Key Starting Material Help Tutorial

Step-By-Step Tutorial

1. Open ChemrytKSM. The default example is 4-hydroxybenzoic acid; replace it when you are evaluating another KSM.
2. Draw the molecule in the structure editor or paste its SMILES into the structure input, then enter a clear structure ID and name.
3. Confirm the displayed structure, salt or charge state, stereochemistry, and tautomer. Resolve any red structure-validation error before prediction; review yellow warnings before interpreting results.
4. Click `Analyze KSM` to open the result workspace for the current structure. Use `Run Prediction` when you need the ML property and safety screens.
5. Review the `Structure` and `ML Predictions` tabs first. Check model applicability, confidence, reactive-site or impurity cues, and any metadata gap instead of treating every number as equally reliable.
6. Use `Route Scouting` to define the API target, KSM entry role, and route count, then generate and compare route candidates and precedent evidence.
7. Complete the `Vendor` and `Analytics` tabs with supplier identity, lot scale, CAS/IUPAC checks, sample purity, reactivity, assay, chiral, moisture, and thermal evidence.
8. Use `Regulatory` and `Process Science` to draft the ICH Q11 designation rationale, model impurity carry-over, record CPP hypotheses, and review residual-solvent, elemental-impurity, stoichiometry, and EHS inputs.
9. Use `Sustainability`, `Integrations`, and `DoE` for green-chemistry metrics, informatics hand-off previews, and an experimental matrix. Save or transfer the reviewed evidence through your controlled project workflow.

Tutorial Notes

• Keep one chemically defined form throughout a comparison. Changing stereochemistry, protonation, salt form, or tautomer can change predictions and invalidate a baseline comparison.
• A validation pass means the notation appears usable; it does not prove identity, purity, or suitability as a regulatory starting material.
• When Live Edit & Compare is enabled, lock the intended baseline before editing the structure and wait for the edited-analog profile to finish updating.
• Treat route, vendor, regulatory, sustainability, and DoE outputs as structured decision support. Replace example or heuristic values with project-specific, traceable evidence.
• Investigate unavailable models or metadata gaps explicitly. Do not silently substitute a missing prediction with an assumed value.
• Record the exact structure, model result date, supplier/sample context, assumptions, and experimental follow-up needed for each decision.