Traditional drug discovery processes often take several years and require millions of dollars to generate a robust lead-optimization stage program. Numerate’s drug discovery process rarely takes more than nine months to get to this stage.

Numerate builds virtual assays for each of the drug design goals to predict the most important molecular properties. These virtual assays are then applied to very large virtual libraries.

Our virtual assays are lab-equivalent, meaning the accuracy of our predictions (50-80%) is similar to many laboratory assays and sufficient for any virtual design-test loop to analyze compounds.

Leveraging our expertise in medicinal chemistry, we define very large and diverse libraries (>10 billion compounds) relevant to each program. These libraries (virtual chemical spaces) are combinatorially encoded, comprising thousands of scaffolds and thousands of pendant substitution groups.

Finally, we apply the panel of virtual assays (activity, ADME, and toxicity) to analyze these large virtual libraries. This process takes about one month (on 1,000 computers) to identify a small set of high-probability compounds (structures of compounds highly likely to satisfy the design criteria).