Features
Determine the synthesis priority
- Hyper Binding uses artificial intelligence to predict drug-protein binding energy.
- The stronger the prediction value (binding score), the higher the probability of showing activity.
- A molecule with a binding score of -10 kcal/mol is more likely to be active compared to a molecule with -8 kcal/mol.
- This process can be easily performed with just a few clicks.
Easier and more accurate than docking
Millions of drug-protein data and binding structures are used, making it more powerful. Unlike docking, it can be easily used by experimental researchers as well.
|
|
Hyper Binding |
Docking |
|
Number of parameters |
~Several miliions |
10-20 |
|
Core technology |
Physics informed deep learning |
A few simple physics equation |
|
accessibility |
Experimental researchers can also use easily. |
Mainly for CADD experts |
|
Accuracy |
High |
Low |
Show high correlation with experiments
Hyper Binding shows a higher correlation with experimental values compared to conventional docking and other deep learning models.
Visualize the three-dimensional binding structure
- With just one click, you can visualize the three-dimensional structure and analyze interactions between ligands and proteins.
- You can obtain ideas for molecular design to improve biological activity.
Artificial Intelligence Model in Hyper Binding
- Hyper Binding uses a physics informed deep learning model.
- By combining physical principles with deep learning, it shows high predictive performance even with limited data.
- Hyper Binding is continuously becoming more accurate by refining physical principles and learning from an increasing amount of data.
