Edward Marcotte

@edwardmarcotte.bsky.social

Systems & synthetic biologist, proteomics, bioinformatics; Professor of Molecular Biosciences @ UT Austin; Cofounder, Erisyon

The brilliant Kris Kuo, with help from Fabian and the Werner lab, has identified a key driver of cyclic gene expression in Sulfolobus, which has a cell cycle but no CDK/cyclins. Amazingly, TFB2 has a cyclin box that regulates its degradation as cells exit division. Another reason to love archaea!

A TFIIB paralog drives cyclic transcription to orchestrate the archaeal cell cycle

Cyclic transcription is a hallmark of the cell cycle. While transcriptional waves in eukaryotes are driven by oscillations in cyclin-dependent kinase (CDK) activity, many archaea have an ordered cell ...

biorxiv.org

The reason Claude did well at binder design is because the prompt hardcoded that it should use ESMFold2 for scoring, and apparently ESMFold2 is really good. It's also only a few months old so not widely used yet. AI in biology keeps disappointing. blog.genesmindsmachines.com/p/anthropic-...

Anthropic has not solved the peptide-binder design problem, but maybe bi[o]hub has

The protein design field is getting pushed forward by human experts in protein design, who would have thought

blog.genesmindsmachines.com

In May it was big news when 100+ suspicious images were found in Thermo Fisher's catalog. Now, that number has expanded dramatically to 18,000+, thanks to some arduous searching by @reeserichardson.bsky.social. My reporting on the massive trove and its implications for antibody reliability:

More than 18,000 questionable images found in antibody catalogues of 15 companies

Science sleuth’s findings raise questions for researchers working to improve the reliability of commercial antibodies.

nature.com

A massive update: At least ~15 companies~ are selling scientists antibodies using faked validation data. We've documented 18,000+ manipulated images on 17,000+ products sold by leading laboratory suppliers including Thermo Fisher, Abcam, Santa Cruz Biotechnology, Millipore Sigma and Bio-Techne. 1/🧵

The window for public comments is still open but closes soon (tonight!) for NSF's proposed changes to require approval by political appointees, add restrictions to publishing & communication of research, and ban international collaborations.

Stand Up for Science!@standupforscience.net · last mo.

🚨WARNING🚨 Today is the LAST DAY to leave your public comment on the NSF Proposal. The Proposal implements some of Vought's Science-Destroying OMB Proposal plans in advance: zurl.co/9nywR Leave your public comment TODAY using our guide & tool: fight2win.standupforscience.net/campaign/Pro...

Can we use human language to predict novel bioactive molecules that work in vivo? In our new paper, we trained a deep learning model on a dataset derived from PubMed text and used it to discover novel antibiotic co-inhibitors that reduced infection burden, even in mouse models!

Bild

Learning from human and chemical languages to predict biological function Novel LLM AI approach trained on the literature Effective compounds in vivo for antimicrobial resistance Congrats UT Austin researchers Mavridou, Ellington, Wilke, Marcotte & colleagues www.biorxiv.org/content/10.6...

Learning from human and chemical languages to predict biological function

Understanding how molecular structure encodes biological function remains a grand challenge in drug discovery. Here, we present PubCheF-1, a deep learning model that predicts literature-derived biolog...

biorxiv.org