A good grasp of condensed matter theory may not seem to be the best preparation for a career in biotech, but the founders of startup Gene Network Sciences are hoping theoretical physics will be the key to their success.
Co-founders Colin Hill and Iya Khalil met as physics graduate students at Cornell University, where Hill was analyzing simple gene networks. The two developed an interest in applying computational modeling to complex biological systems, drew up a business plan, and launched GNS with the goal of combining computation with quantitative experiments to achieve in silico drug discovery.
The approach that physics takes to solving scientific problems is driving the way we do things, said Khalil, vice president of research and development at GNS. Its not just mathematical conjectures or going in the lab and blindly poking away, but its about coming up with theories that can be tested experimentally and vice versa.
Hill, CEO and president of GNS, noted that a number of key concepts from physics, such as chaos theory and dynamical systems theory, are important in describing the enzyme kinetics that control cellular behavior. Acknowledging the companys bias towards theoretical physics, Hill said that since its launch in August 2000, the companys 15 employees now hail from the fields of mathematics, genetics, biology, chemistry, engineering, linguistics, and computer science.
Once it began accumulating the data necessary to simulate a living cell, Hill said the GNS team was disappointed with the quality of available data mining tools and set out to create its own. The result, BioMine, is based on the work of GNS director of systems biology Vipul Periwal.
BioMine gives researchers the option of using 14 clustering algorithms to group similar experiments and genes, or of applying BiCluster, a proprietary one-step process that goes directly from normalization to correlation of genes and samples. BioMine also offers five statistical tools to provide validation and determine the P-value or probability of relationships.
It was never our plan to sell software, said Hill, but recognizing the demand for better gene expression analysis tools, GNS decided to make BioMine commercially available in order to support the longer-term development of its Digital Cell simulation platform.
A trial version of the software is available from the companys website (www.gnsbiotech.com), and Hill said that a pharmaceutical company and two biotechs are beta testing the software in addition to the 30 or so who have downloaded the trial.
GNS has several other offerings in the pipeline: Diagrammatic Cell Language (DAC), a diagrammatic language for describing cellular interactions, large-scale network visualization, and simulation; and Cell Object Language, a modular programming language for modeling and analyzing large-scale biological systems.
In addition, GNS is readying its Digital Cell simulation platform and Digital Disease Models for simulating networks and pathways controlling the onset of disease. The first GNS disease model is colon cancer.
In addition to the computational side of the equation, GNS is addressing the experimental side of systems biology. The company launched its own wet lab around two months ago and is supporting a team of biologists to mine the literature for additional data to feed into its model.