Publications

OpenADMET generates open ADMET data and structures, and its datasets and blind challenges are being used across the community to benchmark new models, test validation strategies, and study ADMET and potency prediction. This page lists OpenADMET's own publications, followed by the papers we know about that use OpenADMET data, grouped by dataset. If your paper uses OpenADMET data and isn't listed, let us know.

As of October 2026, this page lists 33 publications:

3 OpenADMET overview and perspective papers
1 structural biology paper from OpenADMET
29 papers that use OpenADMET datasets: 3 challenge overview papers from the organizers and 26 from the wider community

Some papers use more than one dataset and appear in more than one dataset section below.

Featured Preprint

December 2025 · Preprints

OpenADMET: Embracing the Avoid-Ome to Transform Drug Discovery

Fraser, J. S.; Edgar, S.; Handly, L. N.; Kosuri, S.; Chodera, J. D.; Murcko, M.; Walters, W. P. A preprint introducing OpenADMET and the Avoid-Ome framework for transforming drug discovery through open science.

OpenADMET overviews

Perspectives on the motivation for OpenADMET, the "avoidome," and the role of blind challenges.

Publication

Blind challenges let us see the path forward for predictive models

Chodera, Walters, Kosuri, Fraser

J. Chem. Inf. Model. 2026, 66(4), 1947–1949

Perspective on why blind challenges are essential for evaluating predictive models

View Publication

Preprint

OpenADMET: embracing the avoid-ome to transform drug discovery

Fraser, Edgar, Handly, Kosuri, Chodera, Murcko, Walters

Preprints.org, Dec 2025

The case for OpenADMET and open data on the avoid-ome

Publications

OpenADMET: Embracing the Avoid-Ome to Transform Drug Discovery

OpenADMET overviews

Blind challenges let us see the path forward for predictive models

OpenADMET: embracing the avoid-ome to transform drug discovery

Structure is beauty, but not always truth

Structural biology

Structural plasticity and ligand promiscuity of CYP3A4 revealed by cryo-EM

Challenge overview papers

Assessing the state of the art in ADMET prediction: lessons from the OpenADMET-ExpansionRx community blind challenge

Operational insights from the front lines of OpenADMET: lessons from our blind challenges

A computational community blind challenge on pan-coronavirus drug discovery data

ASAP-Polaris-OpenADMET antiviral data (2025)

Mol-JEPA: a multimodal joint embedding predictive architecture for molecules

GRASP: graph representation learning with assay supervision for molecular properties

Benchmarking uncertainty and improving risk ranking in multi-task bioactivity prediction

Benchmarking experiments with external data, domain-specific features, and learning strategies for modeling ADME and potency of coronavirus main protease inhibitors

BoltzMol-1: towards reliable virtual screening for fast and cost-effective hit discovery

Surrogate machine learning models for ADME prediction: from synthetic labels to experimental data

ASTRA: statistically robust model selection from cross-validation

XIMP: cross graph inter-message passing for molecular property prediction

Machine-learning methods for pH-dependent aqueous-solubility prediction

Improved ADME prediction by multitask pretraining on predicted data: insights from the ASAP-Polaris-OpenADMET blind challenge

A comparative study of deep learning and classical modeling approaches for protein–ligand binding pose and affinity prediction in coronavirus main proteases

Deep learning vs classical methods in potency and ADME prediction: insights from a computational blind challenge

Fingerprint-based machine learning for SARS-CoV-2 and MERS-CoV Mpro inhibition: highlighting the potential of Bayesian neural networks

Polaris challenge: data-driven priors to improve docking for pose prediction

Deep-learning vs physics-based docking tools for future coronavirus pandemics

TEMPL: a template-based protein–ligand pose prediction baseline

Robust prediction of protein–ligand binding potency with multi-modal customized gate control

Superior molecular representations from intermediate encoder layers

OpenADMET-ExpansionRx data (2025–26)

Human intestinal absorption revisited: the 1998 descriptors survive, and the method extends to permeability

GRASP: graph representation learning with assay supervision for molecular properties

Evaluating autonomous LLM agents across molecular prediction and optimization benchmarks

Beacon-2: completing the reward function for chemical optimization in drug discovery

MolLedger: an additive graph neural network with chemically grounded ADME attributions

Mol-JEPA: a multimodal joint embedding predictive architecture for molecules

Model validation protocols for machine learning in small molecule drug discovery

BoltzMol-1: towards reliable virtual screening for fast and cost-effective hit discovery

Probabilistic contrastive pretraining for multi-task ADME property prediction

QuADMET-Former: pre-training on quantum mechanical properties improves ADMET prediction

PXR challenge data (2026)

Evaluating autonomous LLM agents across molecular prediction and optimization benchmarks

Predicting CYP450 induction and PXR activation with machine learning approaches: challenges and opportunities

Mol-JEPA: a multimodal joint embedding predictive architecture for molecules

Fingerprint-based machine learning for SARS-CoV-2 and MERS-CoV M^pro inhibition: highlighting the potential of Bayesian neural networks