PanXome is software technology company with a platform which enables a broader spectrum of drug discovery than conventional computer-aided drug discovery tools.  The platform is grounded in a unique approach to building predictive models of druggable protein targets, based on homology and ligand similarity.  The platform is computationally efficient and currently covers 97% of the human exome, which opens a heretofore inaccessible region of target exploration and the ability to predict off-target interactions.

TARGET Discovery and ID

Screen Shot 2017-08-25 at 3.09.20 PM.png


  • Identification of protein targets responsible for disease mechanism of action

  • Deconvolution of targets responsible for both on-target and off-target effects

  • Method includes computational inference and forward chemical genomics

  • Precomputed database of predictive binding models for 97% of the human exome


  • Average hit rate of 24% in top 100 predictions

  • 24X better than competing algorithms

  • No requirement for experimental structures

  • Higher quality compounds moving through hit-finding to lead-identification

  • Screening covers 97% of human proteome

  • Uses distant evolutionarily relationships and local pocket similarity to identify ligands

  • Does not require any known ligand binding information to achieve comparable hit rates

  • Can screen 22 million compounds/per hour on a single CPU Core




  • Probability of passing a Phase 1 safety trial with 80% precision and recall

  • Prioritized list of side effects ranked by severity

  • List of proteins causing the side effects

  • Design away from the severe side effects based on the off-target protein pocket properties

  • Applicable to all small molecules including natural products


  • Repurpose drugs that are on the market  to treat new diseases

  • Repurpose drugs that have failed efficacy to target new proteins and diseases to rescue the asset

  • Identify new target proteins associated with diseases with 30% precision