We’ve seen in the last couple of years increased interest in building whole cell models, that is computer models that attempt to simulate a whole organism. The most recent attempt was by Karr et al, published in Cell, 2012, 150(2), 389-401 under the grand title of ‘A whole-cell computational model predicts phenotype from genotype’. Whole cell models are interesting projects to attempt for a number reasons. They allow us to discover how much we really know about an organism, they can sometimes stretch computing technology, they most certainly stretch modeling approaches (with all the pitfalls that entails), they are fun projects to do and last but not least they should give us predictive models that can be used in a whole host of practical situations ranging from biomedical to environmental..
Here I would like to summarize some of the past attempts at whole cell modeling. The field surprisingly goes all the way back at least to the 1960s when computers started becoming more widely available and thoughts about whole cell modelling go back at least to the 1930s to Sewall Wrights seminal paper on physiological networks: Wright S. 1934. Physiological and Evolutionary Theories of Dominance. Am. Nat. 68: 25–53.
While browsing in the library as an undergraduate I accidentally came across a book by Heinmets with the title ‘Concepts and models of biomathematics : simulation techniques and methods’. As as flicked through the pages I came across the stunning figure shown below (The long strip of rectangles on the right-hand side are the genes). I was quite mesmerized by this. This was a whole-cell model that incorporated a host of systems ranging from metabolism, DNA replication, protein synthesis, transport and even viral infection. Now obviously it was not possible for Heinments to fit the model as well as the recent Karr model but it was a significant piece of work nevertheless. In addition to Heinments, other authors around the same time, including Stahl (1967) – who give a review of work up to 1967 – and Weinberg in this 1970 thesis (available online) on Computer simulation of a living cell: Interdisciplinary Synergism were attempting the same thing. More recently we also have the work by the ECell team in Japan and arguably the whole cell metabolic models from the Palsson group. Earlier this year I had the good fortune to attend the 3D Virtual Cell workshop at San Diego. There is a growing interest in diverting more effort into whole-cell modeling. Whether we will quickly obtain reliable whole models is doubtful but in the long term I would say yes. In the short term however such a project would yield a number of important results. As with the Karr model, such a large project would help drive:
1. New measurement technologies
2. Give us a better understanding of what measurements should be made
3. Development of new modeling approaches (without the pitfalls)
4. Development of new computing methodologies, perhaps grounded in parallel architectures
5. Develop new standards for model validation, exchange, and data
6. New ways to visualize such complex simulations
7, And last but not least, create excitement among young scientists and hopefully laypeople (who will after all foot the bill)
In a sense it would the systems biology version of the Apollo program but perhaps on a small scale 🙂