When using 3D Structure Based Settings in a generator, the generated molecules are docked in the specified target protein binding pocket and optimized with respect to the Docking Score and the Contact Score.
The contact score is a proprietary score developed by Iktos. It has been successfully deployed in multiple projects to identify novel scaffolds with good activity. It rewards the presence of key interactions between the generated molecule and the protein and ranges in value from 0 to 1. Interactions are given a weight by the user that is proportional to their relative importance:
- If interactions with a strong weight are present between the generated molecule and the protein pocket, the molecule is given a high contact score (close to 1).
- If interactions with a strong weight are absent, the molecule is given a low contact score (close to 0).
- Interactions with a low weight have little impact on the final contact score.
NOTE: weak H-bond interactions have a global weight set at half of the other kinds of interactions, so the custom weight values for H-Bond weak interaction will be divided by 2.
Illustration of the Contact Score. The size of spheres is a function of the importance given to each interaction.
Interaction weights
The Contact Score is a function of the weights given to ligand-protein interactions by the user during the setting of the structure-based parameters (for more information on this setting, see the corresponding page in the documentation).
Interaction weight represents how important the interaction is compared to other interactions of the same type. It is used by the Contact Score to compute a ratio between the number of interactions detected in a specific pose, versus the number of interactions formed by a reference ligand (or specified by the user).
References
- , , , , , , , Exploring isofunctional molecules: Design of a benchmark and evaluation of prediction performance, 2023, Mol. Inform., 42, 2200216.
- C. Bissantz, B. Kuhn, M. Stahl A Medicinal Chemist's Guide to Molecular Interactions, 2010, J. Med. Chem., 53, 5061-5084.
- R. Ferreira de Freitas, M.Schapira A systematic analysis of atomic protein-ligand interactions in the PDB, 2017, Med. Chem. Commun., 8, 1970-1981.
- E. Nittinger, T. Inhester, S. Bietz, A. Meyder, K. T. Schomburg, G. Lange, R. Klein, M. Rarey Large-Scale Analysis of Hydrogen Bond Interaction Patterns in Protein-Ligand Interfaces, 2017, J. Med. Chem., 60, 4245-4257.
- N. K. Shinada, A. G. de Brevern, P. Schmidtke Halogens in Protein-Ligand Binding Mechanism: A Structural Perspective, 2019, J. Med. Chem., 62, 21, 9341-9356.
- B. Kuhn, E. Gilberg, R. Taylorn J. Cole, O. Korb How Significant Are Unusual Protein-Ligand Interactions? Insights from Database Mining, 2019, J. Med. Chem., 62, 22, 10441-10455.