A Word With Jonathan Karr - Biology Week


Photo courtesy of Erik Jacobsen, Threestory Studio Inc.

Last summer, a quietly reported triumph in bioengineering was achieved when a team at Stanford University successfully completed the first computational model of an organism. Lead by Dr. Markus Covert, researchers accounted for every molecular interaction possible for a single cell organism known as the Mycoplasma genitalium. Their work takes the first negligible steps in conducting experiments outside of live laboratory environments and into the more controlled atmosphere of computer desktops.

Dr. Markus Covert talking about his initial inspiration for the project and the progress being made in the area of bioengineering thanks to his team.

Matthew DeMello, co-host of BTR’s Third Eye Weekly podcast, had the chance to speak with two of the lead authors on the study documenting the Syracuse cell mapping project, Jonathan Karr and Jayodita Sangvhi. In the first segment airing today at noon, Mr. Karr gives us more details on the cross-disciplined approach required to complete this cell model how comprehensively it can replicate all the complexities of the actual bacterium.

BreakThru Radio: To begin, tell us how you first became involved in this project.

Jonathan Karr: I was a graduate student at Stanford in 2006 and was very interested in cell biology and the intersection of cell biology and mathematics and physics. So I joined the laboratory of Marcus Covert and soon after that we got very interested in trying to uncover everything that is known about single organisms and trying to assemble that into a single computational model.

BTR: And the organism of choice for your computational model project was a human pathogen, the Mycoplasma genitalium. Your team choose this organism, I read in the literature, because it is the worlds smallest free-living organism. Can you tell us exactly what you mean by free-living and how that makes computational modelling easier.

JK: Yes that is correct. We decided to model the Mycoplasma genitalium because it has the smallest number of genes among all freely-living organisms. Which means that it lives in the absence, or it can live in the absence, of any other organism. So you can grow it in isolation from other organisms, in particular, from humans or other bacteria. That means that when you are developing a computational model like we’ve planned to do, you can model just the bacterium, and you don’t have to be concerned with the human host in addition, or any other bacteria or any other cell types. So it really simplifies the challenge in terms of what is needed to be modeled.

The model represents each major cellular function including DNA, RNA, and protein synthesis. Photo courtesy of Covert Lab, Stanford University.

BTR: Not to sound crass, but is it significant at all that this is a sexually transmitted disease? Does that particular type of pathogen make it a more ideal candidate for modelling?

JK: From a computational perspective, the fact that its a human pathogen really isn’t significant. Its still interesting from a medical perspective that we did build a model of an organism which is also a human pathogen, but the primary motivation for building the model of this particular organism, is that it has a small size and a small genome.

BTR: Going back to what you were saying before, in a video interview I was watching with Dr. Covert [see above], available on the Stanford engineering website. He’s the lead researcher behind this project. He describes biology as the science you learn if you’re not good at math, which I know is kind of like a scientist joke, but how much would say working on this project required a cross discipline approach that can only be farther developed from here in the area of bioengineering.

JK: That’s a great question. I would characterize this as really fundamentally interdisciplinary research. What Dr. Covert was referring to is that historically, biology has had very little influence on mathematics as compared to chemistry and in particular physics. But that’s been changing over the last, say, 20 to 30 years, as biological research has gotten more complex and there’s been a greater need for a biologists to analyze larger and large data sets using automated computational and mathematical methods. So our research really falls into that more modern flavor of biological research which has a very heavy emphasis on math and computation. The lead authors on the project at least have extensive backgrounds both in biology as well as in math and computation.

BTR: Now that a computational model exists for this cell, what knowledge does this allow us to pursue that wasn’t available before? What kind of simulations can we conduct now that we have this computational model mapped out?

JK: Using the encyclical model we can really simulate any potential experimental condition and we can simulate these experimental conditions with infinite precision and specificity. In contrast to a live experiment where it can be difficult to perturb the system in exactly the way you want, in the computational model, its very easy to completely specify the perturbations that you want to make. So for example experimentally, if you wanted to change the medium or delete a gene, you could do that, but the experiments are usually not quite as clean. Sometimes you’ll end up deleting more than just the gene that you wanted or a little bit less of the gene that you wanted. In sylico, you can specify exactly which part of the chromosome that you want to delete, exactly the start point and the end point so you can perform experiments with a very high degree of precision. Then in addition, because it is a computational model, we also can record the complete fate of the cell after these perturbations which is impossible to do experimentally given our current technologies. For example experimentally we can measure a few properties of individual cells for example we can look at the expression of a couple of genes at once in individual cells. But in a computational model we can look at the expression of all the genes, we can look at the computations of all the metabolites in the cell. The model really gives you the ability to do experiments with more precision than we had previously. And of course you can do them cheaply and you can do them quickly because they are all done in sylico.

Stay tuned into Third Eye Weekly on BreakThru Radio for Viral Week (starting Monday, February 4th) when we’ll be hearing from fellow lead author of the Syracuse University study on cell mapping, Jayodita Sangvhi.