Brown CS Blog

In addition to advancing the field at a local and national level, Brown CS community members have a global impact on education and research. For more articles on Brown CS around the world click here.

Brown University’s Department of Computer Science (Brown CS) and Center for Computational Molecular Biology (CCMB) are looking forward to giving a record number of talks at one of the most prominent conferences in computational biology. Current and former students and post-docs of professor Ben Raphael presenting at the twenty-second annual International Conference on Intelligent Systems for Molecular Biology (ISMB 2014) in Boston include Iman Hajirasouliha, Max Leiserson, Layla Oesper, Anna Ritz, and Hsin-Ta Wu, featuring work co-authored with Ahmad Mahmoody, Gryte Satas, and Suzanne Sindi.

Ben credits the strong Brown CS representation to a little bit of fortunate timing and a lot of ongoing effort from a dedicated group of researchers. “It’s surprising that all these projects came together at the same time,” he says, “but it shows that we have a hard-working group with a strong culture of mentoring to help each person reach their full potential. Our small size requires excellence across the whole team, and I’m proud that we’re able to compete successfully against much larger groups, not only in CS but also in medical schools and research institutes.”

Layla Oesper and Gryte Satas are happy denizens of what they consider to be a technological leading edge. “Understanding/analyzing sequencing data from cancer genomes is a difficult task,” they explain. “There are many factors that make identifying the landscape of mutations in a heterogeneous tumor sample hard.  Our algorithms are aimed at quantifying this type of information, an important first step in determining what mutations drive cancer. This wasn’t possible even five years ago, and it lets us make use of the vast amount of data that has been accumulating."

“We’re also looking forward to the inaugural Raphael Reunion!” they laugh, explaining that the conference will allow all of Ben’s current and former students to reunite in Boston. What do all these colleagues have in common? “Momentum, collaboration, and energy,” says Layla. They’re qualities that are evidently shared by their mentor. “Ben is as passionate about our work as he is about his own,” Gryte says. “How does he find the time to sleep?”

Ben shrugs. “What makes me happy is that all these projects include multiple authors who are great team players. Because our group is part of a CS department, we can recruit people with strong skills in algorithm development and program design, which really gives us an edge.”

Max Leiserson’s collaborators, in addition to Raphael, include two researchers from Tel Aviv University. His highlight talk focuses on a paper they published a year ago, about an algorithm developed by Brown CS called Multi-Dendrix. Without prior information, it searches for genes with approximately exclusive mutations and high coverage in a cohort of tumors, which enables the identification of “driver” genetic pathways that cause cancer when mutated.

Asked for his goals for the conference, Max says, “I’m looking forward to all the talks, the chance to learn from others. My biggest hope is that people will get excited about our results and even more cancer researchers will use our software.” He explains that external researchers will have an even easier time using Brown CS tools in the future: currently, some of the applications necessary to run Multi-Dendrix are proprietary and need to be purchased, but an upcoming transition to entirely open-source software will allow maximal ease of use.

While working under Raphael, Iman Hajirasouliha, Anna Ritz, and Hsin-Ta Wu’s collaborators have included colleagues from the bioscience industry as well as academia (Brown University and elsewhere). Hsin-Ta will present a paper on detecting copy number aberrations in cancer co-authored with Iman, and Iman will present a paper on intra-tumor heterogeneity co-authored with Brown CS’s Ahmad Mahmoody. Both were among 29 papers directly accepted in the first round of peer reviews, out of 204 submissions.

“Only recently,” Iman comments, “have scientists realized that mutations that we formerly thought of collectively, such as breast cancer or lung cancer, actually vary considerably from person to person. Our task now is to characterize heterogeneous mutations across a tumor. It’s a major step forward.”

“I’m really eager to introduce people at the conference to our new research about finding driver recurrent copy number aberrations,” adds Hsin-Ta. “Our new method has advantages in not only accurately identifying candidate copy number aberrations which could drive cancer but providing an algorithm which is simple and fast, and readily adaptable for high-throughput sequencing data. For us as computer scientists, in the future it will be exciting to apply this method to larger cancer datasets as more and more cancer patients are sequenced.”

Ritz and Hajirasouliha’s shared research into structural variants across a single genome, not specific to a particular disease, was partially aimed at the challenges caused by software limitations. “Second-generation sequencing platforms,” Anna explains, “are slow but quite accurate, with an error rate of one or two percent. Third-generation technology gives a more in-depth analysis, but has a fifteen percent error rate. Our algorithm is probabilistic, and it shows that combining the two platforms allows you to reduce the impact of errors while maximizing the third-generation benefits.”

Anna sees real parallels between Brown CS’s progress in computational biology and the field’s increasing opportunities: “Ben’s first year at Brown was my first year as a graduate student. His great ability to locate the next big problem is helping us make major contributions to the field. Just as one example, our results are helping overcome the hesitation about third-generation sequencing paradigms. If we can prove their worth, they’ll get used, and medical progress will be made.”

Ben agrees. “It’s so exciting,” he says, addressing prospective ISMB 2014 attendees and celebrating the members of the research group that he’s put together. “We’re thrilled to be building on our long-term record of strength in this field, to share our results and demonstrate what a great environment we have for students of all levels. I hope my team gets to see the appreciation for their research.  I think it will help them understand how much all their effort, day in and day out, has achieved.”