CACHE

CRITICAL ASSESSMENT OF COMPUTATIONAL HIT-FINDING EXPERIMENTS

DONATE

  • About
    • WHAT IS CACHE
    • Conferences
  • CACHE News
  • CHALLENGES
    • Challenge #1
      • Announcement
      • Computation methods
      • Preliminary results
      • Final results
    • Challenge #2
      • Announcement
      • Computation methods
      • Preliminary results
      • Final Results
    • Challenge #3
      • Announcement
      • Computation methods
      • Preliminary results
      • Final Results
    • Challenge #4
      • Announcement
      • Computation methods
      • Preliminary results
    • Challenge #5
      • Announcement
      • Computation methods
    • Challenge #6
      • Announcement
    • FAQ
  • PUBLICATIONS
  • CONTACT

Challenge #1

Hit Identification
Method type (check all that applies)
Machine learning
Hybrid of the above
training ML with LigandEvolution results for QSAR
Other (specify)
Evolutionary optimization - RosettaLigandEvolution
Description of your approach (min 200 and max 800 words)

Due to the lack of available screening data for molecules targeting the WD40 repeat, we will start by searching the Enamine Make-On-Demand library with our own novel evolutionary optimization algorithm RosettaLigandEvolution (LigEvol) and ligand docking with RosettaLigand. Both are part of the Rosetta software suite and developed by us. RosettaLigand optimizes ligandprotein complexes with complete small molecule, side chain, and backbone flexibility to capture all aspects for determining binding conformation and affinity. LigEvol reliably optimizes ligands through fragment search in combinatorial libraries like Enamine and identifies local molecule minima. Not only does it detect promising compounds with ease, but it also strictly remains inside Enamine make-on-demand space. We aim at sampling 100,000 molecules and expect, based on previous experiments, to derive 1,000 potential hit candidates. Next, we will seed LigEvol with promising compounds to further explore the chemical space close to areas of high interest and train a QSAR and an applicability domain (AD) model with our own BioChemicalLibrary (BCL) software on the predictions from RosettaLigand. The QSAR model allows us to screen the entire database to identify areas of chemical space of higher priority, and we can reiterate LigEvol while seeding it with reagents that are closest to areas of interest. At this point, we expect to have a large number of diverse compounds with a high predicted affinity, covering most if not all of the chemical space of interest. The final selection will be picked through the appliance of multiple BCL filters for drug-likeness, solubility, and more. We will ensure a high diversity of scaffolds through clustering. Finally, our medicinal chemists will manually inspect the compounds and prioritize 100 candidate molecules, which can be excellent starting points for hit optimization, for testing.

Method Name
RosettaLigand, RosettaLigandEvolution, BCL
Free software packages used

Rosetta Suite, Biology and Chemistry Library (BCL), PyMOL, Python– RDKit and other free libraries

Relevant publications of previous uses by your group of this software/method

1) Schüß C, Vu O, Schubert M, Du Y, Mishra NM, Tough IR, Stichel J, Weaver CD, Emmitte KA, Cox HM, Meiler J, Beck-Sickinger AG. Highly Selective Y4 Receptor Antagonist Binds in an Allosteric Binding Pocket. J Med Chem. 2021 Mar 11;64(5):2801-2814. doi: 10.1021/acs.jmedchem.0c02000. Epub 2021 Feb 17. PMID: 33595306. 2) Bertron JL, Duvernay MT, Mitchell SG, Smith ST, Maeng JG, Blobaum AL, Davis DC, Meiler J, Hamm HE, Lindsley CW. Discovery and Optimization of a Novel Series of Competitive and Central Nervous System-Penetrant Protease-Activated Receptor 4 (PAR4) Inhibitors. ACS Chem Neurosci. 2021 Dec 15;12(24):4524-4534. doi: 10.1021/acschemneuro.1c00557. Epub 2021 Dec 2. PMID: 34855359. 3) Mendenhall J, Brown BP, Kothiwale S, Meiler J. BCL::Conf: Improved Open-Source KnowledgeBased Conformation Sampling Using the Crystallography Open Database. J Chem Inf Model. 2021 Jan 25;61(1):189-201. doi: 10.1021/acs.jcim.0c01140. Epub 2020 Dec 22. PMID: 33351632; PMCID: PMC8130828. 4) Butkiewicz M, Rodriguez AL, Rainey SE, Wieting J, Luscombe VB, Stauffer SR, Lindsley CW, Conn PJ and Meiler J; Identification of Novel Allosteric Modulators of Metabotropic Glutamate Receptor Subtype 5 Acting at Site Distinct from 2-Methyl-6-(phenylethynyl)-pyridine Binding. ACS Chem Neurosci. 2019;10:3427-36. 5) Vu O, Mendenhall J, Altarawy D, Meiler J. BCL::Mol2D-a robust atom environment descriptor for QSAR modeling and lead optimization. J Comput Aided Mol Des. 2019 May;33(5):477-486. doi: 10.1007/s10822-019-00199-8. Epub 2019 Apr 6. PMID: 30955193; PMCID: PMC6824857. 6) Brown BP, Mendenhall J, Meiler J. BCL::MolAlign: Three-Dimensional Small Molecule Alignment for Pharmacophore Mapping. J Chem Inf Model. 2019 Feb 25;59(2):689-701. doi: 10.1021/acs.jcim.9b00020. Epub 2019 Feb 12. PMID: 30707580; PMCID: PMC6598199. 7) Schubert M, Stichel J, Du Y, Tough IR, Sliwoski G, Meiler J, Cox HM, Weaver CD, Beck-Sickinger AG. Identification and Characterization of the First Selective Y(4) Receptor Positive Allosteric Modula-tor. J Med Chem. 2017;60:7605-12. 8) Butkiewicz M, Wang Y, Bryant SH, Lowe EW Jr, Weaver DC, and Meiler J. HighThroughput Screening Assay Datasets from the PubChem Database. Introduction Chem Inform. 3.1 (2017). doi: 10.21767/2470-6973.100022 9) Sliwoski G, Schubert M, Stichel J, Weaver D, Beck-Sickinger AG, Meiler J. Discovery of SmallMolecule Modulators of the Human Y4 Receptor. PLoS One. 2016;11:e0157146. 10) Butkiewicz M, Lowe EW, Mueller R, Mendenhall JL, Teixeira PL, Weaver CD, and Meiler J. Benchmarking ligand-based virtual High-Throughput Screening with the PubChem database. Molecules 18.1: 735-756 (2013). doi: 10.3390/molecules18010735 11) Nguyen ED, Norn C, Frimurer TM and Meiler J. Assessment and challenges of ligand docking into comparative models of G-protein coupled receptors. PLoS One. 2013;8:e67302.

Hit Optimization Methods
Method type (check all that applies)
Machine learning
Hybrid of the above
training ML with experimental results and LigandEvolution results for QSAR
Other (specify)
Metropolis-Monte-Carlo-Optimization (RosettaMMC)
Description of your approach (min 200 and max 800 words)

In the optimization step, we will append our aforementioned pipeline with a multi-pronged approach to focus on scaffold explosion and remain nearby hit scaffolds but deeply explore the structure-activity relationship (SAR). We will build a new applicability domain (AD) incorporating substructure sensitive features derived from our confirmed hits. Furthermore, we might utilize a newly developed deep neural network to investigate the interaction pattern. These training models can further improve the scoring method, adapt the peculiarities of the protein-ligand interactions, and evaluate the docking results in more detail. Including BCL Alchemy functionalities in Rosetta will utilize the Enamine reaction set and further fragment libraries to mutate our best hit scaffolds. To further expand the chemical space in collaboration with our institute's medical chemistry synthesis group, we will build a library of potential alterations that they can carry out efficiently. These rules of replacements and reactions will define our own synthesizable molecule space, which we will test through ligand docking and QSAR models. Furthermore, an in-house reaction-based combinatorial library will be applied in iterative loops of RosettaLigand and BCL. Promising compounds will be synthesized in our facilities and sent to the challenge organizers. Additionally, we will again seed LigEvol with promising compounds, now based on confirmed hits, to search again through chemical space and remain focused on confirmed areas of interest. We will add another method to search through Enamine make-on-demand space utilizing MonteCarlo-Metropolis optimization (RosettaMMC) of single fragment replacement to explore local minima of hit compounds. Again, the candidates yielded by this mixture of approaches will be filtered by BCL functionality and valuable feedback from medicinal chemists to form a final selection of 100 compounds.

Method Name
RosettaLigand, RosettaLigandEvolution, RosettaMMC and BCL
Free software packages used

Rosetta Suite, Biology and Chemistry Library (BCL), PyMOL, Python– RDKit and other free libraries

Relevant publications of previous uses by your group of this software/method

See above

Cache

All rights reserved
v5.47.19.49

Footer first

  • Login
  • Applicant Login
  • Terms of Participation
  • Privacy Policy
  • FAQ
  • Docs
This website is licensed under CC-BY 4.0

Toronto website development by Rebel Trail