Our platform allows all known empirical data to drive us to the best possible drug leads.

Numerate has created a new innovative drug design platform that can rapidly deliver novel leads with no need for a crystal structure and with very limited SAR data. Our approach consists of modeling the phenomena that are critical to the success of hits, leads, and candidates. We then apply these models to spaces of drug-like chemistry large enough to contain many, many potential solutions.

This comprehensive data-driven approach allows us not only to model phenomena like binding to a target, but also to model the efficacy of compounds, and their ADME properties and off-target toxicities, as well as other properties. How does our platform affect the way we run programs? Here is an example:

At the start of a typical program, we virtually assay 25 million compounds from a bespoke, focused virtual library of 1 trillion (or more) compounds, against a handful of accurate activity, selectivity and ADME models at a cost of one-one hundredth of a penny per compound, in about one week. These results are analyzed by our chemists, and the most promising series are commissioned for synthesis and testing.  The laboratory results are then fed back into the models to further improve their accuracy, and the cycle repeats. Each cycle takes less time as the search focuses in an iterative fashion on the more promising spaces of chemistry.

At the core of the platform is our machine-learning-based modeling capability. This is really our secret sauce.  What distinguishes our modeling approach is its ability to leverage all types of data (both ranged and discrete) and all sources of data (including literature references and patents), combining many kinds of experiments (cell-free binding experiments, whole-cell experiments), and putting them together into a single model which appropriately deals with the bias, noise, and complexity present in any SAR data set. These models are very fast, and, more importantly, are very general. As a result, we are able to apply them in spaces that are truly novel, patentable, and ultimately, valuable.