molclaw-denovo-sampling

This Skill generates novel drug-like molecules from scratch using a REINVENT4 de novo generative model to sample unbiased SMILES strings, applying optional Lipinski filtering to ensure drug-likeness for downstream screening.

De novo molecule generation samples entirely new molecular structures from scratch without requiring a starting structure, making it ideal for pure chemical space exploration, building diverse scaffold libraries, and initial hit discovery.

Yes, you can apply multiple filter presets and optional Lipinski Rule of Five filtering during de novo molecule generation to control chemical quality and drug-likeness of the sampled SMILES strings.

Yes, REINVENT4 supports unbiased chemical space exploration by using a de novo prior to sample entirely new SMILES strings, making it suitable for initial hit discovery where no starting molecule is provided.

Save generated SMILES strings and computed properties as a CSV file during the de novo generation process, enabling direct integration into downstream screening and optimization workflows.

Build a diverse scaffold library without a starting molecule by using de novo generation to sample unbiased SMILES strings, applying filter presets to ensure quality and diversity across the generated compound set.