Advertisement

If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.

ENJOY UNLIMITED ACCES TO C&EN

Careers

From Lab Bench to Laptop

Organic and medicinal chemists find a niche in computer-aided drug discovery

by BETHANY HALFORD, C&EN WASHINGTON
April 18, 2005 | A version of this story appeared in Volume 83, Issue 16

Shanmugasundaram
[+]Enlarge
Credit: PFIZER PHOTO
Credit: PFIZER PHOTO

Most chemists know that organic solvents and computers don't mix. Likewise, medicinal and synthetic organic chemists working in drug discovery generally tend to toil at the bench, leaving the virtual work of computer-aided drug design (CADD) to the theoreticians.

"Organic chemistry provides a well-rounded background for people to do a number of things," remarks J. Phillip Bowen, a chemistry professor and director of the Center for Drug Design at the University of North Carolina, Greensboro. As one of the first organic chemistry professors to devote his research and teaching careers to CADD, Bowen would know.

The students in Bowen's lab range from theorists doing CADD to organic chemists pursuing purely synthetic projects. All students get a grounding in both practical organic chemistry and computational drug design, Bowen says.

According to Bowen, the success of multidisciplinary labs like his indicates a change in chemists' attitudes toward the combination of these areas. In the mid-1980s, when Bowen--with a Ph.D. in organic synthesis and a postdoc in computational chemistry--was looking for his first academic job doing CADD, he found that jobs in chemistry departments were few and far between. In fact, his first academic job was in a school of pharmacy, where there is considerable interest in drug discovery.

That's all changed now, Bowen says. "Essentially, every pharmaceutical company that's involved in drug discovery has CADD scientists, and chemistry departments are focusing more on drug design." And as more structural data about biologically active sites become available, he thinks opportunities in CADD will continue to grow.

John H. Van Drie, director of CADD at the Novartis Institutes for BioMedical Research in Cambridge, Mass., concurs. CADD has been expanding steadily over the past 15 years, he says. He thinks organic and medicinal chemists will continue to find job opportunities in this area, although he's quick to point out that the positions are mostly for chemists with Ph.D.s in computational chemistry.

"In general, if you look around the industry, maybe 10% of CADD scientists come from an organic or medicinal background," he says. But he adds that those few chemists who can make that career switch find that their unique perspective is valued.

Case in point: When Veer Shanmugasundaram decided to leave India and study in the U.S., he had to decide between two different careers. With a bachelor's degree in electronics engineering and a master's degree in chemistry, he could have pursued a career in the semiconductor industry or the pharmaceutical industry.

He chose the pharmaceutical industry and entered a medicinal chemistry Ph.D. program. The students in his group were really "jacks of many trades," Shanmugasundaram says. "We were required to do multiple things as part of our thesis: design, synthesis, and even biological assays."

Even so, Shanmugasundaram admits that he found himself increasingly attracted to computational chemistry, auditing classes in the subject when he had the opportunity. Then he saw Van Drie give a presentation about CADD. "I liked this method of doing drug discovery on computers," he recalls. But without a Ph.D. in computational chemistry, he wondered if he'd ever have the opportunity to work in the field.

Van Drie
[+]Enlarge
Credit: NOVARTIS PHOTO
Credit: NOVARTIS PHOTO

SEEKING ADVICE, Shanmugasundaram contacted Van Drie, who encouraged the young scientist to pursue an industrial postdoc in cheminformatics at Pharmacia, where Van Drie was working at the time. The postdoc led to a permanent position.

Pharmacia has since been taken over by Pfizer, but Shanmugasundaram is still working as a CADD modeler at Pfizer's Ann Arbor, Mich., facility. He finds that his multifaceted training makes him an excellent team player, able to interact with theorists, bench chemists, and biologists. "With an interdisciplinary background, you understand what they are thinking, and you can speak their language," he says.

For many chemists, the move from wet chemistry to CADD is a gradual one. "You don't wake up and say, 'I'm going to change careers today,' " jokes Kent D. Stewart, an associate research fellow at Abbott.

Stewart started dabbling in computational chemistry when he was an assistant professor at Atlanta's Emory University. Moving from academia to the pharmaceutical industry, Stewart spent about seven years gradually increasing the amount of computational work in his job description until he moved to Abbott to take a purely computational position. That was 13 years ago. His work at the firm includes de novo drug design, drug-docking experiments, and modeling both his own ideas for potential drug candidates as well as proposed targets that come from other medicinal chemists in the company.

Many older computational chemists also started out as bench chemists, Stewart points out, cottoning to computer modeling as the programs gained acceptance for pharmaceutical development.

Now, though, Stewart agrees with Van Drie when he says it's uncommon for a drug company to hire a CADD scientist who doesn't have a Ph.D. in computational chemistry.

Hickey
[+]Enlarge
Credit: BOEHRINGER INGELHEIM PHOTO
Credit: BOEHRINGER INGELHEIM PHOTO

Starting out as a bench chemist "is not the most common way to approach computational chemistry," Stewart says, "but at least from Abbott's perspective, it's a necessary niche." Stewart adds that the combination has worked out well for him, although he credits Abbott for giving him "ramp-up time" in his new position. "When you switch careers like this, it is a difficult thing to do," Stewart explains. "You're changing your graduate training, and for a while you're going to be less productive than you were in your old job."

Eugene Hickey managed to switch from a medicinal chemist's position to one in CADD without even changing his company parking spot. As a principal scientist at Boehringer Ingelheim, Hickey's research includes molecular modeling, de novo design, virtual screening, docking, and library design, even though he originally joined the pharmaceutical company as a medicinal chemist.

"Part of my thesis had computational aspects to it," Hickey notes, but his doctoral research was focused on organic synthesis. Any computational training he gleaned was basically self-taught.

While working in industry as a medicinal chemist, Hickey says he often employed modeling to design new compounds and decide which leads to pursue. On one project that he's particularly proud of, Hickey used computational methods to help prioritize structure-activity relationships. The compound that his group eventually made from that work went on to the clinic, and his representation of its crystal structure was used as the Journal of Medicinal Chemistry's cover for the first half of 2003.

Hickey's facility with CADD didn't go unnoticed at Boehringer Ingelheim. Eventually, he had the opportunity to join a computational group. "I really wrestled with it," he says. He liked the idea of addressing multiple projects--as a medicinal chemist he typically worked on one project at a time. He would have to give up managing his medicinal group, though.

Hickey also didn't think it would be fair to his computational coworkers for him to join the group without any formal computational training. He took a weeklong training course. "The problem is that it's only a week, and you need a lot more training than that," he points out, adding that his move was made much easier by his CADD colleagues, who took the time to teach him new approaches and software. "I think the transition would have been very difficult without their help."

Stewart
[+]Enlarge
Credit: ABBOTT LABORATORIES PHOTO
Credit: ABBOTT LABORATORIES PHOTO

AT FIRST, not having a formal computational chemistry education was an obstacle for Hickey, but certainly not an insurmountable one. And he's been able to use his organic chemistry background to sell ideas from the CADD group to chemists who are still working at the bench.

Although he's pleased with his career switch, noting that it's given him an entirely different perspective on drug design, he says he misses making compounds sometimes. "It's when you have that one crazy idea that no one wants to work on," he explains.

Pharmaceutical companies aren't the only place where organic and medicinal chemists find work in CADD. Because many of their products are targeted at bench chemists, companies that make computational chemistry software are always on the lookout for chemists who can speak about functional-group modifications with the same confidence that they use to discuss programming languages.

The software industry proved to be a good fit for Leah Frye. With a Ph.D. in organic synthesis, a postdoc in pharmacology, and stints in academia and the pharmaceutical industry, Frye moved with her family to Oregon with no particular job prospects lined up.

While she attended a regional American Chemical Society event in Portland, her diverse background and astute questions caught the attention of Shi-Yi Liu, vice president of marketing for Schrödinger. At the time, the company wasn't looking to hire, but Liu thought Frye's varied experience would be a tremendous asset to the growing chemistry software company. Liu decided then and there to recruit Frye for the company's Portland office.

Frye has worn a number of different hats during her time at Schrödinger. She was instrumental in the software's first proof-of-concept studies, and these days, part of her job is to help other firms perform computational studies using Schrödinger's software. "We can do really large-scale docking experiments," she says. "Or, if there's a small company that doesn't have a screening section, we can give them a start."

Frye thinks she has less computational background than all of her colleagues at Schrödinger. But she says it's never been a hindrance for her--if anything, it's been a benefit. "Computational chemists and medicinal chemists don't necessarily speak the same language," she explains.

Understanding how computational chemistry software works lets her put computational studies in perspective. "Every medicinal chemist has more ideas for potential drugs than they could ever pursue." Frye points out. "So how do you decide what you're going to make? Computational methods can really help you prioritize what molecules will be your best drug candidates."

Clark
[+]Enlarge
Credit: TRIPOS PHOTO
Credit: TRIPOS PHOTO

Likewise, knowing organic chemistry helps her recognize which computational results are practical from a bench chemist's perspective. "You have to know what you can actually make," Frye says. "You have to know if it's a reasonable molecule, and you have to know if it will make a good drug or not."

Frye reckons that as pharmaceutical companies learn more about the structure of active-site targets, the industry is going to want more people who can do computational chemistry but who can also understand organic and medicinal chemistry.

Along the same lines, she believes that computational chemistry will become an everyday tool for medicinal chemists working at the bench. "In the not too distant future, I see computational methods often being used before people make molecules," Frye says. She likens it to creating a new automobile line: The car gets designed in a computer program long before anyone ever picks up a tool to build it.

Some chemists manage to gain expertise in medicinal chemistry and CADD during the course of their doctoral training. Jeremy R. Greenwood, Frye's colleague at Schrödinger, caught the CADD bug about halfway through his Ph.D. work in synthetic medicinal chemistry. "I became more and more interested in why things worked and didn't work," he explains.

Greenwood had a little experience with quantum chemistry, so he took on an extra thesis adviser and made computational work part of his dissertation.

Although he had less time at the bench than did some of his colleagues who have switched over from organic synthesis, Greenwood says his wet lab experience has still been very valuable. When a CADD program suggests a series of molecules that would fit well within an active site, he can still say what is a reasonable target and what just won't work in a reaction flask. "There's no point in making a prediction of something that's impossible to make," he says. "You'll just get laughed at."

Greenwood also thinks that having some computational know-how is helpful to bench chemists who want to use the technology to explore their own ideas. A medicinal chemist could explore the relationship between an active site and a potential drug, or a synthetic chemist could use spectral prediction programs to figure out what he or she has made.

Greenwood believes that having a background in two such different areas is unusual but also highly marketable. "It's a real niche area, but it's very good," he says. Still, he adds that it's not a career that suits everyone, and he doesn't expect to see medicinal chemists stampeding to become full-time CADD scientists. "Obviously, if people decide that synthesis is too hard and just want to play with computers, then nothing would ever get made," he quips.

Advertisement

Chris Van Dyke tells people that he has the best job at Tripos. As director of scientific training at the chemistry and drug discovery software company, Van Dyke takes his computational and medicinal chemistry know-how and uses it to train Tripos' clients to use the company's software.

Frye
[+]Enlarge
Credit: SCHRÖDINGER PHOTO
Credit: SCHRÖDINGER PHOTO

Van Dyke did his doctoral research with Bowen, but he initially did enzymology research. "When I first started my graduate work, it seemed more like a prison term than a degree program," he recalls. He took a hiatus from graduate school to work for the Peace Corps in Gabon. Upon returning to graduate school, he found his old lab shuttered, and after taking a few classes, he discovered medicinal chemistry. "It was almost as if this was where I was supposed to be from the beginning," he says.

After a couple of years in technical support at Tripos, he joined the scientific training team. To do the job effectively, Van Dyke draws upon his teaching experience--his Peace Corps duties included teaching 60 eighth graders--as well as his work as a bench scientist and a computational chemist. If anything, Van Dyke wishes his background were more varied. "To be successful as a trainer, you have to understand your target audience," he explains. "If I'm teaching medicinal chemists, then I need to understand what medicinal chemists do on a daily basis."

Bob Clark didn't seek out computational chemistry; rather, he was wooed into it. Although his graduate degrees are in bioenergetics, after working as a biochemist at Monsanto in St. Louis for a few years, he switched to doing herbicide and fungicide synthesis within the agroscience giant. "I got tired of telling the synthetic chemists what to make and them not making it. So I became one," he jokes. "I really enjoyed doing synthesis, although my wife wasn't thrilled with me coming home smelling of pyridine and thiol."

Six years later, Clark found himself being courted by Tripos. Clark admits that he was surprised that a software company even tried to recruit him. "The only programming experience I had was with Fortran in the late 1970s," he says. But the folks at Tripos were welcoming, Clark adds, so he traded in his smelly solvents for a mouse and a monitor.

These days, Clark is the company's senior director of research. His job is to develop tools that help other synthetic chemists be more effective. To do that, Clark draws upon his years at the bench as well as his computational expertise. "Typically, with the kinds of programs that we write, the only way to address bugs is to have [a programmer] who has an idea of what the program is supposed to accomplish and at the same time has some grasp of how the computer is doing its job."

Clark has only the highest praise for his classically trained computational and molecular modeling colleagues, but he thinks there's a real need for more chemists with wet lab experience in the computational chemistry realm. "There aren't enough of us," he says, particularly in software development. In fact, Clark notes that for the past few years, he's been told he needs to hire two more chemists just like himself.

So what's a bench chemist looking to become a CADD connoisseur to do? All the chemists with whom C&EN spoke said that finding a mentor was key. "It's a big cultural change, and you're going to have to work hard to make it happen. But usually, you're also going to need somebody on the inside to make that happen," Clark offered.

CHEMISTS WHO are still in graduate school should try to incorporate computational elements into their thesis projects and seek out computational postdocs. Pfizer's Shanmugasundaram says his company's internship programs are a great way for students to find out if they're suited to CADD work.

For chemists who are already gainfully employed in the pharmaceutical industry, finding the right opportunities in computational chemistry can be trickier, but Novartis' Van Drie says that's always the case when making a major career transition. "The main thing I would suggest is to simply start doing modeling on the side," he says. Develop a good rapport with the computational chemists at your company and express an interest in the field, Schrödinger's Frye adds.

More formal educational opportunities are available for working organic and medicinal chemists who want to learn more about CADD. ACS annually offers short courses in the subject, for example.

Tripos' Clark advises getting some sort of credentials, either through education, a specific project, or as part of an official job title. "More chemists are getting involved in modeling at some level," he notes, "but until it's been in your job title at some point, it can't help your career. You need those credentials."

Still, even though the transition may not have been easy, all the chemists who spoke with C&EN find their CADD work fulfilling. "It's nice not to smell of pyridine at the end of the day, but I also think I've contributed something back to the people at the bench, and that's important to me," Clark says proudly. "I went to Monsanto hoping to feed the world. I feel that I'm still doing that. I'm at a greater remove, but I also have a broader impact."

 

Article:

This article has been sent to the following recipient:

0 /1 FREE ARTICLES LEFT THIS MONTH Remaining
Chemistry matters. Join us to get the news you need.