Project 7. Improved computational methods for prediction of mutagenicity

 

In close collaboration with prof. Buydens and dr Ron Wehrens, Chemometrics, RUN

 

Key project members: Olaf Othersen, Lars Ridder, Markus Wagener

 

Sponsor: Schering-Plough

 

Project Goals

Prediction of mutagenicity is an area of interest in pharmaceutical research as genotoxicity of a drug candidate is detrimental in development. Since mutagenicity is directly related to the chemical properties of a compound, i.e. its reactivity and affinity towards DNA, it is generally believed that prediction of mutagenicity on the basis of chemical structure should be feasible at a reasonable accuracy. Goal of the project is to develop a reliable and maintainable alerting system for genotoxicity in drug-like compounds.

 

Approach

We develop an automated and systematic method to derive molecular fragments related to mutagenicity from experimental data. Such an automated approach will result in an improved mutagenicity prediction method which is easy to maintain and update when new data becomes available. In addition to structural alerts, other descriptors and statistical techniques will be applied to improve the predictability of mutagenicity caused by non-covalent drug-DNA interactions. It will also be investigated if the accuracy of prediction can be improved by training on specific compound classes.

 

Results to date

A web-based visualization system able to show the resulting structural alerts and relevant information in a (hyper-)linked manner has been developed. The initial substructure alerts were refined by combining related mutagenic fragments into more general alerts and by incorporating the chemical context into the alert description (i.e. R-groups). This results in more predictive and interpretable alerts for mutagenicity. In the future, we plan to incorporate descriptions of non-covalent DNA interactions into the alerting system as well as to examine the performance of the automated approach on different compound classes, in order to see if we can improve the alerting system by focusing on these specific classes.

 

References

2005 R.D. Snyder, M.D. Smith. Computational prediction of genotoxicity: room for improvement. Drug Discov. Today 10, 1119-1124.

2004 R.D. Snyder, G.S. Pearl, G. Mandakas, W.N. Choy, F. Goodsaid, I.Y. Rosenblum. Assessment of the sensitivity of the computational programs DEREK, TOPKAT, and MCASE in the prediction of the genotoxicity of pharmaceutical molecules. Environ. Mol. Mutagen 43, 143-158.

2005 J. Kazius, R. McGuire, R. Bursi. Derivation and Validation of Toxicophores for Mutagenicity Prediction. J. Med. Chem. 48, 312-320.