Jake Lever

Accelerating Cross-Encoders in Biomedical Entity Linking Javier Sanz-Cruzado, Jake Lever. Proceedings of the 24th Workshop on Biomedical Language Processing. 2025.

Here's the paper link: aclanthology.org/2025.bionlp-... #acl2025

Time for #BioNLP! @javiersanzcruza.bsky.social kicks off the workshop with our work on accelerating crossencoders for biomedical entity linking #acl2025

Our Glasgow AI4BioMed lab has some great work to present at #acl2025. Come along and chat!

We're kicking things off at #ACL2025 with a keynote from UW's Luke Zettlemoyer

And that's the end of the text mining track at #ISMBECCB2025. So much cool work going on in this area

Caralyn Reisle (@creisle.bsky.social) shows how language models can be used to verify claims in a big cancer knowledge base. A huge effort to create a fantastic benchmark set #ISMBECCB2025

Lynn Schriml ( @diseaseontology.bsky.social ) gives a great talk on the Disease Ontology's philosophy of building a sustainable resource so that others can be confident to build upon it. 20+ years of the DO! academic.oup.com/nar/article/...

Physically impossible molecule examples

Some beautiful and physically impossible predicted molecules from some AI models. Brilliant problems with AI from Charlotte Deane's keynote. The ring within a ring is particularly good! #ISMBECCB2025

Creating a reasonable baseline (without fancy ML) is so important! Sometimes our ML methods aren't actually that good. Great example from Charlotte Deane in her Day 4 keynote #ISMBECCB2025

Quick shout out for the Make Data Count challenge to improve data citation from @gigascience.bsky.social's Scott Edmunds. Looks cool and 💲prizes! #BOSC2025 #ISMBECCB2025 www.kaggle.com/competitions...

The O3 guidelines: open data, open code, and open infrastructure for sustainable curated scientific resources - Scientific Data Curated resources that support scientific research often go out of date or become inaccessible. This can happen for several reasons including lack of continuing funding, the departure of key personnel...

Keeping a data resource alive is a huge challenge. Lots of different guidelines being discussed in the #BOSC2025 panel for data resources including the O3 guidelines #ISMBECCB2025 www.nature.com/articles/s41...

Guy ending his talk telling the #BOSC2025 audience if they value global data infrastructure sign their open letter globalbiodata.org/open-letter-... #ISMBECCB2025

OPEN LETTER CAMPAIGN - Global Biodata Coalition Careers at the Global Biodata Coalition.

Biomedical data resources are essential for research but they're underfunded. Great to hear about the Global Biodata Coalition's mission to safeguard them. Time to sign their open letter! #ISMBECCB2025 globalbiodata.org/open-letter-...

Dystopian discussions about dealing with censorship in ontologies with the NCI deprecating transgender terms in the NCIT. Damien Goutte-Gattat gave a necessary but scary BOSC/BOKR talk about managing ontologies across borders. evsexplore.semantics.cancer.gov/evsexplore/c... #ISMBECCB2025

Ontologists do so much more than add new terms and curate papers! So many things to do, especially refactoring! Great BOSC/BOKR keynote from @cmungall.bsky.social #ISMBECCB2025

Are AI agents having virtual lab meetings the future? Day 3 kicks off with a keynote from @jameszou.bsky.social. #ISMBECCB2025

Undervaccination and severe COVID-19 outcomes: meta-analysis of national cohort studies in England, Northern Ireland, Scotland, and Wales Rates of undervaccination against COVID-19 ranged from 32·8% to 49·8% across the four UK nations in summer, 2022. Undervaccination was associated with an elevated risk of severe COVID-19 outcomes.

Some of the challenges in health data in the COALESCE study described by Angela Wood: people who died before they were born, pregnant men and having more patients registered at GPs than people in England #ISMBECCB2025 www.thelancet.com/journals/lan...

Inspirational talk from @profbuchan.bsky.social on the amazing work going on in Liverpool, especially with the city's support for COVID work even in the face of fearmongering press news.liverpool.ac.uk/2025/03/07/5... #ISMBECCB2025

Nucleotide dependency analysis of DNA language models reveals genomic functional elements Deciphering how nucleotides in genomes encode regulatory instructions and molecular machines is a long-standing goal in biology. DNA language models (LMs) implicitly capture functional elements and th...

Paper link: Nucleotide dependency analysis of DNA language models reveals genomic functional elements | bioRxiv www.biorxiv.org/content/10.1...

Brilliant results from DNA language models including an influence score that can differentiate benign and pathogenic variants from ClinVar. Great MLCSB talk from Julian Gagneur (@gagneurlab.bsky.social) #ISMBECCB2025

Funding bioinformatics resources is getting even harder. Should we be charging industry for access? If we do, it needs to be seamless. Great keynote from Amos Bairoch to kick off Day 2 of #ISMBECCB2025

Biomedical text mining for knowledge extraction @ ISMB 2025 A hands-on tutorial introducing biomedical natural language processing including named entity recognition, relation extraction and the power and limits of large language models.

Had fun talking about biomedical information extraction with @javiersanzcruza.bsky.social today. Check out the materials from our #ISMBECCB2025 virtual tutorial. It's a mini intro course on #BioNLP! ai4biomed.org/ismb2025tuto...