Prediction of HPLC Retention Index Using Artificial Neural Networks and IGroup E-State Indices

Abstract

A back-propagation artificial neural network (ANN) was used to create a 10-fold leave-10%-out cross-validated ensemble model of high performance liquid chromatography retention index (HPLC-RI) for a data set of 498 diverse druglike compounds. A 10-fold multiple linear regression (MLR) ensemble model of the same data was developed for comparison. Molecular structure was described using IGroup E-state indices, a novel set of structure-information representation (SIR) descriptors, along with molecular connectivity chi and kappa indices and other SIR descriptors previously reported. The same input descriptors were used to develop models by both learning algorithms. The MLR model yielded marginally acceptable statistics with training correlation r(2) = 0.65, mean absolute error (MAE) = 83 RI units. External validation of 104 compounds not used for model development yielded validation v(2) = 0.49 and MAE = 73 RI units. The distribution of residuals for the fit and validate data sets suggest a nonlinear relationship between retention index and molecular structure as described by the SIR indices. Not surprisingly, the ANN model was significantly more accurate for both training and validation with training set r(2) = 0.93, MAE = 30 RI units and validation v(2) = 0.84, MAE = 41 RI units. For the ANN model, a total of 91% of validation predictions were within 100 RI units of the experimental value.

CE50: Quantifying collision induced dissociation energy for small molecule characterization and identification

“…A model with a higher correlation coefficient and lower standard error will likely be produced with a larger dataset. The development of a predictive model using our methods was demonstrated previously in a model developed for predicting chromatographic retention index values [39]. That model was based on a much larger dataset (n ϭ 498, 394 in training set), permitting the use of a greater number of structure descriptors and also allowing the use of a nonlinear artificial neural network analysis.…”

“…Predictive models have been developed using these methods for a range of data types [41][42][43][44][45][46][47]. Thus, although the present model is not useful for predictive purposes, its development clearly suggests that modeling of a larger dataset could be predictive [23,28,39,[41][42][43][44]. Furthermore, the anticipated improved model may also assist in identifying further structural features important in determining precursor ion stability and fragmentation mechanisms [23,28,[42][43][44][45][46].…”

Optimizing Artificial Neural Network Models for Metabolomics and Systems Biology: an Example Using HPLC Retention Index Data

“…The IGroups methodology [7] explicitly represents every atom in a molecule by grouping atom level contributions from atoms that undergo similar noncovalent interactions in solution. Atom level contributions are quantified by the Atom-level Electrotopological State (E-State) [19].…”

“…MolFind/BioSM applies five additional orthogonal filters to reduce the number of false positives returned from the exact mass search [2–6]. Filters are: HPLC retention index (RI) [3, 6–7], Ecom 50 [2,3] drift index (DI) [5], biological/nonbiological (BioSM) [8] and CID (MS/MS) spectra [5, 9]. In filtering, predicted values for RI, Ecom 50 and DI are made for candidate compounds using computational models.…”

“…In previous studies [3, 6–7], RI was evaluated for the purpose of structure identification in the MolFind method. An HPLC-RI is generated using a homologous series of reference compounds with increasing lipophilicity.…”