PXR Challenge Train/Test Dataset A high-quality experimental dataset for predicting human Pregnane-X Receptor (PXR) induction, comprising over 11,000 compounds screened using a high-fidelity in-house assay. This is the largest publicly available PXR activity dataset, released as part of the OpenADMET PXR Induction Blind Challenge. Blog post: Announcing the Next OpenADMET Blind Challenge: Predicting PXR Induction Challenge Space: openadmet/pxr-challenge Challenge period: April 1 – July 1, 2026 Produced by: OpenADMET CHANGELOG Updated 2026-07-03 Added phase 2 unblinded labels for the default test set subset in phase_2_unblinded config. Updated 2026-05-27 Added phase 1 unblinded labels for the default test set subset in phase_1_unblinded config. Updated 2026-05-27 Additional crude data added, see here for more detail. Updated 2026-04-09 dropping some compounds, fixing minor confidence interval issues and improving naming join. See here for more details. Dataset contents Config Split Description default train Primary assay training set (pEC50, Emax) default test 513-compound blinded test set counter_assay train PXR-null counter-assay training data structure test 184 molecules with X-ray crystal structures single_concentration train Single-concentration screening data (log2 fold change) crudes_htchem train Direct to biology assays of crudes (see our blog post here) phase_1_unblinded test Phase 1 unblinded subset of the primary blinded test set with assay labels phase_2_unblinded test Phase 2 unblinded subset of the primary blinded test set with assay labels