Meet Dr. Russell Schwartz
Tell us about your background. Where are you from?
I’m originally from New York and grew up in the suburbs around New York City. I went to high school in Greenwich, Connecticut. Afterwards, I went to MIT for my undergraduate studies as a computer science major. I stayed there for my Masters and PhD, also for CS. Although my degrees were in computer science, my research was actually related to biology since my undergraduate days. I did not go to college planning to do biological research but when I went looking for an undergrad research project, I happened to find one in computational biology. Once I started working on it, I realized what a great area it was, and I’ve been working on computational biology ever since.
What are your current research projects?
I work on a lot of different areas of biology, such as simulation methods, biophysics, biochemistry, and genetics. My group has a few primary projects in individual areas but also a lot of side projects, often at the intersections between these topics. I think that a big advantage of being in computational biology relative to experimental biology is that it’s much easier to try out new ideas. There are so many interesting problems to be solved that it’s never hard to find something to work on.
One current project I’m excited about is studying tumor evolution—learning to understand how a healthy cell mutates to become a tumor cell and how that initial tumor cell progressively evolves into more aggressive forms. This general question of tumor evolution is something a lot of groups are working on from different directions. We’ve specifically been using phylogenetics (the computational inference of evolutionary relationships) as our approach by treating a tumor as a population of cells in a common environment. We then try to infer evolutionary trees on tumors and in the process identify the major events in tumor evolution in order to understand what evolutionary steps tend to be shared by tumors across many patients. We want to know how tumors become more aggressive, what genes are the underlying factors, and how we can use all of that information for diagnostic tests and new drugs.
We hear you’re the new Co-Director of the Ph.D. Program in Computational Biology. What does it involve?
The Computational Biology Ph.D. Program is a joint program with the University of Pittsburgh, so they have a Co-Director at Pitt, Panayiotis (Takis) Benos. He and I are responsible for supervising the structure of the program and the progress of the students. There is a lot of administrative work, such as meeting with people to discuss changes to the program structure or teaching of program courses, working with others on student support, meeting with the students to keep track of their progress, and discussing with them what the expectations of the program are and what changes they want to see to improve the program. A major part of the work is representing the program to prospective students or interested faculty members, as well as to others in the field outside of Pittsburgh.
Are you seeing big changes in your department? If so, what changes have you helped make? And what are your plans for the department in your new role?
I think to a large degree, our formal computational biology program has been centered on the PhD program so far. The PhD program is relatively young, ending its fifth year now. So a lot of what we’re trying to do is settle it down to a steady state in terms of the number of students and the structure of the program. Hopefully in the next few years, we won’t need large changes. We are, however, looking to expand our efforts with regards to undergraduates, which includes better serving our undergraduate computational biology majors, a small group we would like to expand, as well as trying to improve computational biology training of undergraduate biology majors.
Right now, we are going through some big changes because Computational Biology recently became a department of SCS. Thus, we’re integrating with the other SCS programs. And in general, we’re just trying to figure out how the program needs to evolve to bring in the right mix of students to meet the research needs of the faculty, how to develop the program’s reputation, and how we can become more competitive in making prospective students aware of what Carnegie Mellon has to offer for those doing research in this area.
What roles have you taken prior to being Co-Director of Computational Biology, including before Carnegie Mellon and at Carnegie Mellon? How did it happen?
I’m one of the directors for the Bachelors of Computational Biology and Masters of Computational Biology Programs. Before coming to Carnegie Mellon, I’ve spent a couple of years in the industry at Celera Genomics, working on the sequencing of the human genome. For someone in computational biology, that was a great opportunity to be involved in a unique and important project at the cutting edge of biological research. Being at a university, however, is a different type of challenge. I think both have a lot of appeal to them.
For me, academia is definitely the preference after having experienced both. Both environments have their pros and cons. In the corporate world, you’re part of a team contributing to a common project. You have more things taken care of for you in the corporate world than in the academic world. As a professor, you’re personally responsible for a lot more things working well, including teaching and research responsibilities and many other administrative and management responsibilities. The main appeals of the academic world are teaching—you have to be in the academic world if you like teaching – and more control over your research agenda. Being able to choose your own research direction and decide for yourself which projects you want to work on are big advantages of being in an academic environment.
Which institutions and departments does the Lane Center collaborate with? How does the collaboration work?
We run a joint PhD program with Pitt, so students in the program are treated as much as possible as members of the program, not a particular university. We run our admissions together, have a common curriculum, and co-teach the courses. It’s really one student body though some students are physically at one school and others are at another. We also have shared seminars and clubs; we really try to make it one program for both universities. Sometimes, there are tensions when trying to meet the requirements for one university and meet the needs of the program the same time. It’s worked out well so far though. It’s very advantageous to have faculty from both universities. We’re able to take the strengths of both universities and put together a world-class program.
We also have many active collaborations within Carnegie Mellon. The Lane Center itself has members with appointments in many other CMU units, including several SCS departments and others such as Biological Sciences, Physics, Statistics, and various kinds of engineering. Some of these collaborations are on programmatic or curriculum issues and in other cases they are collaborations on research. The Lane Center helps to facilitate research collaborations, but most arise spontaneously from faculty members learning about one another’s work and seeing opportunities to combine their expertise.
How do you test the applicability of the theoretical models created to a real system?
It varies from problem to problem. This is always an important question: when you’re doing theoretical work, how does it connect to the real world? It’s more of a challenge with biology than with many other disciplines as biological systems tend to be so complex. For my own work, it’s often a matter of working with collaborators who run experiments and gather valid data since I only personally do computational work. A computer model can often suggest novel experiments whose outcomes can be compared with the oredictions of the theoretical work. In other cases, testing a model consists of going to literature and coming up with benchmark data sets where some ground truth is known that one can use to validate a model. In other cases, there may not yet be an experimentally feasible way to test a model and we may have to make a judgment call that the theory is promising enough to pursue even though it may be years before it can be rigorously tested. There’s really no one right answer.
Does cancer research mostly aim to treat, prevent, or to detect cancer? Which aim relates most to your own research?
We’d like to be able to do all of them. There is a lot of interest in the field of therapeutics and helping to treat cancer. Even if you have a good idea of what genes you want to target, though, it’s a long, high-risk, and expensive process to actually develop drugs to target them. And usually, it’s going to end in failure. People like myself working on the basic research side hope that by better understanding the process of tumor development, we will be able to come up with different target genes that might be influenced by a hypothetical drug and then leave it to others to actual develop and validate drugs.
My work is often upstream even of that process, focusing on developing the informatics tools we need to understand at a basic level how tumors develop. From that understanding, we can make decisions that will help us identify therapeutic targets or come up with new diagnostics to figure out who has what particular type of cancer and whether any given person might respond well to any particular treatment. It can also be useful to be able to predict someone’s prognosis, even if it doesn’t tell you what treatment to give to the person.
How will understanding genetics and phylogenetics help improve society’s wellbeing?
Genetics and phylogenetics come into improving society’s wellbeing in various ways. Our cancer work is one of the ways that we hope to improve the wellbeing of society by helping us identify sub-populations of cancers and improve our ability to develop therapeutics and diagnostics. But that’s a fairly specific application; there are other directions as well. Other people are trying to use phylogenetics to find genes that may underlie people’s risk of getting common diseases. Variations in one person’s genome relative to another’s can influence how likely each is to get some disease or how well each person is likely to respond to a certain type of treatment. Identifying genetic risk factors of disease is a large problem in statistical genetics, and understanding how human populations are structured and how these populations have developed can help us with very practical questions about identifying the genes and genetic variations involved in common diseases.
More generally, phylogenetics really underlies a whole approach to many questions in biology known as comparative genomics. In a comparative genomics approach, one looks for similarities in the genomes of different organisms to help make predictions on how their genomes function. Having a good model of the evolution of organisms helps almost any kind of analysis on any biological sequences or structure. In that sense, phylogenetics is really at the heart of many of the most important research questions in biology today.
You’ve designed several courses. How do you design classes to effectively teach students both biology and computer science when students are likely to enter at different levels?
It’s definitely a challenge to develop a class that’s suitable for people with different backgrounds and that always is an issue when you work in an interdisciplinary field. To some degree, that’s something we can handle in class. I may have to present a lot of background material that is review for some people and new material to others. It’s not ideal, but sometimes there’s no way around it. In other cases, we try to identify specifically the material that be will useful for many different groups, and focus on material that’s new to everyone. To some degree, it’s something you have to handle when targeting any class. You can’t have a class suitable for every population, though, and that’s one reason we often come up with new classes. For example, in the past term, Robert Murphy and I designed a new introduction to computational biology targeted toward undergraduate biology students because it became apparent that we simply couldn’t offer the same introduction to computational biology to biology undergraduate students as we would to undergraduate or graduate computer science students.
What are some of your hobbies and activities you participate in outside of work?
I enjoy hiking, though I don’t have a lot of time for it these days. I usually exercise by running or biking. I’m not sure if you can consider this a hobby, but I enjoy walking or wandering around. It can be hard to find time for hobbies when you’re a professor.
Who is your role model?
There are a lot of people I admire and learn from, but it’s hard to pin down any one person I’d say is my role model. I guess I would say that I view personal inspirations much like scientific inspirations; if you try to develop a broad perspective and keep an open mind, you can find good ideas in many places.