Prediction of Chemical Carcinogenicity

Abstract

The base for predicting the carcinogenicity by methods of the computer chemistry is similar to those algorithms applied in the drug design to prognosticate the structure/activity relationships. However, it is well knows that there is no biological activity of carcinogenicity or mutagenicity or mutagenicity. The carcinogenicity rather appears or does not appear as a result of various processes such as cell penetration, several enzymatic metabolization reactions, DNA modification and possible repair processes. The question which one of these processes does actually occur in an individuum and at what time, largely depends on its disposition ie for example on its individual enzyme level. Yet, attemps have been made vor about 40 years to predict carcinogenicity of chemical structures through discussion of the electron density of chemical structures (aromatic substances (1)), thermodynamic consideration such as the HANSCH analysis (2), statistic analyses of the effect of functional groups (3, 4) or through so-called connectivity values (5) in connection with methods of pattern recognition. Considering the fact that out of more than 6 million of known chemical structures, only several hundred of them are deemed to surely have an epidemiological effect, this intent seems to be justified.