OpenPep

Open data to accelerate the discovery of new peptides.

OpenPep is building an open foundation dataset of peptide–protein binding to accelerate AI for peptide medicines. We measure diverse peptides across disease-relevant targets, share the data, and train foundation models for peptide discovery.

Read the Science

How we produce targets, measure binding and train models.

Propose Target for Discovery

Nominate up to five human proteins for OpenPep to study.

Research approach

  1. Produce targets

    Use cell-free protein synthesis to build a broad panel of disease-relevant targets.

  2. Measure binding

    Screen chemically diverse peptide libraries and collect quantitative binding measurements.

  3. Train models

    Learn from the shared dataset and evaluate predictions on targets held out of training.