datamol

You can convert SMILES, InChI, or SELFIES strings into standardized RDKit molecules by parsing them through a Pythonic wrapper that applies sensible default cleaning. This generates native RDKit Mol objects ready for featurization and cheminformatics analysis.

Computing molecular fingerprints and descriptors in batch is supported through high-level functions that calculate common drug-like descriptors and generate ECFP-style fingerprints. This workflow supports parallel processing controls to handle large datasets efficiently.

The best way to extract Murcko scaffolds and cluster molecules for diversity analysis is using scaffold-aware workflow functions that extract scaffolds and cluster molecules. This enables you to pick diverse subsets and perform structural similarity calculations across datasets.

Yes, you can generate 3D conformers and compute SASA using RDKit defaults. The process generates ETKDG-based 3D conformers, allows clustering of conformations, and computes Solvent Accessible Surface Area for comprehensive conformer analysis.

Yes, this cheminformatics approach supports parallel processing for large molecule datasets. It exposes practical high-level functions with parallel processing controls, allowing you to specify the number of jobs to optimize batch descriptor and fingerprint computation.

To apply SMARTS-based reactions to transform molecular structures across a dataset, you can use reaction application functions that process SMARTS patterns. This allows you to systematically transform structures while returning native RDKit molecules for further analysis.