SimBioSys Blog

SimBioSys is a distributor of the CLiDE software, a software package for converting chemical structure images to chemical structures that can be read and interpreted by chemistry software packages, such as ChemDraw and ISIS Draw for example. The software package has been developed in the past by two of our founders: Aniko Simon PhD, Computer Scientist, currently VP of Business Development at SimBioSys, and over many years by Prof A. Peter Johnson (http://www.chem.leeds.ac.uk/People/Johnson.html), an expert in the field of de-novo structure design, synthetic chemistry and the applications of software to chemical structures. CLiDE is installed in organizations around the world and, for many years held a unique position. A new publication on CLiDE just came out a few weeks ago, by the current development team headed by: Aniko T. Valko, see the full citation at:
SimBioSys scientific publications page or ACS - JCIM page

This recent paper systematically evaluates CLiDE Pro’s performance on a large variety of structures, that surpasses our previous validation set for CLiDE. The authors are offering this new, carefully selected test set for base-lining and testing other optical chemical structure recognition (OCSR) tools. They suggest that this test set could be the starting point for a community-based effort to establish a benchmarking test set which would include different categories of images each of which dealt with specific problem types.
This new OCSR baseline testset is available from the publisher of the CLiDE paper as supporting information to the paper as well as downloadable from our web-site: http://www.simbiosys.ca/clide/validation.html

Scoring is undoubtedly the most challenging aspect of docking. A new, comprehensive scoring evaluation paper was recently published by Nicolas Moitessier’s group from McGill in the Journal of Chemical Information and
Modeling. The group which is actively pursuing development of its own docking and scoring methods (Fitted and RankScore), evaluated the effect of protein flexibility and water molecules on the performance of 18 different scoring functions, and placed eHiTS among the top performers.

http://pubs.acs.org/doi/abs/10.1021/ci8004308

Docking Ligands into Flexible and Solvated Macromolecules. 4. Are Popular Scoring Functions Accurate for this Class of Proteins?
Pablo Englebienne and Nicolas Moitessier
Publication Date (Web): May 15, 2009 (Article)
DOI: 10.1021/ci8004308

TOC picture:

We greatly value this recognition, and it is definitely reassuring us as developers that our special approach to scoring is delivering good results. However, we also bear in mind that scoring functions are still not performing at a desirable level, and that the docking paradigm critically depends on the ability to rank poses, and to evaluate binding energies in a way that will enhance the predictive capabilities of in-silico models. We are therefore continuously working on improving our algorithms, and our scoring function, and we believe that the scoring of our newest release, eHiTS 2009, is already an improvement over the 6.2 version that was used in the comparative study.

Our commitment to our users and to high scientific standards is among our core values, and we trust that the next release of eHiTS will raise the bar of scoring even higher.