseases. “Is there a chicken and egg? Does one precede the other?” she says. His lab has a lot of data, she says; if she wants to take on this question, her lab has to figure out the best way to analyze it.
“Sometimes I think the view people have about data science or machine learning is a very oversimplified one,” Ramachandran says, where there’s one tool that can be applied in any data context. But in truth you need a whole toolbox, and her lab is devoted to building the right tool for the right job. There’s PEGASUS; and pong, which she devised with Aaron Behr to rapidly visualize and analyze population structure; and SWIF(r), a technique her postdoc Lauren Alpert Sugden ScM’10 PhD’14 developed to identify adaptive mutations in human genomes.
“You have to get pretty enamored with the data. And instead of having a hammer, you actually end up developing a screwdriver, and it has to fit in really well with the data. You have to understand the nuances of the data—the errors that might be generated in the data, the fickle nature of the data,” Ramachandran says. To do what she does, she adds, you also have to get excited about the context. “For anyone who is in computational biology, I think the complexities of organisms and the diversity of life are what really draw them and keep them in the field,” she says.
As energized as she is by the questions she wants to answer, and the methods she develops to answer them, Ramachandran says the key to her research success has actually been the people she’s worked with. “I was once given the advice of starting a collaboration based on the person, not based on the problem,” she says. “When you find people that you get along with well, where you learn a lot from each other in conversation, where you have some shared interest that keeps you talking to each other, over time you will find the ideas, and you will do the projects.”
She met her future St. Jude’s collaborators in 2009 when she gave a talk there, and they knew they wanted to work together. Yet it took years for them to find a project; they finally published their first paper last year. She similarly found collaborators among her colleagues at Brown, and that pushed her to delve into fields she’d never explored before, like epidemiology. “That’s what draws me to science, is collaboration,” she says. “Letting the personal interaction produce the ideas has been really fruitful for me.”
Leading the Way
The importance of people to Ramachandran is evident with a glance around her office. Photos of students, family, and colleagues decorate doors, shelves, her desk, even the lab homepage. When talking about her research she always credits collaborators, including students. She says her work on PEGASUS with Raphael and Nakka took her “whole lab in this new direction.” Of a chance meeting with Alpert Sugden when she was an applied math grad student, which turned into a productive postdoctoral fellowship, Ramachandran says, “Those kinds of accidents are really influential in one’s career.”
As a woman in science, Ramachandran is conscious of the role she can play in other women’s careers. She says her own path was relatively frictionless, but it took her awhile to see what she’d taken for granted. “I didn’t appreciate until really I became a professor how much my mother made an impression on me,” she says. Even though she had only two female professors in college, “it really didn’t occur to me that I was in a minority somehow, or to think, oh, I couldn’t … be the person teaching. I just thought I could do it because I’d seen it.”
Now Ramachandran tries to be that inspiration. She’s been a faculty speaker for the Artemis Project, a free computer science camp for ninth-grade girls that’s run by Brown undergrads. “I wish they had had that kind of camp when I was little,” she says, “just to be in that environment and have it be all female.” When her 9-year-old niece recently participated in a summer coding camp, Ramachandran recalls, “She said, there’s just one other girl besides me. And I said, well, you can think about me and my lab. There are a bunch of women who do this. And she said, I wish there were more girls. She really enjoyed it, but it was sad to hear her remark on that. I wish that would’ve been different.”
Sahar Shahamatdar says she was lucky: even though her science and technology magnet high school and the undergraduate engineering program at Brown were heavily male, “I always felt I had strong mentors I could rely on.” Joining Ramachandran’s lab last year was another stroke of luck. “Everything just clicked. She was extremely helpful in helping me set up and learn about genomics,” Shahamatdar says. “She was more than willing to meet with me at any time to talk about research, to talk about what I envision my MD/PhD to look like.” She adds, “It’s really inspiring to have her as a role model.”
Ramachandran is universally admired, Aaron Behr says. “Something that’s really special about her as an adviser—she really takes the time to think about the needs and learning styles and skills of everyone in her lab, and her students as well,” he says. “I think it shows in that she’s very adored by all of her students and her lab members.”
And they genuinely have fun together. They go on weekend retreats to Ramachandran’s in-laws’ house in Mid-Coast Maine, to cook and hike and share downtime. They’ve done Escape the Room puzzles, and they play trivia games in the lab. When most of the group gathered on a cold day in early April to take a selfie in front of the CIT building for this story, they were crying with laughter.
That cohesiveness, both Shahamatdar and Behr say, is thanks to Ramachandran’s open, engaged leadership style. “It always really impressed me how much Sohini was thinking about other ways to improve, even though she does such a good job,” Behr says. Her husband, Assistant Professor of History Jeremy Mumford, PhD, and their daughters, ages 5 and 1, come to Maine with them, but the retreats aren’t all recreation. “We each prepared three talks throughout the weekend, which was really intense,” Behr says of one retreat. But he got to learn more about his labmates’ research, and “it was a great way to practice giving a lot of talks on the fly.”
Besides team building and professional development, Ramachandran looks forward to the retreats as a chance to discuss the direction the lab’s going and plan the year ahead. “It helps solidify the themes that we’re focusing on,” she says.
Shahamatdar says she respects how openly Ramachandran shares her vision and takes feedback. “She fosters a very collaborative environment,” she says. “We’re expected to come in every day, and that’s not so that we sit there so that she can just see us do work, but it’s so that … we can interact with each other and we can interact with her.”
Feldman, whose inclusive, cooperative lab culture is a model for Ramachandran’s, says ever since his mentee was in high school, she was “always upbeat, always easy to interact with other people, generous with her time.” It’s been a “treat,” he says, to watch her become a consummate scientist, wife, and mother. “I don’t think I’ve ever met anybody who is as personally accomplished in every dimension of life,” he says.
And now she’s paying that forward, to the next generation of scientists. “I hope that one day I could follow in Sohini’s footsteps,” Shahamatdar says. If she someday has her own lab, she says, “I would definitely think back on what the lab culture was here, and the types of things that Sohini did, and try to emulate that.”