A novel descriptor based on atom-pair properties

Abstract

BackgroundMolecular descriptors have been widely used to predict biological activities and physicochemical properties or to analyze chemical libraries on the basis of similarity. Although fingerprints and properties are generally used as descriptors, neither is perfect for these purposes. A fingerprint can distinguish between molecules, whereas a property may not do the same in certain cases, and vice versa. When the number of the training set is especially small, the construction of good predictive models is difficult. Herein, a novel descriptor integrating mutually compensating fingerprint and property characteristics is described. The format of this descriptor is not conventional. It has two dimensions with variable length in one dimension to represent one molecule. This format is not acceptable for any machine learning methods. Therefore the distance between molecules has been newly defined for application to machine learning techniques. The evaluation of this descriptor, as applied to classification tasks, was performed using a support vector machine after the features of the descriptor had been optimized by a genetic algorithm.ResultsBecause the optimizing feature is time-intensive due to the complicated calculation of distances between molecules, the optimization was forced to stop before it was completed. As a result, no remarkable improvement was observed in the classification results for the new descriptor compared with those for other descriptors in any evaluation set used in this work. However, extremely low accuracies were also not found for any set.ConclusionsThe novel descriptor proposed in this work can potentially be used to make highly accurate predictive models. This new concept in descriptors is expected to be useful for developing novel predictive methods with quick training and high accuracy.

Adverse drug reactions triggered by the common HLA-B*57:01 variant: virtual screening of DrugBank using 3D molecular docking

“…In the absence of extensive HLA-related chemogenomics data in the public domain, the development of virtual screening models that can accurately forecast drug-HLA interactions is extremely difficult. As we noted in our proof - of - concept study [ 44 ], there appears to be an inconsistent application of the molecular docking methodology when studying HLA systems. Indeed, the literature tends to be rather scarce regarding the pre-processing of HLA protein variants prior to modeling as well as which considerations (if any) were made with regards to the co-binding peptide and its ability to stabilize the bound drug.…”

“…In our recent study [ 44 ], we explored the complex intermolecular interactions between the binding pocket of HLA-B*57:01, the co-binding drug abacavir, and three co-binding peptides from the three available X-ray crystals (PDB: 3VRI, 3VRJ, 3UPR) by Illing et al [ 15 ] and Ostrov et al [ 16 ]. After conducting structural alignments of the individual components of our system (HLA-B*57:01 peptide binding pocket, bound abacavir, and co-binding peptide), we concluded that the most significant differences between binding pocket, abacavir, and peptide occurred from the peptide amino acid sequence [ 44 ]. Performing a peptide backbone alignment revealed that the 3D-structure of the peptide backbone was highly conserved [ 44 ].…”

“…After conducting structural alignments of the individual components of our system (HLA-B*57:01 peptide binding pocket, bound abacavir, and co-binding peptide), we concluded that the most significant differences between binding pocket, abacavir, and peptide occurred from the peptide amino acid sequence [ 44 ]. Performing a peptide backbone alignment revealed that the 3D-structure of the peptide backbone was highly conserved [ 44 ]. We also conducted molecular docking using Glide from the Schrodinger Suite to self-dock abacavir with and without the three co-binding peptides P1 (PDB: 3VRI), P2 (PDB: 3VRJ), and P3 (PDB: 3UPR).…”

“…We also conducted molecular docking using Glide from the Schrodinger Suite to self-dock abacavir with and without the three co-binding peptides P1 (PDB: 3VRI), P2 (PDB: 3VRJ), and P3 (PDB: 3UPR). Interestingly, we found that the co-binding peptide provided ~ 2 kcal/mol of stabilization as shown by their respective Docking Score (DS) and also proceeded to conserve the binding mode orientation of abacavir [ 44 ]. When docking was performed without co-binding peptide, abacavir was observed in two stable binding modes, but when peptide was included in the docking procedure, there was only one stable binding mode remaining [ 44 ].…”

“…Interestingly, we found that the co-binding peptide provided ~ 2 kcal/mol of stabilization as shown by their respective Docking Score (DS) and also proceeded to conserve the binding mode orientation of abacavir [ 44 ]. When docking was performed without co-binding peptide, abacavir was observed in two stable binding modes, but when peptide was included in the docking procedure, there was only one stable binding mode remaining [ 44 ]. Next, we docked a small test set of predicted HLA-liable drugs including two HLA-B*57:01 actives: flucloxacillin and pazopanib [ 17 , 18 ].…”