Spotlight - Jose Vilar

Fluctuating Subject Matters -- A Switch to Biophysics

I completed my doctorate in less than three years, finishing in 1998. At the time, the mapping of the human genome was on its way to be completed and high-throughput technologies in molecular biology capable of monitoring thousands of genes simultaneously had emerged. This was an exceptional turning point. It was clear that enough information was soon going to be available to treat the whole cell as a physical system -- meaning that one could try to fully understand how cells behave in terms of the interactions of their molecular components.

Intrigued, I decided to pursue this direction, which meant that I had to move to the US, where the field was more advanced. At the time, some of the most interesting work in the burgeoning field was taking place at Princeton University, which was where I decided to do my postdoctoral research in biophysics, working in Stanislas Leibler's laboratory.

It was an exciting time. All this data was being produced and needed to be processed and put into context. If you want to fully understand biological systems just by doing experiments, it's very difficult. It is almost impossible to handle their complexity. The difficult part is figuring out what is important and which experiments need to be done. I wanted to use the techniques of physics and engineering to predict the behavior of biological systems, which are far more complex than physical nonliving systems. 

Birth of a Field

It wasn't exactly the birth of a new field in precise terms, but at the time there weren't many people trying to attach mathematical formulas to cellular processes. There were basically no books, no academic courses, no review articles, and typical biology journals would look suspiciously at any type of work that involved complex math calculations. Since my arrival in the US, the number of people interested in quantitative biology has grown tremendously. Now there are entire departments, institutes, and many specialized journals that cover the fields of systems and computational biology.

At Princeton, I studied cellular networks, looking at the way in which certain intracellular processes affect each other to get cells going. The system we chose to study was a well known DNA sequence governing lactose metabolism in the E. coli bacterium known as lac operon. This is one of the systems that in the 1950s led to the whole field of gene regulation and it is extremely well characterized. There was all this molecular information about it, including sequence and the detailed three-dimensional structures of all its components, and yet it was not possible to predict what the effect of precise mutations would be in the behavior of E. coli. We developed mathematical models that can tell you how molecules bind to DNA, how cells change their state in response to lactose, and how populations of cells grow together. More importantly, these models can reveal how everything fits together.

After three years at Princeton, I moved with the Leibler lab to The Rockefeller University in 2001. One of the reasons for my making the move was geographic. It was attractive to me, coming from Barcelona, to live in New York City. For the next two years at Rockefeller, I continued the research I had begun in Princeton.

Memorial Sloan-Kettering

When I considered my next steps, I knew I wanted to stay in New York. The newly created Computational Biology Program at Memorial Sloan-Kettering, together with an impressive group of scientists, provided the perfect environment. 

My own lab started here in 2004 and I wanted to study more human health-related systems. The system we had studied, the lac operon in E. coli, was fairly complex, but still it is one of the simplest biological systems. If you wanted to study something more complex, like human cells, you couldn't use conventional computational methods, which meant we had to develop new tools.

We want to understand how a system works quantitatively so that we can be able to control it. Take a car as an example. Stopping a car on a slippery road can be very challenging. If you brake strongly and the wheels get locked, you  can lose control of the car and it may spin. Anti-lock braking systems, known as ABS, have been designed to substitute regular braking under such slippery conditions. In the most sophisticated setups, each wheel is monitored independently, and a control mechanism computes the pressure that should be applied to slow down each wheel independently to stop the car. This type of control requires a concrete model connecting the speed of the wheels with the motion of the car.

I wanted to create computational models of cellular processes to be able to predict the behavior of the cell and to learn how to control it. For instance, it is now understood that many diseases cannot be cured with a single drug, the so-called "silver bullet" approach. They are more complex and most likely will need a combination of drugs and therapies tailored to each particular individual. If you have a patient, you can't try many pharmaceutical combinations to see which ones work, in the same way as you can't try braking in different ways when you are driving on a slippery road. Using computational and mathematical tools, we hope someday in the not-too-distant future to build and use models to test which drug combinations will work on various cancers. We are quite far from these models, but we have the resources dedicated to get there.

Integrating Approaches

Now we're developing mathematical and computational methods. We have just finished an important method connecting the molecular properties with the behavior of the cell. Whereas one group of researchers used to study the details of the molecular components, another studied how the components work with one another within the cell, we are looking to develop a technique that connects both perspectives. This method will give us an idea of the behavior of the system from the bottom up and will uncover how precise changes in the molecular details affect the properties of the cell. We want to take our mathematical methods to more complex systems.

For example, we have used these techniques to study the elastic properties of DNA inside the cell. Elasticity of DNA is remarkably important because many cellular processes require two DNA regions to come together. It turns out that the whole story is much more complicated than originally theorized because DNA can readily bend to form loops in many different ways. It was the first time that scientists were able to infer the molecular properties of DNA within a living cell. In essence, this method is based on the effects that DNA's flexibility, through many intermediate steps, has on the color of cell populations in a test tube: the more yellow you see, the easier it is for cellular components to loop DNA. The computational models we developed allowed us to zoom in with extraordinary precision and connect the intensity of yellow color with the molecular properties of looped DNA. Once we had the model, we only had to use data from experiments that were done over ten years ago! Now, we can use this information to predict the behavior of other systems that have this type of DNA.

When I was young, I took toys apart to understand how they work. Now I am trying to understand how the complex system of the cell works, using that knowledge to build a model that can be experimented with. So, from that perspective, I've returned full circle to where I began, and I'm excited to see where I go next.