DeepPurpose-MD: An End-to-End Colab-Based Ligand Discovery Pipeline Integrating Docking, Molecular Dynamics, and Deep Learning

We present DeepPurpose-MD, a unified and fully executable pipeline for ligand discovery that seamlessly integrates deep learning–based target screening, molecular docking, and molecular dynamics (MD) simulation—all within a Google Colab environment. Designed for modularity and automation, this workflow enables users to go from SMILES and FASTA sequences to binding affinity predictions, docking poses, and nanosecond-scale MD analysis in under three hours. The pipeline features a customized fork of DeepPurpose enhanced with ProtTrans embeddings and PCA compression, achieving a Pearson correlation of 0.80 on the BindingDB COVID dataset. With GPU-accelerated simulation using OpenMM and one-click ligand alignment via the Kabsch algorithm, DeepPurpose-MD eliminates the need for manual receptor and ligand preprocessing. This tool bridges wet-lab and dry-lab workflows and is ideal for both novice and advanced users in cheminformatics, structural biology, and AI-assisted small molecule design. This work describes an in-silico computational pipeline and does not provide medical advice, diagnosis, or treatment recommendations.