By John Russell

May 12, 2005 | While much of the systems biology community is focused on human physiology and therapeutics, there is also a tremendous opportunity in the microbial world, where many paying customers would dearly love to use in silico simulation to help build better “bugs” for bio-processing, chemical manufacturing, and anti-microbial research. 

Genomatica is one of a few systems biology companies that focuses squarely on the microbial world, though its long-term aspirations are even broader. It is the brainchild of systems biology pioneer Bernhard Palsson, UC San Diego, and one of his former PhD students, Christophe Schilling, who is now Genomatica’s president and CSO. Palsson has been researching microbial metabolism for years, and his paper, “Thirteen Years of Building Constraint-Based In Silico Models of Escherichia coli” (Journal of Bacteriology, May 2003), provides a good overview of the approach underlying Genomatica.

Incorporated in late 1998 with operations commencing in July of 2000, the San Diego-based company raised a modest $3.5 million in the early years. Its headcount grew steadily to about 20 today, and the company has been cash flow neutral for the past year or so, according to Schilling. “We didn’t feel that the business models around systems biology were mature enough or the technology [was] mature enough to support tens of millions of dollars of investment. I think you can still say that today about systems biology,” says Schilling.

As the title of Palsson’s paper suggests, Genomatica is a predictive model-maker. Its first two models were of E. coli and Saccharomyces cerevisiae and derived largely from work in Palsson’s lab. SimPheny (for simulated phenotype) is the company’s modeling platform, and Genomatica’s business model spans software provider to research collaborator.

“Companies [often] have their own proprietary organisms, so depending on the size of the bug and the amount of information available, we’ll work from 9 months to 18 months to build the model of how that bug works. We’ll then go to the next stage of trying to implement the model to drive process improvement. So far, we’ve generated handfuls of predictions to improves processes and are now in collaboration to improve them,” Schilling says.

An ongoing 2.5-year effort with Dow is the longest of four commercial collaborations. Specific projects, cited by Schilling, include models to improve production of a commodity chemical, amino acids, antibiotics, and proteins. Genomatica has built roughly a dozen models covering a range of prokaryote and eukaryote organisms.

The so-called constraint-based model incorporates many hard and soft limits such as reaction stoichiometry, thermodynamics, enzyme kinetics and genomic and proteomic data. “It’s really a comprehensive description of metabolism in a genetic and biochemically consistent manner,” says Schilling. Researchers are able, in silico, to delete genes, vary biochemical parameters, and perturb the environment to generate hypotheses. Iterative experimental results — done by collaborators since Genomatica has little wet lab capability — refine the model.

Hardest to get right, says Schilling, are gene regulatory elements, but “exhaustive characterization of regulatory systems is probably not necessary.”

The IT technology used is vanilla — primarily Linux-based systems. The initial development team was “built with people from the software development industry and [we] decided we’d teach them all the biology they needed to know.” The company also has scientists organized around application areas such as chemical production and anti-microbials.

“It’s interesting because one of the things we think about is whether or not the actual process of maintaining a model, say curating an E. coli model on an ongoing basis, is more of an engineering project than a science R&D project,” notes Schilling.
The company’s scientific advisory board is impressive — including, for instance, George Church (Harvard) and Lee Hood (Institute for Systems Biology) — and it has several flourishing academic collaborations. Indeed, Schilling believes there is a large community of under-resourced academic investigators willing to collaborate with the company.

Proof of success is still a year or two off, but so far, characterizing the lives of bugs is turning into a good life for Genomatica.