MULTIVARIATE QSAR MODELING OF THE RATE OF REDUCTIVE DEHALOGENATION OFHALOALKANES

The pseudo first-order rate constants for reductive dehalogenation und er anoxic conditions have recently been reported for a series of halog enated aliphatic hydrocarbons. In this paper it is shown that multivar iate quantitative structure-activity relationship (QSAR) modelling of these data is possible. Based on a training set of nine compounds and using information from 36 chemical and biological model systems, a QSA R was developed explaining 82% (R-2) and predicting 53% (Q(int)(2)) of the variation in reductive dehalogenation. The QSAR was calculated us ing the partial least squares (PLS) method. The derived QSAR was valid ated using an external validation set of six compounds for which exper imentally determined rate constants were available. The external valid ation showed that the QSAR could predict 62% (Q(ext)(2)) of the respon se variation in the validation set. in addition, the validity of the Q SAR was explored using a combination of cross-validation and permutati on. Fifty repetitive randomizations of the real response were made and for each resulting reordered response the full PLS analysis was condu cted. This yielded 50 pairs of Q(int)(2) and Q(ext)(2) which were comp ared with those values of the 'real' model. Thus it was demonstrated t hat the QSAR produced Q(2)-values significantly exceeding those of the random models and consequently the model was concluded to be valid Th e interpretation of the QSAR suggested that the mechanism of reductive dehalogenation involves elements of hydrophobicity and electron trans fer processes. (C) 1996 by John Wiley & Sons, Ltd.