Combining Human Intelligence with AI/ML to discover novel small molecule therapeutics

Drug Discovery: Navigating a Vast Molecular Universe
to Optimize Multiple Properties Simultaneously

With over 10e33 potential drug-like molecules and the need to balance multiple optimization parameters, successful drug discovery requires advanced data integration and innovative approaches that exceed human capabilities alone.

Drug discovery is a multiplex
optimization process

Concomitant optimization of six or more
properties is required for success

Arrepath-Home-sphere-plot-1

There are an estimated
10e33 drug-like molecules

Even within a single molecular class
(e.g. a lead program), there are 100s of
millions of molecules

Arrepath-large_chem_space-1
http://ArrePath_S4F-100x

Discovering Novel Small Molecule Therapeutics to Address Unmet Medical Needs

ArrePath’s AI/ML platform is at least 3X more efficient in identifying new compounds with progressible attributes than traditional approaches because:

  • Our ML strategy combines data across disparate assays
  • We conduct rigorous evaluation of models in real-world use cases
  • Our AI is tightly integrated into the drug discovery process at all stages

 

To date, our approach has identified:

  • Two classes of antibiotics against common bacterial pathogens, each with a distinctly novel mechanism of action
  • Multiple compound families active against NTM
  • Pathways to expand our platform into other therapeutic areas beginning with respiratory medicine
Our Solution