KITP's newest Permanent Member, Boris Shraiman. Photo by Kevin Barron.
Boris Shraiman is the newest of the five permanent members of the Kavli Institute for Theoretical Physics, and the first to specialize in biological physics.
His conversion from statistical and condensed matter physics, with research interests ranging from pattern formation and turbulence to superconductivity, took place gradually in the early 1990s at Bell Labs in New Jersey, where he worked for 17 years after completing his 1983 PhD at Harvard and a postdoctoral fellowship at the University of Chicago. At Bell Labs, David Kleinfeld, a friend and a computational neuroscientist (now at UC San Diego), persuaded Shraiman to attend a journal club called “Brains R Us.”
“I really knew nothing about biology,” said Shraiman, who left St. Petersburg at the time it was called “Leningrad.” “All my education was in mathematics and physics. I was largely educated in the Russian system. There you go one or the other way. So I was ignorant of biology, and I really got excited that there were all these things in the world that I knew nothing about.”
Brains R Us provided Shraiman with his introduction to biology, “largely along the lines of computational neuroscience.” His appetite whetted, Shraiman went off to Woods Hole for, in effect, a crash course on neuroscience.
There he discovered what really interested him was not computational neuroscience, but molecular biology. “I got interested in the molecular mechanism of vision. How does a photon trigger a chain of events in the retina, which culminates in the firing of a neuron? How does a photon turn into an electrical signal first in the retina and then in the brain?” Engagement with those questions, said Shraiman, “set me on the slippery slope of molecular biology, though for some time I continued working in turbulence.” Eventually, biology took over.
“How is physics relevant to biology?” Shraiman muses. “There are certain things that are relevant directly. We are trained to model — to look at nature and to extract and distill some simplified quantitative approximate description. We are not trying to capture all the details, only the essential aspects so additional details can be added that won’t perturb the description."
“But I find my engagement with biology exercises a different part of the brain, so to speak, than my engagement with condensed matter physics questions. When I worked on turbulence or anti-ferromagnetic insulators, I always dealt with hard, but well-posed problems that had been formulated by someone else.”
In the case of turbulence, the equations describing fluid flow were written down by Euler, and by Navier and Stokes, and have been known for more than a century. “These people had already formulated the correct problem,” said Shraiman, “which happened to be mathematically complex; i.e., it is difficult to figure out how these equations describe observed behavior."
“The situation in biology is very, very different. When it comes to problems, we now are often pioneers. As the first ones stepping into these forests, we have to find our own way. Of course we are not the first ones in the sense that we work on phenomena that have been studied in great depth experimentally, but there are no quantitative models, no quantitative descriptions equivalent to the Navier-Stokes equations. We have to find our own way."
“In many ways even worse,” said Shraiman, “very often it is not entirely clear what question to ask, what property to understand. With materials, we know it is important to understand conductivity, magnetic susceptibility, viscosity, momentum, or heat transport. We know that there are well-defined measurable quantities that describe the properties of the material. Just exactly what is the most insightful way of quantitatively describing biological systems is a big question.”
What are the parts?
“On one level we want to know,” he said, “what are the parts. What genes and what proteins are important for a given behavior?” This “parts” approach has been, according to Shraiman, the key emphasis in molecular biology, and very often experiments answer parts questions in a binary "yes" or "no" fashion — Is a given gene important for a given phenomenon?
“But once we know what parts are important—the genes and the proteins—then, we want to ask more quantitative questions,” said Shraiman. “How do the parts interact? How does a bunch of genes and proteins behave on a systems level?”
How do parts interact?
“In phototransduction, for example, it is relatively simple to know the input and the output; the input is the photon, and the output is electrical current. You can ask what happens to the output as a function of light intensity. In other words, you can start describing phototransduction almost as a physical system.”
The next level of understanding encompasses the adaptive properties of the system. What happens, for example, when “eyes” adjust to seeing in a dark room? At first we see nothing. Then the phototransduction system adapts to a low light level. How does adaptation occur? Or, as Shraiman metaphorically asks the question, “What internal knobs do the photoreceptor cells tune in order to adjust properly their signal transduction response?”
How do systems adapt?
“Then,” said Shraiman, “there is the third layer of questions. These systems are very complex; they have lots of bits and pieces; how is it that they work with so many parts and parameters; is there some internal regulatory mechanism which adjusts them until they function properly?"
“The dogmatic response focuses on genetic hardwiring. But perhaps genes hardwire not the exact parameters of a system, but a program that adaptively adjusts these parameters till the system functions well enough."
How do networks evolve?
Finally, to understand how biological networks of systems operate requires thinking in terms of evolution — long-term. Where do the networks come from, and where are they going?
“Life, as we know it,” said Shraiman, “is a snapshot of some particular corner of the living universe at a particular time. Understanding why phototransduction, in the rod cells of vertebrates, involves a particular set of interacting proteins in a particular fashion requires a comparison of signaling systems between different cells in different contexts to identify common aspects.”
It would seem from Shraiman’s interrogatory mode of discourse that the real quest of biological physics is for interesting, insightful questions.
Shraiman identifies two types of questions:
- How does a particular system work in order to cure or treat a disease and to develop drugs, and
- What are the forces shaping networks of these systems. Much, if not most, biological research now focuses on the first type of question. The second type tries to get at general aspects of biological systems design that can only be understood in the context of evolution. Examples of this type of question are: What forces shape biological networks? Are certain designs more likely to have evolved than others?
The first type of question, Shraiman points out, is homo-centric; the second type is not.
Whither the biological physics quest for questions? “We are still very much in the dark,” said Shraiman. “The game here is to try to narrow the questions, focus on particular systems which are rich enough for making general inferences, yet specific enough for experiments. As physicists, our interests are biased towards the general and fundamental. The challenge for biological physics is to reach for those general principles, while standing firmly on the ground of biological experiments past, present, or future.”