John Wesley Cain, 34, started graduate school with a mathematician's aversion to biology. He took a course in his first semester at Duke University with David Schaeffer, an applied mathematician who was just beginning to study models of cardiac rhythms. In the class, Cain had to choose from a list of projects and ended up working on mathematical models of cardiac action potential. "I think that was secretly his favorite project," Cain says.

Cain himself took quickly to the work. "I thought the mathematics was cool. I thought the applications were cool." Eventually, Schaeffer became Cain's Ph.D. adviser. Now, Cain is an assistant professor at in the Department of Mathematics and Applied Mathematics at Virginia Commonwealth University in Richmond. There, he works in applied mathematics with an emphasis on cardiac electrophysiology.

Much of the work he does is in interdisciplinary teams. In fact, he is a co-principal investigator on a training grant in computational cardiology that focuses on teamwork. "The idea is to try to get clinicians, basic science researchers, mathematicians, computer scientists -- you name it -- to actually talk to each other," Cain says. The culmination of that grant will be the World Congress on Mathematical Modeling and Computational Simulation of Cardiovascular and Cardiopulmonary Dynamics at the College of William and Mary from 31 May to 3 June.

This summer, Cain will move to a new position as an associate professor at the University of Richmond, which, he says, is more geared toward undergraduate education. "I have a lot of projects that I have been really itching to get some of their undergraduates involved [with]," he says. He will continue his collaborations with VCU, in part for its medical center and team of cardiologists.

Cain spoke with Science Careers earlier this spring. Below is a partial transcript of the conversation, edited for clarity and brevity.

J.W.C.: You can take a natural phenomenon, such as a heart attack or anything that you would like to understand the mechanisms for ... [and] you can ... run mathematical and computer-based simulations ... to gain intuition, as opposed to actually having to maybe excise a heart from a rabbit or a sheep or something like that to run an experiment. I ... try to gain intuition so that I can then report to the biomedical engineers and then tell them, "These might be the sorts of experiments that you might want to run."

J.W.C.: Sure. Sometimes, for example, regions of damaged tissue can anchor abnormal wave patterns in the heart, and those sorts of things can be simulated. ... So the idea would be: ... take an example of a substrate for anchoring some particular type of arrhythmia. Try to describe that set-up using a system of equations that could be solved using a computer. And then try to see if we can actually reproduce the phenomenon at least qualitatively and hopefully quantitatively with some computer assistance. Because then we can experiment with different types of damaged regions of tissues to see which ones might anchor abnormal rhythms a little bit better than others and then try to simulate or design an experiment for telling a clinician how you might want to use techniques that they use to terminate the arrhythmia.

[Clinicians] have their own range of techniques, of course. They usually use things like radio frequency ablation ... to try to fix the heart when they detect this type of abnormal rhythm. Our idea would be to just try to simulate such rhythms on a computer and then try to figure out, well, how would we correct those sorts of rhythms?

J.W.C.: We do. The groups tend to be very interdisciplinary. ... Cardiologists are the ones who really keep us honest and they tell us exactly what ... they see when they look at a damaged heart and ... then would need to go in and ablate or do something of that sort. The biomedical engineers are very useful folks to talk to because they are very nice in bridging language gaps. ... I am trained as a mathematician and sometimes it's nice to have somebody who can speak ... from an engineering standpoint and talk about quantitative things, and also understand some of the more technical physiology that's going on.

Physics folks tend to be good at running experiments, just like the biomedical engineers. And then the mathematicians like myself would be more geared towards analyzing mathematical models and computational models that are designed to try to explain different types of rhythm [and] explain mechanisms for generating arrhythmias. So there is a big spectrum of scientists who are involved in these sorts of projects, and it actually makes for very fun and lively discussion groups.

J.W.C.: All of my training was in mathematics. Biology, of all the basic sciences, was the one that I had the least interest in when I first started graduate school.

J.W.C.: Right when I was getting ready to start graduate school, I thought briefly about maybe trying to take an excursion out of academia and getting a job in the tech sector. But that was around the year 2000 when the tech sector was in the process of going belly up, so I decided, well, I will try the graduate school route. And in my first semester I took a math course, an applied mathematics course in ordinary differential equations. And during that course, we all had to choose a project to work on, and one of the projects that we were able to choose from was mathematical models of cardiac action potentials in a single cell.

As it happened, the professor had just gotten involved in that, and I think that was secretly his favorite project. By happy coincidence, he asked if I would be willing to work on that as an independent study with him the following semester. And I said "Sure, okay." I thought the mathematics was cool. I thought the applications were cool. We continued it as an independent study the following semester and then we decided to continue into the following summer. And by the end of the summer, I was ... ready to pop the question of, "will you be my academic advisor? Will you advise my dissertation?" It seemed like a great match.

The mathematics is very rich and the physiology -- there is just absolutely no end to the types of questions you can pose about dynamics in cardiac tissue. So it's kind of a happy story where I believe several nice coincidences happened to just put me on the right track.

J.W.C.: I did have to learn a fair amount of electrophysiology to make sure that I wasn't doing anything too outlandish, because the idea isn't just to come up with some sort of mathematical model that is purely phenomenological to the point that all it's designed to do is to try to reproduce different graphs. If you design your model, you can make a model do anything you like. That doesn't necessarily mean that it's grounded in reality.

So we would typically have roundtable discussions where cardiologists, biomedical engineers, physicists and mathematicians would sit around the table and discuss the physiology. This is how I learned it: from roundtable discussions with a research group where the cardiologists and biomedical engineers were the ones who would really keep us honest and made sure that what we were doing was grounded in physiologically reasonable assumptions.

I tried taking a course one time but after a couple of lectures decided to abort that because the roundtable discussions were a lot more fruitful. I found that it was a lot easier to learn some of the physiology background just by directly asking a cardiologist or a biomedical engineer.

J.W.C.: I would say all of the above. It's not necessarily essential that the clinicians understand the math. It's trying to figure out how to ask the right questions of the clinicians to make sure that your model stays honest.

Usually you want to use the mathematics to try to gain insight as to what you should be looking for, and the more complex the model, ... the less amenable to mathematical analysis it's going to be. So you really have to try to convey to the clinicians what a mathematician's limitations would be so that they can craft their questions in such a way that it helps design an experiment. And that's a really delicate tight-wire act to try to walk.

J.W.C.: Biology used to be the science that was not traditionally the one that mathematicians would gravitate toward. There was a nice article that appeared several years back in the Notices of the AMS, "Why Is Mathematical Biology So Hard?" And in that, [Mike Reed] points out that in biology, there is no Newton's Second Law. There is no F = ma that you can resort to. Physics has traditionally been one of the fertile grounds for using mathematical modeling. ... But I would say that as computing has improved over the last several decades, biological modeling has become a very hot area. And a lot of the mathematical techniques are independent of the underlying applications.

So I try to tell students, even if biology is not necessarily your favorite science, it really can grow on you and the techniques you are going to use will be very similar to the techniques you would use to attack certain problems in physics or chemistry or economics. A lot of those techniques are independent of the underlying application and ... a lot of the most exciting questions are biologically motivated.

J.W.C.: Other than just to drive home this point: that I really encourage any student who is interested in getting into this sort of career, you have to be willing to meet with people in fields ... very different from yours, and you have to listen. It took work at first. The electrophysiology literature has a lot of long words that I didn't understand and a lot of difficult things that pushed the limits of what I had remembered from basic chemistry. There was some start up involved but, boy, the rewards sure did make up for that. It's fun to talk to people from a variety of scientific fields. I encourage anybody who is interested in getting into this, don't hesitate to immerse yourself in any sort of quantitative training that you can and immerse yourself in any other sorts of science. ... It is a very, very fun field to be in.

In 2001, Howard "Jack" West was in his final year as a medical oncology fellow at the Fred Hutchinson Cancer Research Center in Seattle, Washington, when he wanted to develop a clinical trial to test a molecularly targeted therapy for advanced bronchioloalveolar lung cancer. He knew that making the transition from clinician in training to lead investigator running a clinical trial would require expertise he lacked and an infrastructure he was unfamiliar with.

"It's a lot for any single person to do and probably overwhelming for most fellows without that experience," West says. So he applied for and was accepted to a then-new program run by the Southwest Oncology Group (SWOG), one of the handful of National Cancer Institute (NCI)???funded cooperative groups that design and run large clinical trials in cancer in the United States. During the training course, each early-career investigator completes a research protocol for a phase II or III clinical trial. "To be able to be walked through [the process] with a team of people who have walked that road before -- that's huge," West says.

The SWOG course and a small handful of others are designed to teach young investigators the skills they need to design and carry out a clinical trial in cancer and to teach them where to find collaborators and mentors. There are other opportunities to learn these skills, too: Fellowships and university courses can also provide a strong foundation in clinical trial design. And there's one key skill that applies whether you take a training course or not: finding mentors to help you through the process.

Feature: Cancer Crusade at 40

This week, Science and its sister publications take a look at where cancer research stands 40 years after the signing of the National Cancer Act. This article is one of two in Science Careers on the topic; see also " Q&A: Finding and Exploiting Cancer's Weaknesses." For more on clinical trials, see " A National Cancer Clinical Trials System for Targeted Therapies" and " Accrual to Cancer Clinical Trials in the Era of Molecular Medicine" in this week's issue of Science Translational Medicine.

See the list of cancer-related articles in all the Science publications at www.sciencemag.org/special/cancer2011/.

Personalizing medicine

Participating in the SWOG workshop allowed West to put together a protocol design for a large multicenter phase II clinical trial, which became known as SWOG 0126. Writing the protocol, which would normally have taken many months, was completed in 3 weeks. Seven months later, the trial, which tested the effectiveness of a targeted epidermal growth factor receptor inhibitor in patients with advanced bronchioloalveolar carcinoma, was active and recruiting patients.

"In a large way, it helped define how I became identified and subsequently defined my area of expertise for the next decade," says West, who is now a clinician at Swedish Medical Center in Seattle and remains an active participant in SWOG.

Participants spend the weeks leading up to the September workshop writing a rough draft of their proposal and completing an online component that introduces them to the intricacies of SWOG group membership and the procedures for reporting serious adverse events and quality-assurance audits. They then go to SWOG's statistical center in Seattle for the intensive 3-day workshop. "At the meeting, a lot of the nuances of the study design are discussed, and all of the components of the protocol are brought together," says Carolyn Hoban, an assistant professor in internal medicine at the University of Michigan, Ann Arbor, who oversees the SWOG protocols in translational medicine.

Modern cancer clinical trials are costly and often take many years to complete. Their design usually has a primary goal or "endpoint" that, in the case of treatment trials, should provide the information required to make an informed decision about whether a treatment is safe and effective for patients. But in an era of targeted cancer therapy, treatments that help some patients may not benefit others. Parsing data requires a sophisticated trial design that includes laboratory work to characterize patient tumors using gene arrays, proteomics, and analysis of regulatory pathways, as well as complex statistical analysis of subsets of patient outcomes.

Such work, often called "translational medicine," is considered the forefront of modern clinical trial design. Each year, Hoban teaches participants how to incorporate translational medicine approaches into their clinical trials. Participants learn how to ensure that a sufficient number of patients are recruited to ensure statistically significant results, how to choose the best endpoints to measure in the trial, how to build a list of resources, and how to meet all the regulatory requirements that accompany research in human subjects, among other topics. Because personalized therapy approaches often require collecting and maintaining tissue and blood samples for interrogation in the laboratory, participants also learn the logistics of biospecimen banking and how to ensure specimens retain their research value for years to come.

Learning how to incorporate molecular characterization of tumor samples into the study design was a particular help, West says. Coming from a clinical background, West didn't have the experience necessary to design a lab-based protocol. But during the workshop, senior SWOG investigators helped make those connections for him. "You are actually plugging into this community of scientists," Hoban says.

Training elsewhere

Beyond the SWOG course, the Methods in Clinical Cancer Research workshop, put on by the American Society of Clinical Oncology (ASCO) and the American Association for Cancer Research (AACR), is probably the best known short course designed to cover the essential elements of designing a cancer clinical trial. "I got to meet a lot of people who have continued to be mentors during my career," says Heather Wakelee, who attended the ASCO/AACR workshop in 2002.

At the course, Wakelee, an oncologist at Stanford Cancer Center in California and an affiliate of the Eastern Cooperative Oncology Group (ECOG), was introduced to representatives from the pharmaceutical companies that made the compounds she wanted to test and provided the funding to support her first trial. "I found that [at ECOG] people were very supportive of me as an eager junior investigator," she says.

At the time, her institution was developing its own immersive course in clinical trial development, which she participated in. Many institutions offer similar courses, including the Johns Hopkins Center for Clinical Trials and the University of Pennsylvania.

There are fellowships available through some professional organizations, such as one offered by the American Society of Hematology. Such fellowships give early-career researchers the opportunity to interact with leaders in their fields and to learn how to design and conduct clinical trials that incorporate a translational research component.

Whether you have attended any sort of training in clinical trials design, you should look within your institution for mentors. Wakelee sought out individuals who had conducted phase I developmental therapeutic trials; her new mentors worked with her to hone her ideas into testable hypotheses.

Through that mentoring process, "I learned to put together a question that could be asked in a clinical trial," Wakelee says. In addition, through the mentorship process, she was given opportunities to lead data-analysis portions of larger clinical trials and get her feet wet without the responsibilities of running the trial. Through these mentored opportunities, she learned all the elements that have to come together to run a clinical trial, she says.

Getting her name out there in smaller studies also helped establish her professionally. And through ECOG mentor Joan Schiller, an oncologist at the University of Texas Southwestern Medical Center at Dallas, Wakelee was able to take the lead in developing a protocol to conduct a large phase III clinical trial examining adjuvant bevacizumab for the treatment of resected early-stage lung cancer.

"Good mentorship is the backbone of academics," Wakelee says. "I've been very fortunate to find incredible mentors both at Stanford and in the larger lung cancer research community to guide me to this point in my career and am thrilled to now be able to act as a mentor for current oncology fellows. It is that continuum of learning and teaching which is part of what makes academic medicine such a wonderful career."

Oncologist David Solit, 41, has some close professional role models: His father was a surgeon and his grandfather a family practitioner. Like many doctors who pursue oncology, he became interested in the disease after a relative died from breast cancer. But it was a laboratory rotation during his oncology fellowship that sealed his interest in cancer research.

"My interest was not to stay in the clinic and try to use the drugs that we had, which, in my opinion, were not very good," Solit says. "I thought it would be best to stay in the lab and to try to actually develop some better treatments that we could bring into the clinic."

Now, Solit holds the Elizabeth and Felix Rohatyn Chair for Junior Faculty and heads his own laboratory in the Human Oncology and Pathogenesis Program at Memorial Sloan-Kettering Cancer Center in New York City. His lab studies a particular signaling pathway, the RAS/RAF/MEK/ERK pathway, which regulates cell growth and survival in several cancers. "We try to identify the underlying genetic basis of different tumor types and then develop novel therapies that will exploit the specific mutations that drive tumorigenesis or cancer progression," he says. Solit is the author of an upcoming Perspective in Science Signaling on MEK resistance, which will be published on 29 March.

Special Feature: Cancer Crusade at 40

This week, Science and its sister publications take a look at where cancer research stands 40 years after the signing of the National Cancer Act. This article is one of two in Science Careers on the topic; see also " Conducting Cancer Clinical Trials." And for more on molecularly targeted cancer drugs, see the news story " Combining Targeted Drugs to Stop Resistant Tumors" (free full text with registration) in this week's issue of Science.

See the list of cancer-related articles in all the Science publications at www.sciencemag.org/special/cancer2011/.

Solit spoke with Science Careers earlier this month about his research and his career path. The following highlights from the interview were edited for brevity and clarity. A full transcript of the conversation is available on CTSciNet.

D.S.: We're very focused on the RAS-BRAF pathway and the AKT pathway. These are two pathways that are very commonly mutated in human tumors, and we try to understand where those pathways are mutated, which type of tumors, what are the mutations that co-occur with mutations in those pathways; and try to understand, if we try to inhibit those pathways, what can we expect in patients. So, can we expect that, if we inhibit, for example, the BRAF kinase, are we going to see the cells stop growing? Are they going to die or are they going to not care at all, depending upon the underlying genetics of a particular tumor type?

D.S.: Exactly, in part. But we also take it a step further. We're very interested in how to actually drug those targets. You can't presume, for example, simply because somebody told you a particular compound is a BRAF inhibitor, that it really just inhibits BRAF or that's how it, in fact, works. So we try to have models where we really understand the genetic basis for the cancer, and then we try to take compounds and figure out, are they going to be able to work in that genetic background?

D.S.: Mostly this is in the lab using an assortment of things. We usually start with cancer cell lines, which we both generate here and obtain from other people. ... We essentially start with those cell lines, try to identify patterns between the mutational status of the cell line and the response to a particular drug, either in terms of the ability of the drug to inhibit a pathway. or to induce cell death. or inhibit growth. We then usually move onto xenograft models and then also, if available, try to test some of these compounds in genetically engineered mouse models that have particular mutations driving tumor formation.

And then the ultimate goal, as you mentioned, was to try to ultimately bring this into the clinic. I've run certain clinical trials but mostly at this point partner with some of my clinical colleagues to test the hypotheses generated in the lab in the clinic, in actual patients, and then actually try to analyze tumors from those patients to see whether the patterns that we identified in the laboratory in fact hold true in patients.

D.S.: I see patients as well. I'm a medical oncologist, so we typically see patients whose cancers have recurred and have spread to other parts of their body. We use chemotherapy or targeted therapies or immunotherapies to try to slow down or shrink down the cancer. For most of the solid tumors that we work with, unfortunately, once the cancer has spread to a distant place, we're at this point unable to cure those patients, although we can oftentimes improve their quality of life or make them live longer. So we obviously have a long way to go to develop effective treatments for most of the cancers we work with.

D.S.: Definitely indirectly -- it puts in perspective an understanding of what type of problems we really should be going after. So, for example, I'm very interested in targeting the pathways that are found in patients whose cancers recur. You can imagine that, if there was a particular mutation but everyone with that mutation was cured by surgery and nobody ever recurred, that, to me, wouldn't be something I would want to spend a whole lot of time on.

So I'm very focused on trying to figure out which of the mutations ... are in the patients whose cancers come back after they get their surgery or initial treatment with radiation, for example, because those are the ones that we are in greatest need for developing new therapies for.

D.S.: It's very difficult to do laboratory research during your clinical training because you typically work, like, 60 to 80 or sometimes more hours a week back then. So it would be very difficult to do any sort of laboratory based research while you're actually doing your internship or residency. ...

In the ... fellowship program in oncology that I did, which was at Memorial Sloan-Kettering, after your first year you have a choice to spend the next 2 years doing clinical research, participating in clinical trials or other clinical aspects of research, or you can go into the laboratory. And at that point I chose to go into the laboratory.

That could have just been for 2 years, but when I went into the laboratory, I really enjoyed the science. My interest was not to stay in the clinic and try to use the drugs that we had, which, in my opinion, were not very good. I thought it would be best to stay in the lab and to try to actually develop some better treatments that we could bring into the clinic. So I stayed not just in the lab for those 2 years but essentially did a postdoctoral fellowship beyond that for another several years even though I had finished my clinical training.

D.S.: Yeah. Without question, I think the institution, at least in this case, did a great job of providing me with a lot of protected time. I would say I was about 30% clinical and I was 70% laboratory. And that's not an uncommon balance for someone in that position. I would see patients one day a week.

D.S.: It's actually not that different. ... I would still say it's about 30% clinical, 70% laboratory.

D.S.: I think without my mentors I would not have been successful. ... I had both a great laboratory mentor and a clinical mentor. And I think that, for someone who tries to do both, that's really important. ... My laboratory mentor was Neal Rosen, and he was very supportive of my career, and he gave me a great environment in which to do laboratory work. But I would say equally as important, I had a great clinical mentor in Dr. Howard Scher, who is head of the genitourinary oncology service at Sloan-Kettering. He's an expert in prostate cancer. [He] made sure that I had adequate protected time to do the laboratory research [and] helped me in terms of my career, in terms of trying to get promoted over time.

D.S.: I think it's both the science and the clinical side. I got interested in cancer because my aunt had breast cancer and, unfortunately, passed away from breast cancer. So that had always had me interested in pursuing this in medicine in particular. ...

In terms of the science side, I think it's an exciting time to be in cancer research. ... The projects that are ongoing, like the Human Cancer Genome Project, [have] really opened up a lot of possibilities to understand the molecular basis of cancer. Right now, we've got the tumor Cancer Genome Atlas that we're part of here at Sloan-Kettering. We are contributing samples actively to this project. This is a project to repeat the Human Genome Project thousands of times using tumor samples instead of normal DNA and really identify what is the full complement of mutational changes or epigenetic changes that actually cause the cancers to develop and progress. ...

So, when these projects are being completed, it really leaves us with just a list of mutational changes that are found in the tumors, but it doesn't really inform us as to which of those are most important or how they cooperate together. So there's a huge amount of opportunity to try to sort through those questions in the lab. And what's exciting to me is that you can directly potentially use that information to impact and improve the care of patients with cancer.

D.S.: I just would say if you're interested in this career path, I would just do it. I think that, like many people, I had many people along the way who probably told me that it's just too hard to pursue this type of career. It's very hard to get your own lab or it's very hard to get this position. But I think persistence is the key in large part. I think you have to be intelligent. You have to be hard working. But I think persistence is really what separates many people who succeed from those who do not.

And there are definitely disappointments that come up in this career path. There's always going to be grants that you don't get and papers that get rejected. Without question, I would say even those who go on to win the Nobel Prize or make huge advances had grants that were rejected and papers that were rejected. But, if you're persistent and you're committed, it's not a guarantee, but there's a good chance that you could achieve what you're interested in or what your goals are.