Introduction
The NVIDIA Generative Virtual Screening blueprint shows how generative AI and accelerated NIM microservices can design small molecules smarter and faster, saving time and money.
Inform the generation of new molecules optimized for multiple user-defined properties simultaneously and reduce the number of synthesis and validation cycles required to discover a true hit. Screen your generated molecules faster using the fastest implementations of accelerated docking AI models deeply integrated into GPU hardware to save time. You can easily download and deploy this workflow anywhere for secure, reliable drug design. It is freely available through the NVIDIA Developer Program for testing or with licensed enterprise support.
The detailed guide of each NIM can be found below:
AlphaFold2 is used to determine protein structure from a protein sequence.
MolMIM efficiently generates molecular structures optimized for multiple desirable, drug-like properties simultaneously.
DiffDock quickly predicts the binding poses of small molecules and their protein targets.
Here, we will show you how Docker Compose can be used to build a workflow from this blueprint and interact with them via the jupyter-notebook Web-UI.