DeepCalici & DeepCalici Plus

DeepCalici is a module that uses a four-dimensional convolutional neural network (4DCNN) AI model to predict binding affinities between proteins and ligands in micromoles (µM), simplifying subsequent experiment planning.

DeepCalici Plus is an enhanced version of DeepCalici.
It provides more accurate predictions, more efficient GPU memory usage, and faster performance than the original DeepCalici module.

Predecessor modules: AI-Dock, AI-Dock Plus

Module usage demonstrations

Select ligand compounds for AI prediction.

When the DeepCalici/DeepCalici Plus module is complete, the predicted binding affinity (µM) values are sorted in ascending order.

Result screen

After completion, the user can press the Download button to download all results for further analysis.
The download folder contains the following files
- Dock_ligand_pdbqt: The three-dimensional structure of the ligand in its bound state with the target protein.
- Org_ligand_pdbqt: the original ligand before binding to the target protein
- Protein_pdbqt: Protein file and related files for docking
- Results: Excel file containing binding energies, binding affinities, and SMILES of selected ligands