This article has been updated to correct a typographical error in the 14th paragraph.
Compugen this week announced that it will lead a consortium funded by the European Commission to develop a computational model of the MAP-kinase pathway.
The three-year, €3.1 million ($3.7 million) project, called SIMAP (Simulation Modeling of the MAP-kinase Pathway), marks Compugen's first involvement in biological simulation.
"This specific type of activity in systems biology — not systems biology in general, in which we do have a few activities — but this specific activity of kinetic modeling of a pathway is something new in Compugen that we wanted to get into, and this was an opportunity for us to do it," Yossi Cohen, vice president of research and discovery at Compugen, told BioInform.
The agreement signals a progression from an agreement that Compugen signed a year ago with Novartis to reconstruct transcriptional networks from gene expression data [BioInform 04-25-05]. While the company has plenty of experience integrating data from multiple experimental platforms, and the Novartis collaboration offered an opportunity to expand that capability into network inference, the move into kinetic modeling presents a new level of computational complexity, and a new set of challenges.
Nevertheless, Cohen said that Compugen is "comfortable" with the project "because we think the other activities that we have done in the past, in which we integrated exact sciences with biology and medicine, are going to give us a really good head start."
"[T]his specific activity of kinetic modeling of a pathway is something new in Compugen that we wanted to get into, and this was an opportunity for us to do it."
The SIMAP consortium includes France's Aureus Pharma, Spain's National Research Council (CSIC) and Vall d'Hebron University Hospital Research Institute, Italy's National Cancer Institute, Germany's Max Planck Institute for Infection Biology, the UK's University of Glasgow, and Israel's Weizmann Institute of Science.
Cohen said that there are three components of the project: computational, biochemical, and clinical.
Compugen will head up the computational component, which involves building the predictive kinetic model as well as using new text-mining methods to extract information on the MAP-kinase pathway from the scientific literature. Aureus Pharma and the University of Glasgow will concentrate on that aspect of the task, while Compugen will focus on the modeling.
The other partners will split the biochemical and clinical tasks, which will generate new data to feed into the model. Specifically, the CSIC, Italy's National Cancer Institute, the Max Planck Institute, and the Weizmann Institute will conduct large-scale biochemical assays such as microarray experiments and RNAi screens, to generate kinetic data related to molecules in the pathway, while the other partners will collect clinical data from cancer patients that will be used to improve and verify the computational model.
The MAP-kinase pathway is implicated in a number of cancer types and already targeted by several cancer drugs, including Irressa, Tarceva, and Erbitux. But Cohen said that much more work needs to be done to elucidate the pathway mechanisms in order to design more effective therapies and diagnostics.
Several years ago, Compugen shifted its focus from bioinformatics software sales toward the discovery of therapeutic proteins and biomarkers, but the company has continued to develop its so-called "discovery engine" predictive modeling technology for its internal research.
Likewise, the company sees the SIMAP project as a way to bolster its predictive modeling methodologies to support its in-house discovery efforts.
"The purpose is not to sell the tool. The purpose is to find and make novel discoveries that can be commercialized," Cohen said.
"Our mission is to combine exact sciences, biology, and medicine to answer unmet clinical needs, and in practical terms, we would like to find novel molecules that are royalty-bearing deals plus milestone payments based on these molecules," he added. "This pathway looked like the perfect match for us because it is well validated, it's thoroughly investigated, and the unmet medical needs are there."
Cohen declined to comment on whether Compugen plans to use the model to validate specific molecules that it has identified as potential therapeutic or biomarker candidates.
Generally, he said, the model will lead to the development of "better theranostic markers that will better predict responders and non-responders; we will identify new targets in this pathway that will be better intervention points; we are going to end up with novel surrogate markers, which are crucial for reducing the attrition rate in clinical trials; and it can also help us determine the right dosage, and the right treatment for drugs already on the market and operating on this pathway."
Other SIMAP partners also view the project as an opportunity to stretch their technology-development muscles.
Francois Petitet, director of life sciences at Aureus Pharma, told BioInform that the drug-target database provider plans to evaluate several text-mining methods under the project that it may eventually use to help automate its database curation efforts.
Aureus also views SIMAP as "a good project for us to try to expand our structure to include more molecular biology," Petitet said.
The SIMAP model will eventually be released as a software tool, but the terms of its distribution have yet to be decided.
Cohen said that the next six to nine months will be spent laying the groundwork for the complex project. Short-term goals include creating a project website, developing a standard dictionary "that covers all the reactions and known model connections," establishing a standard set of protocols for transferring biochemical and clinical data, evaluating available text-mining tools, and setting up ethical guidelines for the collection of clinical data.
Experimental data is expected to start coming online in about a year. In the meantime, Cohen said that Compugen is already building an initial model using available data.
— Bernadette Toner ([email protected])