A Collaborative Cloud-Based Biological Modeling Platform

Here’s a little blast from the past. As the VP of Biology at a startup funded by two West coast venture capital firms, Gordon was involved in the development of a collaborative, cloud-based biological modeling platform, designed to allow biologists to create and share dynamic, stochastic, agent-based models of complex biological systems like cell-signaling pathways. Alas, the development of this platform was ultimately derailed by the financial crash of 2008, and it is no longer available - but it was a really fun and exciting project that was definitely ahead of its time.

The platform was based upon Kappa, a domain-specific language that was designed for modeling biological systems. In addition to the creation of dynamic models, the semantic formalism in the underlying language allowed biologists to do some very powerful static analysis of the systems that they modeled. For example, our agent-based model of the Epidermal Growth Factor Receptor (EGFR) pathway (shown in the video) was capable of generating some 10**30 possible states, and even questions like “Is state x possible?” are decidedly non-trivial in such a model. The failure of even the most extensive rounds of simulation to generate a particular state of interest, could easily be attributed to the fact that the simulation has yet to sample the vast, potential state space of the model enough to have a reasonable probability of observing that state - or it could equally be possible that the set of rules upon which the model is based, are collectively unable to produce that state.

One of the strengths of this platform therefore, was that in addition to being able to model the dynamic behavior of the assembled system, it also allowed for the kind of powerful static analysis that can provide analytical answers to difficult questions of this sort.

Everything you see in the brief demo video below, is running live in a web browser. We hope you enjoy it.