Skip to content

ProTox 3.0

Charité - Universitätsmedizin Berlin

ProTox 3.0 is the Charité web service for prediction of toxicities of small molecules, the successor to ProTox-II, offering web-based predictions and API usage documented in its FAQ.

Catalog updated ·

Overview

ProTox 3.0 is a web service for the prediction of toxicities of small molecules, described on its site as a virtual lab for computational toxicity estimation. It is operated by the Charité group and is the successor to the ProTox-II web server; this entry records the current 3.0 service rather than the older version.

The official FAQ describes access for academic and non-commercial users and limits the API to 250 queries per source IP per day. API use is documented through a POST interface and a downloadable Python example. Submitted structures are sent to the service operator. See: https://tox.charite.de/protox3/index.php?site=faq

Predictions are model-based estimates whose applicability depends on the underlying training data and methods; they are screening signals that complement, not replace, experimental toxicology. Submitted structures and queries are subject to the service's own terms.

Key Features

  • Web interface for submitting compounds and reading predicted toxicity results.
  • FAQ-documented usage including an API section and prediction workflow descriptions.
  • Successor to the ProTox-II web server operated by the same group.

Use Cases

  • Screen candidate compounds for toxicity-related signals before ordering or synthesis.
  • Support teaching discussions on computational toxicology with an accessible web tool.

How to Use

Submit compounds through the web interface and read the result report; the FAQ describes usage and API options. Predictions are computational estimates derived from trained models and similarity searching, not experimental toxicity data, and the service's own terms apply to submitted information. See: https://tox.charite.de/

The official FAQ describes access for academic and non-commercial users and limits the API to 250 queries per source IP per day. API use is documented through a POST interface and a downloadable Python example. Submitted structures are sent to the service operator. See: https://tox.charite.de/protox3/index.php?site=faq

Sources and verification

This entry describes capabilities documented by the linked project sources. Catalog edits are not evidence of a successful run, compatibility test or scientific performance evaluation.

Repository source reviewed for this catalog version:

Read the editorial and verification policy

Related resources

ADMET-AI is an ADMET prediction platform built on Chemprop models trained on TDC ADMET datasets, offered via command line, Python API and a live web server, under the MIT license.

Open sourcePython

Molecular Property Prediction · Drug Discovery

CYPMode

Open Source

CYPMode is a research Python toolkit for time-dependent CYP3A4/CYP2D6 inhibition prediction and CYP3A4 ligand-complex modeling, developed around the OpenADMET blind challenge.

Open sourcePython

Molecular Property Prediction · Computational Chemistry · Drug Discovery

DrugAgent is a research multi-agent framework that combines LLM planning with domain-guided code generation to build and evaluate machine-learning workflows for drug discovery.

Molecular Property Prediction · Drug Discovery

SwissADME

Platforms

SwissADME is a free web tool operated by the University of Lausanne's Molecular Modelling Group and the SIB Swiss Institute of Bioinformatics that computes molecular parameters related to pharmacokinetics and dynamics properties, plus druglikeness indicators, from a list of SMILES.

Molecular Property Prediction · Drug Discovery

Therapeutics Data Commons provides therapeutic machine learning datasets, Python loaders, data splits, evaluation metrics and benchmarks for prediction and molecule-generation research.

Open sourcePython

Molecular Property Prediction · Drug Discovery

TorchDrug

Open Source

TorchDrug is a PyTorch-based toolkit for graph and molecular machine learning, with SMILES input, graph operations, property-prediction building blocks and multi-CPU/GPU execution.

Open sourcePython

Molecular Property Prediction · Drug Discovery

Related guides

Workflows

AI for Drug Discovery: Agents, MCP Servers and Platforms

Choose tools by workflow stage: target evidence, protein structures, known binding sites, molecular generation, property analysis and screening. Compare research agents, MCP integrations, hosted APIs and platforms, with a proposed end-to-end example and practical evaluation gates.