GNINA Tutorial: CNN Scoring Explained (rescore vs refinement vs all)

By BioDockify Computational Research Team · September 3, 2026 · Tutorials

Use GNINA CNN scoring correctly: rescore versus refinement versus all modes, reading CNNscore and CNNaffinity, affinity mapping and the ranking pitfalls to avoid.

Read the complete article with tables, code and references on BioDockify.

Key References

• McNutt AT et al. GNINA 1.0: molecular docking with deep learning. J Cheminform. 2021;13:43. doi:10.1186/s13321-021-00556-0
• Ragoza M et al. Protein-ligand scoring with convolutional neural networks. J Chem Inf Model. 2017;57:942-957. doi:10.1021/acs.jcim.7b00240
• Open Babel documentation: SDF data fields, open-babel.readthedocs.io.

Scope & Limitations

CNN models were trained on PDB-derived complexes; performance on unusual pockets or metalloproteins should be validated locally