Research Spaces
Chemistry Space
Chemistry Space is an AI-powered molecular intelligence workspace for drug discovery chemists. It interprets SMILES strings, predicts physicochemical and ADMET properties, plans retrosynthesis routes, and scores drug-likeness — all through a natural-language interface.
SMILES Structure Analysis
Paste any SMILES string directly into the chat. Chemistry Space will parse the structure, display a 2D depiction, calculate molecular properties (MW, LogP, TPSA, HBD, HBA, rotatable bonds), and flag any structural alerts (PAINS, reactive groups).
Example query
"Analyse this molecule: CC(=O)Oc1ccccc1C(=O)O"
What Wawe returns
Aspirin identified. MW: 180.16, LogP: 1.19, TPSA: 63.6 Ų, 1 PAINS alert (carboxylic acid). Lipinski Ro5: compliant. 2D structure displayed.
You can also describe a molecule in words and ask Chemistry Space to generate a plausible SMILES:
Example query
"Generate a SMILES for a fluorinated pyridine with a sulfonamide group at position 3."
What Wawe returns
Proposed SMILES with 2D depiction, IUPAC name, and property table.
ADMET Prediction
Predict absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiles for any compound. Results include confidence scores and plain-language interpretations.
- Absorption — Caco-2 permeability, human intestinal absorption (HIA), P-glycoprotein substrate/inhibitor
- Distribution — BBB penetration, plasma protein binding (PPB), volume of distribution (VD)
- Metabolism — CYP substrate/inhibitor panel (CYP1A2, 2C9, 2C19, 2D6, 3A4)
- Excretion — Half-life estimate, renal clearance
- Toxicity — hERG inhibition, AMES mutagenicity, hepatotoxicity, carcinogenicity
Example query
"What is the ADMET profile of imatinib?"
What Wawe returns
Full ADMET table with traffic-light scoring: Good oral absorption, high PPB (95%), CYP3A4 substrate, moderate hERG risk (IC₅₀ ~5 µM), no AMES positivity.
Retrosynthesis Planning
Chemistry Space can propose multi-step retrosynthetic routes from a target molecule to commercially available starting materials. Each step includes the reaction type, reagents, and literature precedent.
Example query
"Plan a retrosynthesis for oseltamivir (Tamiflu) using commercially available reagents."
What Wawe returns
A 6-step retrosynthetic tree with disconnection logic, reagent suggestions, reaction confidence scores, and reference DOIs for each step.
Drug-likeness Scoring
Compounds are automatically scored against multiple drug-likeness rule sets:
Lipinski Ro5MW ≤ 500, LogP ≤ 5, HBD ≤ 5, HBA ≤ 10Veber rulesTPSA ≤ 140, rotatable bonds ≤ 10Ghose filterMW 160–480, LogP −0.4 to 5.6, MR 40–130QED score0–1 composite desirability score
Tips for Chemistry Space
- ✓Use canonical SMILES — RDKit canonical form prevents parsing errors.
- ✓Ask for SAR suggestions — "How would a methyl group at position 4 affect LogP?" works well.
- ✓Compare multiple compounds — "Compare the ADMET of compound A vs compound B" generates a side-by-side table.
- ✓Request synthesis feasibility — Add "using Suzuki coupling" or "without protecting groups" to constrain routes.