Anyone Can Dock: an online molecular docking tool for everyone

Molecular docking predicts how small molecules bind to protein targets, but performing a docking calculation has traditionally required software installation, command-line scripting, and manual molecular preparation. Anyone Can Dock is a ready-to-use online platform developed to reduce the technical and procedural burden of conventional ligand-protein docking through four complementary interfaces. First, the Streamlit web application provides a simple four-step workflow: users enter a protein, specify the target ligands, initiate docking with a single click, and view the resulting predictions. Second, the Google Colab notebook supports four docking engines, AutoDock Vina 1.2.7, VinaXB, GNINA, and Vina-GPU, enabling batch virtual screening, teaching, and method development. Third, Anyone Can Dock GPT provides conversational workflow guidance and docking-result interpretation through natural language interaction in ChatGPT and a command-line interface (CLI) version is also available via PyPI. These interfaces streamline several routine yet technically important steps in docking workflows, including receptor preparation for metal-containing structures, pH-aware protonation-state assignment using pK a NET Cloud+ with Dimorphite-DL for initial microstate enumeration, chemically consistent pose export following PDBQT conversion, and built-in RMSD-based redocking validation when a co-crystallized reference ligand is available. The web platform is freely available at https://nyelidl.github.io/anyone-docking/ . Scientific Contribution Anyone Can Dock addresses three practical barriers in accessible molecular docking: metal-preserving receptor preparation that preserves selected metal ions and heme-containing active sites during PDBQT conversion, an integrated pK a NET Cloud+ workflow for pH-aware ligand microstate generation and selection, and SMILES-template-based bond-order reconstruction of docked poses for visualization and downstream cheminformatics. RDKit standardizes the input and handles salts, optional experimental pK a information is incorporated when available, and RDKit tautomers are enumerated and filtered before Dimorphite-DL performs initial pH-dependent microstate enumeration for each accepted tautomer. pK a NET Cloud+ then supplements, evaluates, scores, and ranks the candidates and selects the final state used for docking. These components are presented as preparation and usability features rather than demonstrated improvements in Vina scoring or sampling. Metal preservation, pH-aware protonation, and bond-order correction are engineering improvements that enhance usability, robustness, and interpretability; none is claimed as a demonstrated improvement in docking score or pose RMSD. The hosted Streamlit/AutoDock Vina interface requires no registration; Google Colab and the optional Anyone Can Dock GPT guidance layer require the relevant third-party accounts. For clarity, registration-free access refers to the hosted Streamlit/AutoDock Vina workflow. Google Colab use depends on Google-account access, and Anyone Can Dock GPT depends on the external conversational-AI service through which it is accessed.