Challenge #6

Hit Identification
Method type (check all that applies)
Deep learning
High-throughput docking
Description of your approach (min 200 and max 800 words)

To identify novel ligands for the Triple Tudor Domain of SETDB1 we will consider two starting PDB structures: 8UWP (complexed with the MR46747 ligand), and 7CJT (complexed with (R,R)-59 ligand). Both structures will be subjected to molecular dynamics simulations to assess the plasticity of the binding site, especially regarding the gate to the Kac binding site that, if opened, could present a new strategy to discover new chemistry by extending the one/two aromatic cage site (especially for 7CJT that has been solved with such site already available). Dominant conformations of both target structures will be separately subjected to structure-based virtual screening of the Enamine REAL database (previously prefiltered to remove non-“greenlighted” compounds, based on CACHE traffic light scoring method, and prepared with OpenEye tools), using Deep Docking in combination with either Autodock-GPU and/or gnina (depending on retrospective performance on known ligands) to accelerate the entire process. Final top scoring molecules will be rescored with a target-tunable machine learning scoring function that we have developed in-house to improve docking of solvent exposed pockets, and consensus scored. Simulation Plus will be used to remove non-soluble compounds from the top ranked list. Expert visual inspection will be used to prioritize compounds for testing. An equal number of compounds will be selected from the virtual screens of both structures.

What makes your approach stand out from the community? (<100 words)

Our expertise in simulating complex biological systems will be instrumental in characterizing the plasticity of the binding sites and select the best conformations from fruitful virtual screening. The Deep Docking pipeline has one of the best track-record for hit finding, having been prospectively validated across multiple targets and secured the first place in the 1st CACHE Challenge, as well as being admitted to the second stage of CACHE3.

Method Name
Deep Docking
Commercial software packages used

OpenEye for ligand preparation (QUACPAC, OEOMEGA)

Free software packages used

Gromacs, Autodock-GPU, gnina, Deep Docking, rdkit

Relevant publications of previous uses by your group of this software/method
  1. Brezinova, M., et al (2024). Identification of high-affinity secondary nucleation inhibitors of Aβ42 aggregation from an ultra-large chemical library using Deep Docking. Research Square. doi: https://doi.org/10.21203/rs.3.rs-4512167/v1
  2. Gutkin, E., et al (2024). In silico screening of LRRK2 WDR domain inhibitors using deep docking and free energy simulations. Chemical Science.
  3. Gentile, F., et al (2022). Artificial intelligence–enabled virtual screening of ultra-large chemical libraries with deep docking. Nature Protocols, 17(3), 672-697.
  4. Gentile, F., et al (2021). Automated discovery of noncovalent inhibitors of SARS-CoV-2 main protease by consensus Deep Docking of 40 billion small molecules. Chemical Science, 12(48), 15960-15974.