liveBook · Manning

Dendrogram

As explained in Build AI Drug Discovery Pipelines

A dendrogram is a tree diagram that represents the hierarchy and merge or split levels produced by hierarchical clustering.

A dendrogram is a tree diagram showing the hierarchy produced by hierarchical clustering. Its merge height records the distance between clusters, while long unmerged stretches indicate separation. section 7.2.3

Use a dendrogram to choose clusters and guide compound selection:

  • Draw an imaginary horizontal cut across the tree; the number of intersected vertical branches gives the resulting cluster count. section 7.2.3
  • Sample compounds across clusters to emphasize chemical diversity, recognizing that equal sampling over-represents small clusters. section 7.2.3
  • Interpret Ward-based results cautiously because Euclidean variance is not the most natural distance for sparse binary fingerprints. section 7.2.3

Compound-library workflow

Convert molecules into RDKit fingerprints, fit agglomerative clustering with Ward linkage, inspect the dendrogram, cut it at a chosen height, and sample representatives from the resulting clusters.

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Explains merge heights, separation, and horizontal cuts, then stops short of detailing dendrogram rendering code.
Examples
Uses clustered molecular fingerprints, dendrogram cuts, and representative compound samples as concrete artifacts.
What you can do after
Apply it via agglomerative clustering to choose clusters and assemble chemically diverse compound subsets.
How it connects
Connects hierarchical clustering with agglomerative clustering, Ward’s method, molecular fingerprints, and Tanimoto similarity.

Also explained in

Compare how 4 books explain dendrogram →

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.