Discovery Engine

Integrating combinatorial chemistry, pico-scale activity-based screening, and AI to accelerate the discovery of small molecule therapeutics.

We get to higher value medicines faster

We get to
higher value
medicines faster

Our Discovery Engine integrates three core technologies: combinatorial libraries, pico-scale activity-based screening, and AI. Central to these technologies is novel experimental data that augments powerful AI models to predict new chemical matter. Through iterative discovery engine cycles, we simultaneously expand chemical space, enhance bioactivity, and optimize for drug-like properties to drive faster drug discovery.

Down arrow Scroll to learn more

Darwin Screening Platform In Action

Capacity to generate 1M+ new data points weekly

1859 Chip Animation

Three Core Technologies

  • Combinatorial libraries

    Combinatorial Libraries

    Our One Bead One Compound (OBOC) combinatorial libraries are built for fidelity, diversity, and pharmacological relevance. With our rapid design and synthesis workflow, we swiftly construct focused libraries with both breadth and depth of chemical space.

  • Activity-based screening

    Activity-based Screening

    Our activity-based screening system leverages micro-fluidic technology to rapidly test molecules against the target at pico-scale volumes, generating a million activity data points per week.

  • Artifical intelligence

    Artificial Intelligence

    Our Artificial Intelligence (AI) models accelerate drug discovery by rapidly searching the vast chemical space to guide our scientists to the most promising drug-like compounds. Powered by multiple data sources including proprietary data, we enhance prediction accuracy all the way through lead optimization. We deploy generative design and multi-parameter optimization to give medicinal chemists unprecedented power.

  • Data

Proprietary data powers our discovery engine

Our experimental datasets are derived from our engine’s combinatorial library screens and encompass a vast number of active and inactive datapoints, enabling comprehensive high-throughput structure-activity relationship (SAR) mapping and high validation rates.