medchem

To filter a compound library for druglike candidates, apply medicinal chemistry rules like Ro5 and Veber to evaluate druglikeness and flag issues. This Skill processes SMILES or RDKit Mol objects in batch, returning structured results to triage molecules for hit discovery or lead optimization.

Yes, you can apply structural alerts like Lilly demerits, NIBR, or CommonAlerts to a SMILES list to prioritize hits. The Skill detects undesirable substructures and assesses molecular complexity, guiding compound selection by filtering out molecules that trigger these alerts during library triage.

The best way to select CNS druglike candidates while avoiding structural alerts is to combine composable RuleFilters like CNS and Lilly alerts in a pipeline. The Skill evaluates your SMILES inputs against these constraints, outputting structured results identifying molecules that pass both druglikeness and alert criteria.

Yes, you need RDKit, datamol, and the medchem Python library installed to run these medicinal chemistry filters. These dependencies provide the necessary molecular manipulation and cheminformatics functions required to process RDKit Mol objects or SMILES inputs for batch filtering and parallel execution.

Available medicinal chemistry rules to evaluate druglikeness include Ro5, Veber, CNS, and leadlike variants. These filters assess compound libraries by applying structural alerts and complexity metrics, enabling high-level workflows that identify druglike candidates suitable for downstream analysis in drug discovery pipelines.