drug-discovery-pipeline

An end-to-end computational drug discovery pipeline automates hit discovery by orchestrating GenMol for molecule generation, DiffDock for molecular docking, and Boltz2 for binding affinity prediction. You provide a target protein sequence to generate and screen novel drug candidates.

Predicting molecular docking and binding affinity using AI involves generating a library of molecules and docking them against a specific protein target. The pipeline uses DiffDock for structural docking and Boltz2 to predict binding affinity, identifying high-potential drug leads.

Yes, you need valid NGC API credentials and access to BioNeMo NIM endpoints to perform sequence and structure-based analysis. These microservice endpoints are required to orchestrate the GenMol, DiffDock, and Boltz2 components in the drug discovery workflow.

Yes, this molecular docking workflow supports both de novo molecule generation and scaffold-based decoration for lead optimization. You can use it to generate novel molecules or optimize existing scaffolds by docking them against a target protein and predicting binding affinity.

The safe-mol and requests dependencies facilitate molecule generation and API interactions but are limited by the requirement for valid NGC API credentials. Without proper access to BioNeMo NIM endpoints, the pipeline cannot execute the structured microservice orchestration needed for screening.