In our experience, spatial statistics methods were extremely useful for the analysis of descriptor representations of chemical datasets. Apart from the impact of dataset topology on validation experiments of virtual screening techniques, the field of chemoinformatics provides lots of opportunities for the application of spatial statistics methods. Chemical space is infinite!

Therefore we provide the toolbox implementing the Refined Nearest Neighbor Analysis functions used for the design of MUV for download. In addition to Refined Nearest Neigbor Analysis and the particular design algorithms for creating MUV datasets, the toolbox provides many more functions and methodologies for the analysis of mapped point patterns.

For an excellent, easily comprehensible introduction to spatial statistics refer to:

Fortin, M.; Dale, M. R. T.
Spatial analysis: a guide for ecologists.
Cambridge University Press: Cambridge, UK, 2005

This book is a pleasure to read and introduces all concepts applied in the Spatial Statistics Toolbox.

Feel free to use the toolbox for your research. If you do so, please cite:

Rohrer, S.G.; Baumann, K.
Maximum Unbiased Validation (MUV) Datasets for Virtual Screening Based on PubChem Bioactivity Data
J. Chem. Inf. Model., in press

It would also be great, if you wrote us a few lines about how the toolbox was useful for you to muv@tu-bs.de.

Download:

File	Version, Date	Comments
SpSt.tar.gz	0.91, 1 Oct 2008	Spatial Statistics Toolbox for MATLAB

Extract the archive and include the resulting directory SpSt/ recursively into your MATLAB path. You can do so by clicking File->Set Path...->Include with subfolders... in the MATLAB workspace.

Currently the toolbox is available only for MATLAB and you consequently need the respective license. We are working on a re-implementation using open source software.