pen(Taka)

Change heterocycle to another isoster with HCIE #memo #cheminformatics #RDKit Aromatic heterocycles are often used drug design as central core of molecule and they are often replaced another isosteric heterocycles for improvement of potency, ADMET or finding new IP space. Th…

Change heterocycle to another isoster with HCIE #memo #cheminformatics #RDKit iwatobipen.wordpress.com/2026/03/08/c...

Work led by industrial PhD student Sofia Larsson and in close collaboration with Susanne Winiwarter and @filipmiljkovic.bsky.social at AstraZeneca.

📄 Paper: chemrxiv.org/doi/full/10....
💻 Code: github.com/tsofiac/drug...
💽 Data: zenodo.org/records/1874...

The new #RDKit blog post is a guest post looking at doing Butina clustering on the GPU using nvMolKit.

greglandrum.github.io/rdkit-blog/p...

Rendering molecular image on Dataframe and Plot with marimo #Memo #RDKit #Cheminformatics As lots of readers know that RDKit has useful functions for coding with jupyter-lab. PandasTools and IPythonConsole is useful because by using these functions rdkit can render mol objects as SVG on…

iwatobipen.wordpress.com/2026/03/07/r...

It’s really surprising that Claude code works really hard :)

Shinagawa station is extremely crowded😱

Avogadro Stories (for 2.0 Release) As we finish up the 2.0 (final!) release, I’d like to feature some stories (and citations) about how Avogadro has shaped research and teaching. So if Avogadro has helped your research, please let me ...

Has @avogadro.cc helped your research? Teaching?

As we prepare Avogadro 2.0 for release this spring, I'm hoping to feature some stories (and citations).

#compchem #opensource #openscience ⚗️

Let me know:
discuss.avogadro.cc/t/avogadro-s...

It’s really difficult to understand for me that mixing quantum chemistry and quantum computing….

Small Molecule Statistics: Growth of Released Small Molecules per Year As a member of the wwPDB, the RCSB PDB curates and annotates PDB data according to agreed upon standards. The RCSB PDB also provides a variety of tools and resources. Users can perform simple and adva...

www.rcsb.org/stats/growth...

Spending lots of time to modify same PDB and finally I could make image with ligting molecule.

still struggling blender :(

Happy to share our article on the PROTAC-Splitter, an ML framework for automated identification of PROTAC substructures, is now finally online in Journal of Cheminformatics! rdcu.be/e4Ulj 🤩

Work led by our amazing senior PhD student Stefano, together with collaborators at AstraZeneca. #chemsky

Blender can generate cool view of molecules :)

Moderna Reversal

This is no way to run a regulatory agency:

終末のフール (集英社文庫) Amazon.co.jp: 終末のフール (集英社文庫) : 伊坂 幸太郎: Japanese Books

なお、私のイチオシは「鋼鉄のウール」の（終末のフール）です
www.amazon.co.jp/-/en/%E7%B5%...

成瀬は天下を取りにいく (新潮文庫 み 73-1) Amazon.co.jp: 成瀬は天下を取りにいく (新潮文庫 み 73-1) : 宮島未奈: Japanese Books

相方におすすめされて一気に読んだ。
自分的には面白かった。
www.amazon.co.jp/dp/410106141...

I don't think that current AI technology can find novel scaffolds which is difficult to design by medicinalchemist but I think it can find solution more efficiently. (IMHO...)

MolVE: An Open-Source Web Platform for Visualizing and Evaluating AI-Designed Molecules to Aid in Prioritization Advances in artificial intelligence and deep generative models have enabled the rapid generation of novel molecular structures for advanced material science and drug discovery. However, the effective evaluation of these candidates still depends, in the end, on expert judgment, which is often fragmented and difficult to scale. This work introduces MolVE, an open-source, web-based platform designed to enable asynchronous, distributed, and collaborative evaluation of molecules by experts. The platform combines secure user authentication, data set management, and interactive 2D/3D visualizations, enabling chemists and pharmacologists to curate, annotate, and evaluate molecules efficiently. Additionally, it offers APIs for seamless integration with programming languages and provides a Python service to run machine learning and deep learning models. MolVE is a comprehensive, ready-to-deploy application built with React, Node.js, Express, and PostgreSQL, and is fully containerized using Docker. This setup enables scalable deployment in both academic and industrial settings.

MolVE: An Open-Source Web Platform for Visualizing and Evaluating AI-Designed Molecules to Aid in Prioritization | Journal of Chemical Information and Modeling pubs.acs.org/doi/10.1021/...

Data-Driven Design of PROTAC Linkers to Improve PROTAC Cell Membrane Permeability Proteolysis-targeting chimeras (PROTACs) are promising next-generation therapeutics for the degradation of disease-associated proteins. However, optimizing the physicochemical properties of PROTACs, p...

The model showed good correlation between data came from PROTAC DB and prediction. But showed moderate correlation between data from their experiments and prediction.
Data-Driven Design of PROTAC Linkers to Improve PROTAC Cell Membrane Permeability | JACS Au pubs.acs.org/doi/10.1021/...

It means ML based linker design approach is useful but there is still room for improvement.

It’s snowing here for the first time in about over 10 years.

You’re welcome ☺️
Thanks for developing useful code!

Split PROTAC molecule to 3 compornents with protac_splitter #RDKit #cheminfomratics #memo #python

Proteolysis-targeting chimeras (PROTACs) are one of the interesting modalities in these days because the modality can engage protein of interest (POI) and E3 ubiquitin ligase and then causes…

A Comprehensive Review of Emerging Approaches in Machine Learning for De Novo PROTAC Design Targeted protein degradation (TPD) is a rapidly growing field in modern drug discovery that aims to regulate the intracellular levels of proteins by harnessing the cell's innate degradation pathways t...

A Comprehensive Review of Emerging Approaches in Machine Learning for De Novo PROTAC Design
arxiv.org/abs/2406.16681

Protap: A Benchmark for Protein Modeling on Realistic Downstream Applications Recently, extensive deep learning architectures and pretraining strategies have been explored to support downstream protein applications. Additionally, domain-specific models incorporating biological ...

Protap: A Benchmark for Protein Modeling on Realistic Downstream Applications
arxiv.org/abs/2506.02052

Improvement in reasoning performance of LLMs over time If you tried using ChatGPT when it first came out and concluded it wasn't much use for a scientific reasoning task, it might be time to try ...

Blogpost that looks at how LLMs have improved at 'reasoning' over time. This is a key capability that enables many scientific workflows.

baoilleach.blogspot.com/2026/01/impr...

A reminder that proteins are highly dynamic molecules.

🟧 We have made much progress in measuring & predicting static protein structures, but the dynamics that animate life remain challenging to measure & model.

Paths to cheminformatics: Q&A with Rajarshi Guha - Journal of Cheminformatics Journal of Cheminformatics -

link.springer.com/article/10.1...

Manager "We would lile to develop expert MedChem AI which can design molecules efficientry! Cheminformatician(like me) should develop it!"

MedChem "We don't beleave AI! It proposes strange molecules"

Me "Hmmmm... it's difficult to solve the issue.."