Title:

Mapping of Biological Function with Shape
Fingerprints: Identifying Shape Space which
Well-Predicts Function

Author:

James A. Haigh(1) and
Alfred A. Rabow(2)

(1)University of Sheffield and
(2)AstraZeneca


Abstract:

This work demonstrates the use of Shape Fingerprints to identify
regions of interest in an self organizing map (SOM) of NCI
anti-cancer compounds.   Over 24000 compounds with comprehensive
growth inhibition data (representing the behaviour towards thousands
of targets) have been mapped.  Fingerprints, both Daylight and
Shape, have then been used to recapitulate this map either based on
chemical structure or 3D shape.  Results are presented and interest
is directed to where Shape performs particularly well in contrast
to Daylight Fingerprints, particular those regions which are of
therapeutic interest, such as cell-cycle targets.

ShapeFPCupPoster.pdf