Clustering of environmentally occuring chemicals according to structure similarity
Two clustering methods using structure screens have been developed and tested. The first method, using PLS-discriminant analysis, distinguishes chemicals according to their mode of action. A model has been developed based on a training set of 157 chemicals. The second method subdivides a group of chemicals into clusters using the Tanimoto coefficient as the similarity measure. The LC50 data were modelled by regression analysis and statistical validation. The correlation between the clusters and the EINECS data was checked by PCA, which makes it possible to define the limits of each model.
Bibliographic Reference: Paper presented: QSAR for predicting fate and effects of chemicals in the environment, Brno, March 13-14, 1995
Availability: Available from (1) as Paper EN 38926 ORA
Record Number: 199510660 / Last updated on: 1995-07-07
Original language: en
Available languages: en