I'm so proud of my research group in how they support each other and are making the most of these crazy times. We were fortunate to have a beautiful late-summer day for a group picnic.
Accurate Pocket Identification for Binding-Site-Agnostic Docking
1. RAPID-Net is a newly developed pocket identification model designed for blind docking workflows. It seamlessly integrates with AutoDock Vina and significantly improves docking accuracy, especially in cases where prior knowledge of the binding site is unavailable.
2. The model outperforms existing algorithms, including DiffBindFR and PUResNet, in terms of docking accuracy and pocket-ligand intersection rates across diverse datasets, such as PoseBusters, Astex Diverse Set, and BU48.
3. RAPID-Net adopts an ensemble-based approach, leveraging five independently trained models to improve the stability and reliability of pocket predictions, addressing issues of precision and recall observed in previous algorithms.
4. Unlike traditional methods that focus on geometric properties, RAPID-Net emphasizes docking performance and guides docking software to generate more reliable results by focusing on well-defined search grids.
5. One key innovation of RAPID-Net is its use of soft-labeling and ReLU activation in the output layer, inspired by medical image segmentation techniques, which enables better differentiation between pocket regions and their boundaries.
6. The model’s capability to identify remote binding sites, including allosteric sites, exosites, and flexible regions, makes it a powerful tool for discovering new druggable targets and advancing therapeutic development.
7. In blind docking scenarios, RAPID-Net outperforms AlphaFold 3 in accuracy, demonstrating its potential to enhance docking workflows without the need for large-scale computational resources.
💻Code: https://t.co/oG0y1tYLpK
📜Paper: https://t.co/FNtslcSTZS
#ComputationalBiology #DrugDesign #DeepLearning #MolecularDocking #MachineLearning #Bioinformatics #AI #RAPIDNet #Docking
Come join our Department of Physics and *Astronomy*!
Applications due Nov 30 to be our newest assistant professor of Astronomy. https://t.co/EIuui6a9sj
@waynestatephys@WayneStateCLAS
@mglerner I diversify my investments and I anti-diversify my data. I give one company everything and hope they don’t have a breach. I am not willing to go without modern conveniences, but keeping it to one company is probably better than sharing it with everyone?
@DanielRoyal@_heyadrienne I kind of like it when urban planners design flex space to be used however the community desides in the moment. Sometimes we need more parking, sometimes it is rush hour or construction so temporary no-parking signs are easy to install
Had an amazing time at the @FASEBorg lipid droplet conference!! Thank you to all of the speakers and attendees for making this such an exceptional meeting. See you in 2024!!! @BobFarese #LDSRC22#LDSRC24#LipidTime#LipidDroplet
Want to learn more about the coupling of lipid phase separation, nanoscale membrane curvature, and single-lipid diffusion? We have the paper for you!
https://t.co/F3JLPoWHyi
@JavanainenM@waynestatephys@WayneStateCLAS@waynestate
It's not too early to start thinking about the Mass Poor People and Low-Wage Workers' Assembly & Moral March on Washington.
June 18, 2022
Take action here: https://t.co/AUKoN4o0ya #PoorPeoplesCampaign