Technology for Growth

Quantitative Structure Activity Relationships, (QSAR) are mathematical models, usually derived with the aid of computer modelling of chemicals, which attempt to explain the biological effect of chemicals in terms of one or more properties of the chemicals concerned. QSARs are often used in the pharmaceutical and agrochemical industries as part of the drug or pesticide discovery programme.

A white paper from the EU on Future Chemicals Policy, (REACH), recommended a more rigorous scheme of testing of industrial chemicals for safety in the environment.

To avoid using animal experimentation, it is being proposed that QSAR analysis could be used to predict chemical safety against a number of ecotoxicological endpoints. There are almost as many ways of conducting QSARs as there are those doing it. The European Chemicals Bureau (ECB) (ecb.jrc.it/qsar/) , have developed a series of tools and guidelines on acceptable standards for QSARs. Issues such as applicability of domains, validation methods and modelling algorithms are all under debate. Where previously only multiple regression techniques on congeneric chemicals were considered robust, now we are exploring such techniques as neural networks, genetic algorithms and machine learning, to develop reliable models on large data sets of unrelated chemical structures.

• TfG has expertise and software to conduct and validated QSARs to any standard likely to be adopted by ECB.
• TfG can build and optimise molecules, calculate up to 2000 molecular descriptors, and develop QSARs using a range of statistical techniques, including genetic algorithms and neural networks. 

Useful Link: QSAR World . QSAR World is a free online resource dedicated to QSAR. It is an effort to build a vibrant and interactive community of QSAR professionals, researchers and students.