An Integrated Virtual Screening Approach for VEGFR-2 Inhibitors

Abstract

In recent years, various virtual screening (VS) tools have been developed, and many successful screening campaigns have been showcased. However, whether by conventional molecular docking or pharmacophore screening, the selection of virtual hits is based on the ranking of compounds by scoring functions or fit values, which remains the bottleneck of VS due to insufficient accuracy. As the limitations of individual methods persist, a comprehensive comparison and integration of different methods may provide insights into selecting suitable methods for VS. Here, we evaluated the performance of molecular docking, fingerprint-based 2D similarity and multicomplex pharmacophore in an individual and a combined manner, through a retrospective VS study on VEGFR-2 inhibitors. An integrated two-layer workflow was developed and validated through VS of VEGFR-2 inhibitors against the DUD-E database, which demonstrated improved VS performance through a ligand-based method ECFP_4, followed by molecular docking, and then a strict multicomplex pharmacophore. Through a retrospective comparison with six published papers, this integrated approach outperformed 43 out of 45 methods, indicating a great effectiveness. This kind of integrated VS approach can be extended to other targets for the screening and discovery of inhibitors.

Fragment virtual screening based on Bayesian categorization for discovering novel VEGFR-2 scaffolds

“…However, from top 0.5 to 0.1 %, the EF values did not show any significant increase which may imply that the increased proportion of actives was almost the same with the increased proportion of compounds on the hit-lists. In terms of AUC values, all the statistics were striking and amazing as for the training and internal test sets it reached 1* [29] from top 1 % to top 0.1 %, demonstrating that the hit-lists only contained actives which is desirable for VS investigation which have been illustrated in our study [29]. For the external test set it also reached this point from top 0.2 % to top 0.1 %.…”

“…In this approach, a structural descriptor ECFP_4 which yields better performance among 13 extended-connectivity fingerprints [29], and 9 physiochemical descriptors including ALogP, number of hydrogen donors and acceptors, rings, aromatic rings, rotatable bonds, molecular solubility and Molecular_FPSA were employed to construct the models as these molecular descriptors are widely used in ADME predictions [16,20]. Compounds of 409 actives and 24,680 decoys within similar physicochemical space ( Fig.…”

“…The self-and cross-docking analysis of VEGFR-2 crystalized complexes in our previous study [29] indicates that the combination of Glide_SP and PDB 3EWH performs best and is considered as the most desirable solution for dockingbased VS of VEGFR-2 inhibitors. Thus, Glide [51] in the Schrodinger 2009 software was adopted for molecular docking analysis [52].…”

“…Most VEGFR-2 smallmolecule inhibitors exert biological effect by competitively binding to the highly conserved ATP pocket [26,27]. Many types of VEGFR-2 kinase inhibitors have been already approved by the Food and Drug Administration (FDA) for clinical use and they consist of diaryl ureas (e.g., sorafenib, regorafenib), pyrimidines (e.g., pazopanib), quinazolines (e.g., vatalanib), indolinones (e.g., sunitinib) and some other types (e.g., axitinib) [28,29]. Despite the huge advances in developing VEGFR-2 inhibitors, there are undoubtedly many issues that are not yet fully explored, from both the clinical and biochemical perspectives.…”

Virtual Screening Strategies: A State of Art to Combat with Multiple Drug Resistance Strains

“…Complete presentation of these approaches is beyond the scope of this review, however herein, we have discussed some of the protocols successfully implemented in recent studies. By combining the molecular docking, pharmacophore and fingerprint-based 2D similarity, a novel two layer workflow was developed to enhance the virtual screening performance [71]. By combining these methods the authors noticed the improvement in the performance of virtual screening process on DUD database.…”