- 500-Gene-&-Protein In Silico Model Includes One-Third Of All Cancer Drug Targets -

ITHACA, NY - June 3, 2002 - At the Beyond Genome Conference, Gene Network Sciences (GNS) today announced that it has created the largest known data-driven computer model of a human cancer cell. The company built the predictive simulation of a colon cancer cell, which consists of more than 500 genes and proteins, using its Visual CellTM and Digital Disease ModelTM software along with experimental data points of mRNA, protein, and phosphorylated protein levels collected in the GNS wet lab and from outside collaborators. The GNS model speeds the drug discovery process by identifying high value drug targets, testing the efficacy of lead compounds, and running virtual clinical trials.

Included in the GNS colon cancer model are approximately one-third of the drug targets that pharmaceutical companies are pursuing for cancer, including BCL-2, Ras, IKK and p53.

"In the next year, we plan to reach the 5,000-gene-and-protein mark and to incorporate all known regulatory pathway information about the cell and every known drug target for cancer into our model. We are quickly moving towards recreating life on the computer," said Colin Hill, CEO and founder of GNS. "While size is a key metric for measuring a model, another is data diversity. We've created algorithms that use the massive amounts of available genomic and proteomic data to constrain the model in a statistically meaningful way."

The dynamic GNS simulation of interconnected signal transduction pathways and gene expression networks controlling human cell growth contains over 2,000 variables. The model describes the processes of endocytosis, receptor signaling, signal transduction, transcriptional control of gene expression networks, and protein translation and degradation mechanisms. It predicts various physiological outcomes such as cell cycle progression and arrest through G1-S and G2-M starting from mitogenic signaling, cell cycle arrest and apoptosis induction via p53, and the interplay between survival signals and apoptosis.

The model has already generated significant results. "We have made predictions on targets that sensitize cancer cells towards apoptosis and secondary targets that can be used in combination to lead to significant cell death in cancer cells, but not in normal cells. Our platform is ideal for finding targets and combinations of targets that provide the required efficacy and have minimal or no side effects," said Iya Khalil, GNS co-founder and vice president of R&D.

GNS attributes its success to the proprietary tools it has created to tame the complexity of biological modeling and to integrate large-scale biological data. The company utilizes its Diagrammatic Cell LanguageTM, the first complete language that describes cellular interactions, to visually represent interactions and parse them into simulation code. GNS also has proof of concept for its optimization and network inference algorithms, which infer unknown cell circuitry. The company created parallel and distributive computational architectures for running the models, and built a 192-processor Linux supercomputing cluster with IBM. In addition, GNS experimentally validates predictions from the models in its own wet lab.

"With its unprecedented level of detail and diversity of data, the GNS model helps us to better understand the molecular biology of cancer," said Dan Notterman, M.D. and chairman of pediatrics at the Robert Wood Johnson Medical School. "Colon cancer is one of the more frequent and more serious malignancies. In silico modeling in general and the GNS effort in particular hold the promise of speeding the discovery of drugs to treat this disease."

In addition to its colon cancer modeling, GNS is working on simulations of additional cancers, E. coli bacteria and Mycoplasma genitalium.

 
How it Works

GNS begins by text-mining the massive amounts of diverse molecular biology data from literature sources. This information is combined with quantitative protein data and mRNA expression data from Western blot experiments, anti-body screens, and DNA microarrays.

The model incorporates this information via Visual CellTM, a drawing toolkit that incorporates the company's proprietary Diagrammatic Cell Language. The diagram automatically parses into computer code and the dynamics of the gene expression networks and protein signaling pathways are simulated. Leveraging massive amounts of DNA microarray and proteomics data with data constraints and network inference algorithms, the cell simulation can then be used to generate predictions about the efficacy and toxicity of drug targets and lead compounds. Finally, GNS experimentally validates predictions in order to continually improve the models in an iterative process.

 
Metrics of Modeling

The metrics that GNS uses to define the size, depth and accuracy of its models are the number of:

• core biological components (gene and protein targets such as p53, Ras and c-Myc);
• chemical species (such as phosphorylated) and parameters (rate of binding, rate of degradation, etc.);
• mechanisms (transcription, translation, receptor endocytosis, etc.);
• physiological processes and cellular outcomes (division, apoptosis, etc.); and
• data types incorporated (DNA sequence, mRNA and protein concentrations, cell localization, etc.).

 
About Gene Network Sciences

Founded in August 2000, Gene Network Sciences (www.gnsbiotech.com) is a privately held biotech company headquartered in Ithaca, New York. A pioneer in the field of systems biology, GNS integrates biological and chemical data to create accurate and robust computer models of cell function and human biology. GNS helps pharmaceutical companies better understand the complex human biological systems that they seek to affect. The company's technology will ultimately increase clinical trial success rates and help bring better drugs to market faster.

Return to top