liveBook · Manning

Inter-cluster distance

As explained in Build AI Drug Discovery Pipelines

Intercluster distance is a rule for measuring separation between two clusters from the distances among their members.

Inter-cluster distance measures separation between two clusters using distances among their member compounds. In agglomerative clustering, the closest clusters merge repeatedly until all compounds form one hierarchy. section 7.2.3

Common linkage rules differ in which member-to-member distances they use:

  • Single linkage uses the minimum pairwise distance. section 7.2.3
  • Complete linkage uses the maximum pairwise distance. section 7.2.3
  • Average linkage averages distances across all cross-cluster pairs. section 7.2.3
  • Ward’s method merges clusters that produce the smallest increase in total within-cluster variance. section 7.2.3

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines four linkage rules and their tradeoffs, then notes Ward’s mismatch with sparse binary fingerprints.
Examples
Uses RDKit fingerprints, Ward clustering, dendrograms, and Taylor-Butina comparisons to evaluate molecular library diversity.
What it assumes
Assumes compound distances or fingerprints and basic clustering vocabulary, while defining linkage choices from first principles.
What you can do after
Apply it via RDKit and scikit-learn to cluster compounds and sample representatives near cluster centers.
How it connects
Links hierarchical clustering, agglomerative clustering, Ward’s method, molecular fingerprints, and Tanimoto-based clustering.

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.