We are writing an article which will explain the computational science behind the THINK software.

In the mean time if you have access to the science literature a research paper describes a related approach to that used by THINK.

Design of Libraries To Explore Receptor Sites Catherine M. Murray ‹ and Stephen J. Cato* J. Chem. Inf. Comput. Sci. 1999, 39, 46-50.

Abstract

Despite rapid progress in both combinatorial chemistry and high-throughput screening, the number of molecules that could potentially be made and tested for biological activity still far exceeds the capacity for synthesis or screening. Consequently, it is potentially valuable to select and synthesize sublibraries that contain rationally selected subsets. When the structure of the protein receptor site is known, this may be used to impose restrictions of the selection on molecules. This paper describes a method for rapid analysis of large virtual libraries to select a subset that can exhibit at least one conformer which will interact strongly with the receptor and fit within the receptor site.

The two authors of the the paper were from Chemical Design Ltd., UK , and Chemical Design Inc., California respectively - Chemical Design was the Chemistry software company which Keith Davies founded.

A research paper from the University of Oxford has just been published, this paper gives some more of the background science to the computational software.

Similarity Calculations Using Two-Dimensional Molecular Representations Benjamin C. P. Allen, Guy H. Grant, and W. Graham Richards* New Chemistry Laboratory, University of Oxford J. Chem. Inf. Comput. Sci. 2001, 41, 330-337

Abstract

Molecular similarity calculations are important for rational drug design. Time constraints prevent these techniques being used on large data sets or on large molecules. By reducing the molecular representation to a two-dimensional form, the alignment of the molecules can be greatly speeded up. The accuracy of the resulting similarity values can be improved by using a neural network.

How long has it taken to develop the software?

Development of the THINK software started in 1998.

Who developed THINK?

THINK has been developed by Keith Davies (Keith.Davies@chemistry.oxford.ac.uk).

Figure. A molecule (GDP) in the active site of a protein (shown zoomed in on the active site with a solvent surface covering the protein).