Stanford Researchers Leverage Deep Learning with GLOW and IVES to Transform Molecular Docking and Ligand Binding Pose Prediction

Deep learning has the potential to improve molecular docking by improving scoring functions. Current sampling protocols often require prior information to generate accurate ligand binding poses, which limits the accuracy of the scoring function. Two new protocols, GLOW and IVES, developed by researchers at Stanford University, address this challenge and demonstrate improved efficiency in pose sampling. Comparative evaluation of various protein structures, including those generated by AlphaFold, validates the methods.

Deep learning in molecular docking often relies on rigid protein docking data sets, neglecting the flexibility of proteins. Although flexible docking takes into account the flexibility of proteins, it tends to be less precise. GLOW and IVES are advanced sampling protocols that address these limitations and consistently outperform reference methods, especially in dynamic junction zones. It holds promise for improving ligand pose sampling in protein-ligand docking, which is crucial for improving deep learning-based scoring functions.

Molecular docking predicts the placement of ligands at protein binding sites, which is crucial for drug discovery. Conventional methods face challenges in generating accurate ligand poses. Deep learning can improve accuracy, but it depends on effective pose sampling. GLOW and IVES upsample for challenging scenarios, increasing accuracy. Applicable to predicted or ligand-free protein structures, including those generated by AlphaFold, they offer curated data sets and open-source Python code.

GLOW and IVES are two pose sampling protocols for molecular docking. GLOW employs a smoothed van der Waals potential to generate ligand poses, while IVES improves accuracy by incorporating multiple protein conformations. Performance comparisons with reference methods show the superiority of GLOW and IVES. Test set evaluation measures correct pose percentages in cross-docking cases. The pose quality of the seeds is vital for efficient IVES, and the Smina docking score and score are used for selection.

GLOW and IVES outperformed baseline methods in accurately sampling ligand poses, excelling in challenging scenarios and AlphaFold benchmarks with major protein conformational changes. Evaluation of the test sets confirmed their increased likelihood of sampling correct postures. IVES, which generates multiple protein conformations, offers benefits for geometric deep learning of protein structures, achieving performance comparable to Schrodinger IFD-MD with fewer conformations. Ligand pose datasets are provided for 5000 protein-ligand pairs generated by GLOW and IVES, aiding the development and evaluation of deep learning-based scoring functions in molecular docking.

In conclusion, GLOW and IVES are two powerful pose sampling methods that have been shown to be more effective than basic techniques, particularly in difficult scenarios and AlphaFold benchmarks. Multiple protein conformations can be generated with IVES, which is very advantageous for deep geometric learning. Furthermore, the data sets provided by GLOW and IVES, which contain ligand positions for 5000 protein-ligand pairs, are invaluable resources for researchers working on deep learning-based scoring functions in molecular docking.

Review the Paper and GitHub. All credit for this research goes to the researchers of this project. Also, don't forget to join. our 34k+ ML SubReddit, 41k+ Facebook community, Discord channel, and Electronic newsletterwhere we share the latest news on ai research, interesting ai projects and more.

If you like our work, you'll love our newsletter.

<!– ai CONTENT END 2 –>