Code, not Drugs: The LLM Revolution in Computational Drug Design Process before product

Code, not Drugs: The LLM Revolution in Computational Drug Design
open.substack.com/pub/medchema...

The Latest in Automated Analog Generation

An LLM-based chemical compound generator prompts some thoughts about what these might be good for:

Rapamycin's Secrets

Never assume that you know all about what any drug molecule is doing in the body - where it's going, what it's binding to. A new example:

Glycosidase-activated prodrugs of a cytotoxic iron chelator for targeted cancer therapy New glycoside-prodrugs based on the iron chelator deferasirox were designed. Selective enzymatic activation by glycosidases was observed within 24 hours, accompanied by cancer cell-selective…

🔓Be sure to read this recent #OpenAccess work by Nils Metzler-Nolte & co: "Glycosidase-activated prodrugs of a cytotoxic iron chelator for targeted cancer therapy"👀

pubs.rsc.org/en/content/a...

📍@ruhr-uni-bochum.de 🧪

Small molecule drug discovery for Ebola virus disease Known for its widespread outbreaks, including the 2013–2016 epidemic that infected almost 29 000 individuals and resulted in approximately 11 300 deaths, Ebola virus (EBOV) and related filoviruses…

🔓Read this #OpenAccess Review by @moorelab.bsky.social and colleagues which explores the historic and recent progress in the discovery of small molecule drugs for Ebola virus disease⬇️

pubs.rsc.org/en/content/articlelanding/2025/md/d5md00533g

📍University of Illinois Chicago 🧪

RSC Medicinal Chemistry outside front cover for vol 16, issue 10.

On the front cover of issue 10 is work by Yosuke Demizu and colleagues on analysing the linker rigidity and improve intracellular behaviour of three novel PROTACs targeting H-PGDS, read more here👇

pubs.rsc.org/en/content/a...

📍National Institute of Health Sciences, Kanagawa 🧪

All stirred up: chemical engineers refute claims that ‘stirring doesn’t matter’ Failing to mix reactions in heterogeneous or industrial systems could cause numerous issues and might even be dangerous, preprint claims

Earlier this year, a paper came out that suggested for many organic reactions stirring has little effect on reactions rates. Now, a group has released a preprint arguing that stirring remains critical for reproducibility, selectivity and scalability.

Finally out!

We have released a #medchem exercise (as a #KNIME workflow) in which you can apply isosteric replacement to a given scaffold and enumerate all the suggestions.

The files required to run it are available on wcair.dundee.ac.uk/training/tra... or hub.knime.com/s/n4K2My0xJX...

#chemsky

Today, we’re spotlighting Ryushi Seo and his Just KNIME It! Season 3 data journey.

Just KNIME It! is the perfect way to build your KNIME skills — whether you're just starting or leveling up.

Season 4 of #JustKNIMEIt! starts this week, register 👉 knime.com/just-knime-i...

Vernalis v1.37.0 Release - Changes to PDB Connector Nodes We have just released v1.37.0 of our community plugin which brings in 1 new nodes, deprecates one node and adds new features to one existing node NEW Node - PDB Connector Hit Count This node does ex...

v1.37.0 of the Vernalis #KNIME community contribution is now available for AP v5.4, with v5.3/v5.2 to follow. New PDB Connector features including #AlphaFold access in the Query executor, new Hit Count node, and the PDB SMILES Query node is retired - see forum.knime.com/t/vernalis-v... for details

Exploring Chemical Databases with KNIME Low-code workflows for cheminformatics

🚀 New to #cheminformatics? Check out #ChacMool — a set of #KNIME workflows by Carlos D. Ramírez Márquez & José L. Medina Franco that make #chemical #library analysis (descriptors, scaffolds, PAINS, fingerprints) easy & accessible! without coding! 🔬💻

📌 #READ → medium.com/low-code-for...

🚀 Registration is open for the #Fribourg #WinterSchool in #DataAnalytics & #MachineLearning, Feb 2–13 2026.
📍Hybrid: at Fribourg University or online
🔍Topics: data analytics, predictive/causal machine learning, deep learning
💻 Software: Python, R, Julia, Knime
👉 Sign up: www.unifr.ch/appecon/en/w...

Save the date! Next year's European #RDKit UGM will take place from 16-18 September in Darmstadt, Germany.
Pre-registration will open in early 2026.

Exploring the Power of R Graphics with KNIME: A Collection of Examples Low-code data visualizations to enrich your data storytelling

If you are interested in more R graphics with knime I can offer this collection
medium.com/p/9241e033e4ac

Great work @cedricgraebin.bsky.social ! Reminds me of the early workflows on myExperiment back in the early days of #KNIME (gulp! 12y ago  😬)
www.myexperiment.org/workflows/26...

diving deep on tools such as KNIME and its chemistry extensions - you can do some nice things with it, and we have been developing some tools with the platform as part of the free resources available at wcair.dundee.ac.uk/training/tra....

PubChem Fast Structural Search – cgraebin001 This component implements the Fast synchronous search engine that is part of the PubChem Power User Gateway (PUG) REST API to search for molecules in the PubCh…

Once more for the #knime #chemistry folks - I implemented the "fast" synchronous search for chemical structures in the PubChem database into a KNIME component.
hub.knime.com/s/bU6bz0lBve...

#chemsky

🔍 Webinar Highlights:
• Molecular descriptor calculations
• 3D coordinate generation
• Molecular fingerprinting techniques
• Structural pattern identification with SMARTS syntax
• Data analysis of calculated descriptors
• Seamless integration with KNIME and Python

It would work the same way for KNIME or DataWarrior. KNIME can read XLSX files directly and can interpret structures submitted as SMILES, MOL, etc, if you add chemistry-enabled extensions (RDKit, Indigo, CDK). DataWarrior can import CSV files if I remember well.

p.s.: both are free and open-source

KNIME, the Konstanz Information Miner -- I had a lot of fun with this at University, and want to revisit at some point. It seemed like the most powerful tool nobody ever talks about.

ChEMBL structure search – cgraebin001 This node takes a table with structures (either SDF/MOL, RDKit or SMILES format) and runs a structural search using the ChEMBL REST API web services. The searc…

For the #knime #chemistry folks out there:

I am releasing some of the things I made for KNIME as standalone components. The first one is a refined, visualization-enabled ChEMBL chemical structure search. Hope it is useful! - hub.knime.com/cgraebin001/...

#ChemSky

ChEMBL Activity Search by ID – cgraebin001 This component uses the ChEMBL REST webservices API to retrieve biological activities reported to a compound represented by its ChEMBLID. If you don't have the…

One more for the #knime #chemistry folks around:
My last custom component for KNIME could find if a given chemical structure can be found on the #ChEMBL database. What about biological activities?
This second component retrieves does this using the ChEMBLID as query: hub.knime.com/s/22f3jdJjRK...

All of that using an open-source software exploring the potential of a open-access database!

Using the other #knime components I published, you can:
- from the chemical structure (SMILES), get the molecule ID
- from the molecule ID, get the biological activities reported for that compound and the document ID where they are
- from the document ID, get the reference!

ChEMBL get Document by ID – cgraebin001 This component retrieves information about a document (Publication, Book, Patent, Dataset) in the ChEMBL using its identifier number (ChEMBLID) as query. Retri…

Still on the #knime #chemistry topic: still exploring the #ChEMBL REST API, I made another component that can retrieve a document data (authors, journal title, year, volume, pages, DOI, PubMED ID, Patent ID) from the correspondent Document ChEMBLID

hub.knime.com/s/8a1Txmjb_R...

WCAIR training workflow 01: Reaction Enumeration – cgraebin001 Solutions for data science: find workflows, nodes and components, and collaborate in spaces.

The second #knime #chemistry workflow is more a customization of sorts - if you want to enumerate your own custom chemical reactions instead of the amide coupling presented in the first exercise (this is why the title is the same, inclusive):

hub.knime.com/s/AarwRqxiW_...

#ChemSky

WCAIR training workflow 01: Reaction Enumeration – cgraebin001 Solutions for data science: find workflows, nodes and components, and collaborate in spaces.

I also made some exercise workbooks using #knime and #chemistry (focused on Drug Discovery). One of them explains how to enumerate a chemical reaction using KNIME - and what you can make of the data you get back: hub.knime.com/s/fw7J8YeZwi...

#ChemSky

PubChem PUG REST Search by Chemical Structure – cgraebin001 This component makes use of the NCBI PubChem PUG REST API to search for compounds in the database with a given chemical structure (provided in the SMILES forma…

Still going (strong?) on the #knime #chemistry custom components, I made one that looks for chemical structures on the PubChem database and retrieves back the PubChem CID if the compound was found there: hub.knime.com/s/SQrBVFVToy...

#ChemSky

There's a new #RDKit blog post introducing some new functionality that I'm really excited about: doing efficient substructure and similarity searches in very large chemical libraries:
greglandrum.github.io/rdkit-blog/p...
#ChemSky