Hello, I am trying to make something similar (hiererchical clustering figures of molecules) to what Rosenthal et al did in their Nature paper (

http://www.nature.com/nature/journal/v465/n7296/full/nature09099.html).

The Supp. Info file from this paper is here:

http://www.nature.com/nature/journal/v465/n7296/extref/nature09099-s1.pdf

On page 26-27 they say:

"Node and edge assignments were made using a custom Pipeline Pilot application.  Scaffolds

were defined from molecules using the “MurckoAssemblies” option in the Generate Fragments

component (exocyclic double bonds and linker double bonds were included, and alpha atoms

were treated as R-attachment points). Cores were defined from molecules using the same

Pipeline Pilot component, but excluding alpha atoms and linker double bonds, and converting all

bonds into single bonds. Five categories of edges were assigned. Molecule nodes were connected

to other molecule nodes that shared the same parent scaffold.  Edges were also assigned between

molecule nodes and their parent scaffold node. Edges were assigned between a parent scaffold

node and a child scaffold that contained the parent substructure.  Scaffold nodes that could not be

assigned a parent were assigned to core nodes. Core nodes were hierarchically clustered using

the McQuitty method and the Dixon-Koehler distance metric (Pipeline Pilot ECFP_4

fingerprint).  This algorithm produced a connected graph (all nodes had at least one connection).

The network was visualized in Cytoscape (v 2.6.3) using the Y-files circular layout method."

Any help in how to make the matrix file as said in BOLD in above paragraph? or the figures like the one shown in the Paper through any other method in PP?

Thank you

nar