Development of a Virtual Screening Method for Identification of “Frequent Hitters” in Compound Libraries

• DOI: 10.1021/jm010934d
• Publication Date: 2002-07-26T04:45:26Z
• Authors: Olivier Roche; Petra Schneider; Jochen Zuegge; Wolfgang Guba; Manfred Kansy; Alexander Alanine; Konrad Bleicher; Franck Danel; Eva-Maria Gutknecht; Mark Rogers-Evans; Werner Neidhart; Henri Stalder; Michael Dillon; Eric Sjögren; Nader Fotouhi; Paul Gillespie; Robert Goodnow; William Harris; Phil Jones; Mikio Taniguchi; Shinji Tsujii; Wolfgang von der ...; Gerd Zimmermann; Gisbert Schneider
• Keywords: Identification; Prioritization; Data set; Chemical database

A computer-based method was developed for rapid and automatic identification of potential "frequent hitters". These compounds show up as hits in many different biological assays covering a wide range targets. scoring scheme elaborated from substructure analysis, multivariate linear nonlinear statistical methods applied to several sets one two-dimensional molecular descriptors. The final model is based on three-layered neural network, yielding predictive Matthews correlation coefficient 0.81. This system able correctly classify 90% the test set molecules 10-times cross-validation study. database filtering, between 8% (compilation trade drugs) 35% (Available Chemicals Directory) frequent hitters. filter will be valuable tool prioritization large databases, compound purchase testing, building new virtual libraries.