Deep MedChem

Online search: EnamineStore: enaminestore.com Chemspace: chem-space.com/searchZINC-20: zinc20.docking.org ZINC-22: cartblanche22.docking.org arthor.docking.org/index.html CHEESE: cheese-new.deepmedchem.com Software Packages: InfiniSee by BioSolveIT: biosolveit.de/infiniSee_xREAL by Cresset: cresset-group.com/software/flare/ Giga-Search by Molsoft: molsoft.com/giga-search.ht… Hyperspace and Pharos-3D by Alipheron: alipheron.com Orion (FastROCS ) by OpenEye: eyesopen.com/orion/database… PharmScreen by Pharmacelera: pharmacelera.com/pharmscreen/ Maestro Suite by Shrödinger: schrodinger.com/platform/produ… Synthon Searching on RDKit: greglandrum.github.io/rdkit-blog/pos…

DM us for the slides.

Missed our webinar on AI-powered chemoinformatics workflows? The recording is now live youtu.be/dPDFrZi2yaQ We covered practical use cases around molecular embeddings, CADD workflows, and cloud/local deployment options. #Cheminformatics #DrugDiscovery #CADD #AIinDrugDiscovery

Google Slides: #slide=id.p" target="_blank" rel="nofollow noopener">docs.google.com/presentation/d… Code: github.com/Deep-MedChem/9… Challenge app: shiny.imtm.cz/apps/polishp/9…

We won the 9ADD DYRK1B Virtual Screening Challenge! 🏆 (Team from Deep MedChem with @miroslavlzicar, Kryštof Škach, Alexander Korolyov) Our Presentation: youtube.com/live/7cuUKSz28…

Benchmarks of SynthonGPT from our CTO!

SynthonGPT Prototype Launch! Try it out at: synthongpt.mireklzicar.com

iFrame embedding for CHEESE is now live! You can drop CHEESE Search as a widget into your webpage, blog, doc, CMS or a dashboard with a single html snippet. Responsive by default, sandboxed for safety, and easy to configure via URL params. mireklzicar.com/blog/cheese_if…

Do your shape-similar hits still hold up when you actually align conformers on the query? I built a demo to explore that 👉 cheese-conformers.mireklzicar.com

We are super happy to see CHEESE Molecular Similarity Search being used in another drug discovery project, this time against SARS-CoV-2: onlinelibrary.wiley.com/doi/epdf/10.10… You can try CHEESE Search here: cheese-new.deepmedchem.com

@BiologyAIDaily Dear all, try updated UI of the CHEESE Search here cheese-new.deepmedchem.com. We want to hear your feedback!

CHEESE: 3D Shape and Electrostatic Virtual Screening in a Vector Space 1. Introducing CHEESE: This framework revolutionizes ligand-based virtual screening by embedding 3D shape and electrostatic properties into a searchable vector space, allowing approximate nearest neighbor search and enabling rapid identification of similar molecules across large databases. 2. Speed and scalability: CHEESE achieves state-of-the-art performance on benchmarks like DUD-E and LIT-PCBA, with search speeds up to three orders of magnitude faster than traditional methods like ROCS, while reducing computational costs by six orders of magnitude. 3. High computational efficiency: By leveraging vector indexing and approximate nearest neighbor search, CHEESE can screen the Enamine REAL 5.5B database in seconds on a standard 4-core CPU with minimal RAM, making it highly accessible for large-scale applications. 4. Advanced shape and electrostatic metrics: CHEESE uses Shapesim and Espsim to capture molecular similarity, achieving improved accuracy in identifying bioactive compounds with minimal false positives, particularly useful in bioisostere screening and drug discovery. 5. Outperforming traditional methods: CHEESE consistently surpasses ROCS and traditional docking in retrieving shape- and electrostatically similar compounds, especially in multi-ligand virtual screening, where it shows a threefold increase in enrichment factors. 6. Robust generalization: Trained on the ZINC dataset, CHEESE demonstrates strong out-of-distribution generalization, effectively adapting to diverse chemical spaces like DrugBank, PubChem, and specialized natural product databases without additional training. 7. Potential applications: CHEESE’s framework supports scalable, fast, and accurate ligand-based virtual screening, offering promising applications in high-throughput drug discovery, bioisostere identification, and molecular optimization workflows. @miroslavlzicar @deepmedchem 💻Code: github.com/Deep-MedChem/c… 📜Paper: doi.org/10.26434/chemr… #VirtualScreening #DrugDiscovery #3DChemistry #Electrostatics #ComputationalChemistry

Revolutionize Drug Discovery with CHEMRIYA™ & CHEESE Search --- otavachemicals.com/products/chemr…

🔬 What does a chemical space look like when visualized? In this prototype, we mapped 10 million molecules from ZINC using UMAP, creating a high-density 2D representation. While the current version of CHEESE Explorer (🔗 lnkd.in/dMgDgxXz) handles up to 100K points, ...

Molecular similarity search with CHEESE API --- fast and now even easier to integrate with a new KNIME node: hub.knime.com/s/JgIqQIeNalQT…

What Can Neural Network Embeddings Do That Fingerprints Can’t? "Fingerprints may retrieve molecules with electrostatically similar fragments or motifs; however, they cannot capture the overall ‘big picture.’..." Read the full blog: 👉 deepmedchem.com/articles/what-…

The preprint about CHEESE Search is here: chemrxiv.org/engage/chemrxi…

CHEESE Search Webinar on Dec 5, 4:30 CET, 10:30 ET---we will dive into the CHEESE Search preprint (technology, benchmarks, and more) on ChemRxiv: Make sure to register: linkedin.com/events/cheese-…

CHEESE is tackling the speed-accuracy tradeoff in ligand-based virtual screening by distilling 3D shape and electrostatic similarity metrics into a vector space. Achieving SOTA on standard LBVS benchmarks. chemrxiv.org/engage/chemrxi…

Explorer 🌐 lnkd.in/eEBRpgi6 A new app which allows to interactively navigate chemical spaces. #chemistry #ai #drugdiscovery

It is this feeling your assets could proceed faster through the drug discovery pipeline... And then there is our expertise---read about it more in our newsletter. deepmedchem.com/newsletters/de…