as Marshall College Professor Jones could have said: "This belongs in a database!" https://doi.org/10.1021/acsomega.4c10413
Well, the first pKa column and all arsonic acids are now in @wikidata :)
#Cheminformatics
6 posts4 participants0 posts today
new BridgeDb Datasources release: https://github.com/bridgedb/datasources/releases/tag/20270728
"Uses UniProtKB as name, removes EcoGene, and multiple small updates"
BridgeDb Datasources is a dataset with metadata about data sources and organisms used by BridgeDb Java and downstream tools like @wikipathways, PathVisio, and others
Join Franciszek Job at EuroSciPy as he presents a scalable framework to unify chemical datasets from sources like PubChem, UniChem & COCONUT.
Canonicalize with RDKit
Scale via Dask
Deduplicate with InChI keys
Ideal for ML pretraining, benchmarking, and chemical data analysis.
Schedule: https://lnkd.in/eaAxwUN2
Tickets: https://lnkd.in/end9aYzE
lnkd.inLinkedInThis link will take you to a page that’s not on LinkedIn
chemical compound identifiers in @wikidata: https://edu.nl/xb8y7
Most cheminformatics code that queries ChEMBL struggles with reproducibility.
chembl-downloader can help:
>>> import chembl_downloader as cd
>>> df = cd.query("""
SELECT chembl_id, pref_name
FROM molecule_dictionary
WHERE pref_name IS NOT NULL
""")
It's even sneaking its way into @wpwalters and @dr_greg_landrum blogs :)
Code/Docs: https://github.com/cthoyt/chembl-downloader
Preprint: https://arxiv.org/pdf/2507.17783
New Preprint Alert!
We're excited to share our latest work on #ChemRxiv! MARCUS (Molecular Annotation and Recognition for Curating Unravelled Structures) is a web-based platform for extracting chemical information from scientific papers.
Preprint: https://doi.org/10.26434/chemrxiv-2025-9p1q1
Try it out: https://marcus.decimer.ai
ChemRxivMARCUS: Molecular Annotation and Recognition for Curating Unravelled StructuresThe exponential growth of chemical literature necessitates the development of automated tools for extracting and curating molecular information from unstructured scientific publications into open-access chemical databases. Current optical chemical structure recognition (OCSR) and named entity recognition solutions operate in isolation, which limits their scalability for comprehensive literature curation. Here we present MARCUS (Molecular Annotation and Recognition for Curating Unravelled Structures), a tool to aid curators in performing literature curation in the field of natural products. This integrated web-based platform combines automated text annotation, multi-engine OCSR, and direct submission capabilities to the COCONUT database. MARCUS employs a fine-tuned GPT-4 model to extract chemical entities and utilises an ensemble approach integrating DECIMER, MolNexTR, and MolScribe for structure recognition. The platform aims to streamline the data extraction workflow from PDF upload to database submission, significantly reducing curation time. MARCUS bridges the gap between unstructured chemical literature and machine-actionable databases, enabling FAIR data principles and facilitating AI-driven chemical discovery. Through open-source code, accessible models, and comprehensive documentation, the web application enhances accessibility and promotes community-driven development. This approach facilitates unrestricted use and encourages the collaborative advancement of automated chemical literature curation tools.
We dedicate MARCUS to Dr Marcus Ennis, the longest-serving curator of the ChEBI database, on the occasion of his 75th birthday.
ha, I managed to convert a CXSMILES in @wikidata via the MDL V2000 molfile and the #inchi webtool into an #InChI (!B) and InChIKey :) https://www.wikidata.org/wiki/Q66421202#P117
heading to the 2nd day of the Technical InChI Meeting in Aachen/Germany
heading tomorrow to the InChI Technical Exchange Meeting Summer 2025 in Aachen/DE
Looking forward to it, and particularly talking about the InChI for inorganics and trying that in @wikidata :) See https://doi.org/10.26434/chemrxiv-2025-53n0w
And also the nano InChI, see https://doi.org/10.3390/nano10122493
ChemRxivScholia Chemistry: access to chemistry in WikidataSharing knowledge on chemicals in the digital age has been the playground of databases such as the Chemical Abstract Services and PubChem. Wikipedia complements this field by providing context to chemicals aimed at a broad audience, but is not easily read by machines. Wikidata was started as a database service to improve the machine readability of the knowledge captured in Wikipedia. Wikidata has an open license, application programming interfaces, and a strong provenance model. Scholia uses the features to provide access to chemical knowledge. This study reviews the chemistry in Wikidata, shows how thousands of new chemicals were added, extends Wikidata with new properties for chemical representation and external links to additional databases, and shows how we extended Scholia to represent the chemistry in Wikidata.
updating some java #cheminformatics libraries... Euclid and CMLXOM, and possibly making a Bacting release... but need to continue working on the CDK Depiction book chapter too...
oh boy... finally getting there... so close to finalizing and release a new SMARTCyp?
See for context https://chem-bla-ics.linkedchemistry.info/2024/04/07/cdk2024.html and https://chem-bla-ics.linkedchemistry.info/2024/06/16/cdk2024-3.html #cdk2024
did you notice I started @Codeberg ?
# Qleverfile for PubChem, use with the
# QLever CLI (`pip install qlever`)
#
# qlever get-data # ~2 hours, ~120 GB, ~19 billion triples
# qlever index # ~6 hours, ~20 GB RAM, ~350 GB disk space (for the index)
# qlever start # a few seconds
Source: https://github.com/ad-freiburg/qlever-control/blob/main/src/qlever/Qleverfiles/Qleverfile.pubchem
I think I am going to try to recover a bit of #cheminformatics / #chemistry #history, and make the index of the Internet Journal of Chemistry (IJC) FAIR in @wikidata
While the journal no longer exists, many articles are cited quite a few times.
I did some exploration some time ago, and for some I found full text "self-archiving" versions online.
And, TIL that Web of Science has entries for the articles too, which I just added for the 9 articles already in #Wikidata: https://w.wiki/Eide
Anyone have views or references on the effectiveness of count/sparse #cheminformatics fingerprints compared to binary/dense fingerprints?
What about comparing methods to turn count/sparse fingerprints into binary ones? I know of several approaches, but nothing methodical or published.
I'm trying to understand how I might add sparse/count fingerprints to chemfp.
new blog post: "PFAS in the blood of the Dutch population" https://chem-bla-ics.linkedchemistry.info/2025/07/06/pfas-in-the-blood-of-the-dutch-population.html
About the recent @rivm report.
"So, what can I do to make this report more FAIR?"
Continued thread
and follows by CMLXOM 4.12: https://github.com/BlueObelisk/cmlxom/releases/tag/cmlxom-4.12
CMLXOM is a Java library for processing CML, implementing the XML object model (XOM) for the Chemical Markup Language (CML).
Euclid 2.11 is released: https://github.com/BlueObelisk/euclid/releases/tag/euclid-2.11
Euclid is a library of numeric, geometric and XML routines.
Version 5.0b1 of chemfp - the comprehensive package for binary #cheminformatics fingerprints - is out and ready for curious beta testers.
Here's are highlights. For more info and Linux install info see https://chemfp.com/chemfp-50b1-available-for-beta-testing.html
- shardsearch to search multiple files
- simhistogram for a histogram of pairwise Tanimoto comparisons
- the FPB file size limit increased from ~250M fingerprints to well over a billion
- new Klekota-Roth fingerprint type for RDKit and OpenEye
chemfp.comchemfp 5.0b1 available for beta testing
Ammar's PhD thesis can be found online at https://cris.maastrichtuniversity.nl/en/publications/enhancing-the-interoperability-and-reusability-of-nanosafety-data and his defense is live streamed, see the link at https://www.maastrichtuniversity.nl/events/phd-defence-ammar-ammar
It contains a lot of #openscience: https://social.edu.nl/@egonw/114766897696096032
And in the afternoon there is a minisymposium on the thesis topic: https://social.edu.nl/@egonw/114705860626722108