y-Randomization and Its Variants in QSPR/QSAR

Abstract

y-Randomization is a tool used in validation of QSPR/QSAR models, whereby the performance of the original model in data description (r2) is compared to that of models built for permuted (randomly shuffled) response, based on the original descriptor pool and the original model building procedure. We compared y-randomization and several variants thereof, using original response, permuted response, or random number pseudoresponse and original descriptors or random number pseudodescriptors, in the typical setting of multilinear regression (MLR) with descriptor selection. For each combination of number of observations (compounds), number of descriptors in the final model, and number of descriptors in the pool to select from, computer experiments using the same descriptor selection method result in two different mean highest random r2 values. A lower one is produced by y-randomization or a variant likewise based on the original descriptors, while a higher one is obtained from variants that use random number pseudodescriptors. The difference is due to the intercorrelation of real descriptors in the pool. We propose to compare an original model's r2 to both of these whenever possible. The meaning of the three possible outcomes of such a double test is discussed. Often y-randomization is not available to a potential user of a model, due to the values of all descriptors in the pool for all compounds not being published. In such cases random number experiments as proposed here are still possible. The test was applied to several recently published MLR QSAR equations, and cases of failure were identified. Some progress also is reported toward the aim of obtaining the mean highest r2 of random pseudomodels by calculation rather than by tedious multiple simulations on random number variables.

Application of quantitative structure–activity relationships for modeling drug and chemical transport across the human placenta barrier: a multivariate data analysis approach

“…The proposed model was considered robust if the new models on the set with randomized responses presented signifi cantly lower r 2 and Q 2 than the original model. In addition, as a more critical test, randomized response data were used to run from the beginning the model development procedure (Rücker et al, 2007). Finally, external validation was performed in the fi nal model by splitting the data set into several training and test sets.…”

A QSAR Model for Predicting the Anti HIV‐1 Activity of TIBO Derivatives Using the Norm Indexes

“…Y-randomization test has been widely used to evaluate the robustness of a QSAR model, 28,29 In this work, five random shuffles of the y vector were performed and results were listed in Table 6. Results shown in Table 6 suggested that values of R 2 and Q 2 of these new random models were significant lower than those of our original model.…”

Integrating Structure and Ligand-Based Approaches for Modelling the Histone Deacetylase Inhibition Activity of Hydroxamic Acid Derivatives

“…At first, the statistical parameters of the LOO-CV were checked. According to the obtained results, it is remarked that the internal predictions are good because the variance explained in the prediction by LOO (Q At last, to demonstrate that the model was not the result of a chance in fitting given data, the Y-scrambling procedure was employed, placing the answers (dependent variable) at random, keeping all descriptors (independent variables) in the model, and then, performing whole model building procedure n times again [53]. If the model performance decays dramatically, it is a sign of a lack of chance.…”