Fig. 7

Hierarchical scaffold clustering in KNIME: Bemis-Murcko scaffolds and associated UniProt IDs (as a list) are forwarded as an input. The ‘Table Row To Variable Loop Start’ node forwards Bemis-Murcko scaffolds per respective target into the loop. Molecular distances are computed for the retained scaffolds (‘MoSS MCSS Molecule Similarity’ node) and hierarchical clusters are generated accordingly (the ‘Hierarchical Clustering [DistMatrix]’ node). Scaffolds are grouped into a specific group by applying the ‘Hierarchical Cluster Assigner’ node. The loop is then repeated for the remaining targets from the input table. The output table contains UniProt IDs, associated scaffolds, and cluster IDs