Structure-carcinogenic activity relationship studies of polycyclic aromatic hydrocarbons (PAHs) with pattern-recognition methods

Recently a new methodology to identify the carinogenic activity of polycycl ic aromatic hydrocarbon (PAHs) was proposed. This methodology named electro nic indices methodology (EIM) is based on the use of local density of state s (LDOS) calculations. In this work we perform a comparative study of this methodology with principal component analysis (PCA) and artificial neural n etworks (ANN), All the physicochemical descriptors were calculated from the molecular eigen-states/spectra obtained through the well-known semi-empiri cal method parametric method 3 (PM3). PCA and ANN results show that EIM des criptors are relevant to identify the carcinogenic activity of methylated a nd non-methylated PAHs. Also, we show that the combined use of these distin ct methodologies can be an efficient and powerful tool in the structure-act ivity studies of PAHs or other organic compounds. We have studied 81 methyl ated and non-methylated PAHs, and our study shows that with the use of thes e methods it is possible to predict correctly the PAHs' carcinogenic activi ty with high accuracy (similar to 80%). (C) 2001 Elsevier Science B.V. All rights reserved.