Executable Cell Biology

From BioNetWiki

Authors: Jim Faeder (anyone is welcome to sign up here)

A paper has recently appeared in Nature Biotechnology that purports to review efforts to develop formal descriptions and models of biological systems. I have yet to read the paper in detail, but I would note immediately that it fails to cite a substantial body of work in this area that was reviewed by us last year in STKE. The authors apologize in the Acknowledgments for not being comprehensive in their review, but it is unfortunate that a large body of work seems to have been missed.

• What is hierarchical structuring?

Figure 5 presents a model in pi calculus that is typical for its obfuscation of simple interactions. Recent advances in simulation methods by Danos and co-workers and the Los Alamos group have shown that other rule-based modeling languages provide a more natural and computationally powerful framework for computational models.

Recent work by Kwiatkowska and colleagues (2006) using probabilistic model checking is of particular interest. This work will be the subject of an upcoming review.

Challenges

Main thrust of executable biology (formal biology?) is "...to shift biology toward an engineering science, where students learn to use formal approaches." I think this is a reasonable goal in many ways, but it does overlook the fact that there is still a lot of biophyics -- which is what gives rise to the rules, the ability to formalize -- that we still have left to learn. In particular, understanding the roles of spatial localization and heterogeneity, allostery, and diffusion all remain major and largely unexplored challenges in the realm of biological systems. Engineering without physics is like a language without an alphabet (DISCLAIMER: I am not a physicist).

• Developing quantitative measures to test system dynamics

It's definitely a good idea to measure things.

• Identifying 'logic gates' in biology

I think this is an elegant re-statement of a relatively old idea that a major goal of modeling is to understanding the wet circuitry of a cell. I know I have read that terminology somewhere; does anyone know of specific references?

• Biology as an engineering science

Shouldn't this be biology as a mulit-disciplinary science?

My main complaint here is really the failure to identify specific technical challenges, which may not really be a deficiency in the context in which the article is presented (A review Nature Biotechnology). I don't think the authors do a very good job of addressing the technical limitations of the work that is reviewed, and I don't think they provide any really new insights. That said, I generally agree that the introduction of formal methods and analysis in a user-friendly package is needed in biological science. So far, most molecular biologists have been relying on their own reasoning abilities and a very fuzzy roadmaps. I think with the aid of some better representational and analysis tools, a lot of new insights can be found.

References

• Kwiatkowska, M. et al. Simulation and verification for computational modeling of signaling pathways. in Proceedings of Winter Simulation Conference, Monterey, California, December 2–6, 2006, 1666–1674 (IEEE, 2006) [1]
• Schaub, M.A., Henzinger, T.A. & Fisher, J. Qualitative networks: a symbolic approach to analyze biological signaling networks. BMC Syst. Biol. 1 (2007).[2]