INTERVIEW

Andrej Sali

Interviewed by Melissa Mertl

Posted October 12, 2001 · Issue 112

Background

Biography
Andrej Sali received his B.Sc. in chemistry from the University of Ljubljana, Slovenia, in 1987. After receiving his Ph.D. in biophysics in 1991 at the University of London, he then went to the Department of Chemistry at Harvard University as a Jane Coffin Childs Memorial Fund postdoctoral fellow in the lab of Martin Karplus, where he continued to develop comparative modeling methods and also studied models of protein folding. Since 1995, Sali has been at the Rockefeller University, New York City, where he is now an associate professor. He focuses on developing and applying methods for protein structure modeling, primarily in the area of comparative modeling.

What event led you into research?

It happened very early, by itself, as far as I can remember. I asked my father for a chemistry kit when I was eight - it was in a very colorful box - and I started playing with it.

Who has most inspired and/or influenced your work?

My father, and all of my mentors - Professors Igor Kregar, Francek Gubensek, Vito Turk, Tom Blundell, and Martin Karplus.

Who awarded you your first grant and what was it for?

It was a scholarship for undergraduate studies in chemistry from Institute Jozef Stefan in Ljubljana, Slovenia. Only a few of those were given each year.

What was your best experiment?

It was a computer experiment almost ten years ago - a simulation of how a simple model of a protein finds its native state if certain conditions are met. I liked it because the picture that emerged was simple. It connected thermodynamics with kinetics and was certainly correct for the model studied, not because there was much evidence that it applied to real proteins.

What was your most disastrous moment in the lab?

I don't remember when I was last in the lab. The last work-related disaster was loosing a manuscript file on which I worked for a day, just before the deadline. Yes, I used Microsoft Word.

Which scientific idea (yours or others') do you regreat the most?

There are no regrettable scientific ideas. There are only bad uses of scientific ideas, generally by nonscientists.

What is the greatest unanswered scientific question?

Please let me know when you find out. What defines the greatest question? Seriously, I do not know, but it seems that the questions about how the brain works or how molecules interact with each other to form cells and then multi-cellular organisms are really big questions.

What is the biggest challenge facing structural biology today?

To get an accurate free energy function for protein structure. The free energy function depends on the structure of the protein and includes entropy as well. We have approximations for free energy function, but they're not really accurate. So we cannot really predict the structure very accurately either by comparative modeling or threading. Errors occur because we have errors in current confirmation of free energy function - which is an indicator of the native state of the protein. The native state of the protein probably corresponds to the global free energy minimum. In addition to knowing what single confirmation is best, [an accurate free energy function] could also tell you what's the dynamic and alternative states of a protein.

What are your current research interests?

The improvement and application of methods for prediction of protein three-dimensional structure, and finding new ways of using protein structure modeling for functional annotation of proteins.

For me, a very helpful core idea in what we do is that comparative modeling is just an optimization problem. So what you do is get models [with] special restraints that can come from any sources and have any shape appropriate. So that's a very general view, but we have written our software to implement this general view. It's that general view which gives us a lot of flexibility in developing our methods and improving them.

What are your scientific plans for the next 5 years?

To move from predicting and using the structure of individual proteins to predicting and using the structures of stable and transient protein complexes.

We need to do what we can to support the structural biology effort. But that's a more technical role. Scientifically, I think we need to move to the next level in the hierarchy of structure. We have to move from looking at individual proteins to the structure of complexes - like the ribosome and the nuclear pore.

The other thing will be to look at functional sites. Now that we have structure, how can we use it to better understand function? What binds better, what doesn't bind, where are the binding sites, how does the specificity change, what are better drug targets, and so on.

What are the qualities of a successful researcher?

This depends on how one defines success. By my definition, [one needs] talent, intelligence, perseverance, a love of science, and luck. In biology, as opposed to mathematics and physics, one should also have communication and political skills. . . . In mathematics, there tend to be more loners, but in biology you need to select the right group of people, you need resources, and you need to be able to convince people at your university and elsewhere that you're worth investing in.

If you could work with any scientist (historical or current), who would it be?

I dislike the "fame" aspect of the question, but I would not mind working with the scientist who solves the protein-folding problem.

Melissa Mertl is a former news editor for BioMedNet's News and Comment section. Before joining BioMedNet, she wrote for Science's Next Wave, an online publication for disgruntled postdocs.