You provide one or more compounds — by name, SMILES, or from a previous job — and select which toxicity models to run. Your model selection most shapes the result: it decides which of the 61 endpoints get predicted, from a single acute-toxicity readout to the full comprehensive profile.

Not available in chat

ProTox-3 is deprecated on the platform and not registered with the orchestrator, so chat cannot call it. Use ADMET-AI for toxicity flags (hERG, Ames, hepatotoxicity) or DeepPurpose for specific endpoint predictions — existing ProTox-3 job results remain viewable.

Webapp input methods

The .bloom* webapp provides three input methods for ProTox-3 toxicity analysis:

Input type selection

Compound Names (PubChem lookup) - Default option

Enter compound names like "aspirin", "caffeine", "ibuprofen"
Automatic lookup via PubChem to retrieve SMILES
Supports common drug names and synonyms

SMILES Strings

Enter canonical SMILES directly (e.g., CC(=O)OC1=CC=CC=C1C(=O)O)
Isomeric SMILES supported for stereochemistry
Automatic validation and standardization

From Previous Job

Import compounds from completed workflows
Select from AiZynthFinder, REINVENT, or ZINC outputs
Automatic SMILES extraction from job results

Compound entry

Add multiple compounds using the Add Compound button
Maximum 50 compounds per job (warning appears at 45 compounds)
Remove compounds using the delete icon
Real-time validation of compound inputs

Webapp parameters

Job configuration

Job Title (optional): Descriptive name like "Drug Candidate Toxicity Screen"
Output Filename: Base name for output files (default: protox3_results)
Quiet Mode: Toggle reduced output (default: enabled)

Model selection

ProTox-3 organizes 61 endpoints into categorical groups. Select models by clicking panels in the interface:

Core Models

Acute Toxicity (LD50): Lethal dose predictions and toxicity classes
Toxicity Targets: Protein binding and toxic pathways
All Models (Comprehensive): Complete analysis - all available models

Organ Toxicity

Hepatotoxicity (DILI): Drug-induced liver injury
Neurotoxicity: Nervous system toxicity
Nephrotoxicity: Kidney toxicity
Respiratory Toxicity: Lung and respiratory effects
Cardiotoxicity: Heart and cardiovascular toxicity
Carcinogenicity: Cancer-causing potential
Immunotoxicity: Immune system disruption
Mutagenicity: DNA damage and mutation potential
Cytotoxicity: General cell toxicity

Nuclear Receptors

Aryl Hydrocarbon Receptor (AhR): Pathway activation
Androgen Receptor (AR): Male hormone disruption
Estrogen Receptor (ER): Female hormone disruption
PPAR-gamma: Metabolic pathway disruption

Molecular Events

Thyroid Hormone Receptor α: Thyroid disruption
Ryanodine Receptor (RYR): Calcium channel disruption
GABA Receptor (GABAR): Neurotransmitter disruption
NMDA Receptor (NMDAR): Glutamate receptor antagonism

Drug Metabolism (CYP450 Enzymes)

CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, CYP3A4

Model selection behavior

Select "All Models (Comprehensive)" to run all endpoints at once
Selecting "All Models" clears individual model selections
Selecting specific models after "All Models" replaces it with individual selection
Selected models appear as removable chips above the category panels

Validation rules

At least one compound is required
At least one toxicity model must be selected
Empty compounds are filtered out before submission
Invalid SMILES or names that fail PubChem lookup will error during processing

Rate limits

⚠︎ ProTox3 has a 250 queries/day limit. Large batches (45+ compounds) display a warning about daily quota consumption.

Workflow integration

Compounds can be imported from completed jobs via the "From Previous Job" option. Compatible upstream tools include:

AiZynthFinder: Retrosynthetic route products
REINVENT: Generated molecules from de novo design
ZINC/PubChem: Database search results
Fragment Libraries: Virtual screening hits

ProTox-3 inputs & parameters