Cheminformatics Tools — QSAR, ADMET & Toxicity Prediction

Build and apply quantitative structure–activity relationship models using SELFIES-encoded structures and ML regression. Predict bioactivity, potency, and physicochemical endpoints from any drawn or pasted molecule.

Generate and screen ML-derived pharmacophore models from SELFIES or drawn structures. Identify hydrogen-bond donors/acceptors, hydrophobic centres, and aromatic features — then match against compound libraries.

Predict AMES mutagenicity, hERG cardiotoxicity, hepatotoxicity, and LD₅₀ using SELFIES-aware ML classifiers. Screen compounds for safety liabilities early — before synthesis or in-vitro testing.

Generate novel drug-like molecules from scratch using generative models. Set target constraints — scaffold, property ranges, or binding site — and explore chemical space beyond known compound libraries.

Simulate thermogravimetric analysis (TGA) curves from molecular structure. Understand decomposition temperature, mass loss steps, and thermal stability — directly in the browser.

Predict differential scanning calorimetry (DSC) thermograms including melting points, glass transitions, and enthalpy values for pharmaceutical and material compounds.

Predict ¹H and ¹³C NMR spectra from any SMILES or drawn structure. Visualise chemical shifts, coupling patterns, and multiplicity for research and structure verification.

Search similar reactions by structure, explore reagent sets, and interpret mechanistic steps. Full ML-powered retrosynthesis model coming soon.