liveBook · Manning

BRICS

As explained in Build AI Drug Discovery Pipelines

BRICS is an RDKit-supported method and module that fragments molecules and recombines the fragments using retrosynthetically motivated chemical rules.

BRICS (Breaking of Retrosynthetically Interesting Chemical Substructures) fragments molecules and recombines the resulting pieces with retrosynthetically motivated chemical rules. It biases compound generation toward synthetically accessible structures, but does not guarantee that every product can be synthesized. section 7.2.2

A practical RDKit workflow is:

  • Convert valid SMILES records into molecules and decompose each molecule with BRICS.BRICSDecompose. section 7.2.2
  • Convert unique fragment strings back into molecules, compute 2D coordinates, and recombine them with BRICS.BRICSBuild. section 7.2.2
  • Limit enumeration, sanitize generated molecules, and discard products that fail valence, kekulization, or other validation checks. section 7.2.2

Retrosynthetic library construction

A Python pipeline collects unique BRICS fragments from a molecular dataset, builds products with maxDepth=1, stops after 10,000 candidates, sanitizes each product, and stores valid molecules for downstream screening. section 7.2.2

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Explains fragmentation, recombination, enumeration limits, and sanitization, but stops short of guaranteeing experimental synthesizability.
Examples
Uses RDKit code that decomposes dataset molecules, rebuilds products, computes coordinates, and rejects invalid sanitized structures.
What it assumes
Assumes RDKit molecule handling and SMILES records while demonstrating the fragmentation and recombination workflow.
What you can do after
Build it from scratch by extracting BRICS fragments and enumerating validated recombinations with RDKit.
How it connects
Connects molecule fragmentation with combinatorial enumeration, retrosynthetic library design, chemical reaction rules, and synthesizability.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.