Streamline Drug Discovery with eToxPred: Now Available on Vecura
This update enables medicinal chemists and drug discovery researchers to rapidly filter and prioritize candidate small molecules by predicting toxicity and synthetic accessibility directly within the Vecura workflow, eliminating the need for complex local infrastructure.
What is eToxPred?
eToxPred is a specialized machine-learning tool designed to streamline early-stage drug discovery by rapidly evaluating small organic compounds. It provides two critical metrics for candidate molecules: a toxicity probability (Tox-score) generated by an ExtraTrees ensemble classifier and a synthetic accessibility score (SAscore) based on fragment contributions.
It helps users filter and prioritize vast compound libraries by identifying potential safety risks and synthetic challenges before expensive wet-lab synthesis is required. It is especially useful for medicinal chemists and computational researchers aiming to optimize lead compounds efficiently.
What can users do with eToxPred on Vecura?
With eToxPred on Vecura, users can:
- Instantly evaluate the toxicity probability of a small molecule by inputting its SMILES string.
- Assess the synthetic accessibility of candidates to gauge manufacturing feasibility.
- Quickly filter large libraries of compounds to focus efforts on the most promising leads.
- Bypass complex infrastructure setup by running these predictions directly within the Vecura interface.
What the output means
The output provides a predictive report containing the Tox-score (a probability of toxicity) and the SAscore (a measure of synthetic complexity).
This output should be used to support scientific decision making. It does not replace experimental validation. Users should interpret the Tox-score as a heuristic guide based on the training distribution, not as a substitute for formal in vitro or in vivo safety testing.
Why this matters
In drug discovery, the transition from computer-aided design to physical synthesis is a major bottleneck. Identifying toxic or synthetically challenging molecules early in the pipeline prevents the wasted time and resources associated with developing compounds that are destined to fail in later stages.
By combining a robust statistical machine learning approach for toxicity with the well-established Ertl & Schuffenhauer heuristic for synthetic complexity, eToxPred offers a balanced, interpretable screening tool that bridges the gap between theoretical molecular design and practical chemistry.
- Developed by: Research group led by Dong et al.
- Source: Official GitHub repository
- Reference: Dong et al. 2018 — eToxPred original paper
Try eToxPred on Vecura.
Open the model workspace and start evaluating it with your own inputs.