Product Engine

Experimental evidence and computation in one discovery loop

We measure disordered proteins in motion, resolve the shapes they move through, and design small-molecule drug candidates against transient pockets.

Translucent protein strands drifting through many overlapping shapes at once, a disordered protein in motion rendered as a soft violet and cyan ensemble cloud on a pale background. Illustrative generated image.
Product Engine loop

Measurement, modeling, design, and validation in one Product Engine

Each stage of discovery hands the next one evidence rather than assumption. Measurement reads a disordered protein directly in solution, enhanced sampling resolves the rare states where a binding pocket opens, and generative models propose molecules against those states, with the laboratory testing every prediction.

That closed loop makes disordered targets tractable. There is no fixed structure to lock onto, so experiment and computation share one evidence layer and stay in step, from the first measurement to a candidate at the bench.

Measurement, modeling, design, and validation in one Product Engine Peptone's Product Engine circulates four capabilities around a shared evidence layer: measurement reads a disordered protein in solution, modeling resolves rare states, design proposes binders against those pockets, and validation tests every prediction at the bench, feeding the next round. Product Engine Measurement Read in solution Modeling Resolve rare states Design Design against the pocket Validation Tested at the bench Measurement, modeling, design, and validation in one Product Engine Peptone's Product Engine forms a closed loop around a shared evidence layer: measurement, modeling, design, and validation, each feeding the next. Product Engine Measurement Read in solution Modeling Resolve rare states Design Design against the pocket Validation Tested at the bench

What holds true across every program

The same commitments carry from the first measurement to the molecule we test in the laboratory.

  • Measured in solution, not forced into a crystal
  • Described as ensembles, not single structures
  • Rare binding-competent states resolved, not averaged away
  • Designs treated as hypotheses, tested at the bench
  • One shared evidence layer across experiment and computation
  • Proprietary ensemble and assay data that compounds across programs
Capabilities

From the first measurement to the bench

See the science behind the Product Engine and the programs it drives

Science Pipeline