Let’s dream a bit. How would you dramatically reorg biotech research / drug development to make dramatically faster progress towards curative therapies?. (Don’t say “use AI”, hypothesis/candidate generation under current structure isn’t a bottleneck).

RT @BenjaminGVincen: NUT Carcinoma patients: Consider the T-Scan clinical trial (link below). T-Scan has TCR-Ts targeting cancer/testis an….

Listening to my first Dwarkesh episode and…it’s pretty good?. Definitely a funny sub-culture of jargon, mannerisms, & priorities but it’s fun to dream big.

"A Screening Study to Collect Samples for TAA, HLA & HLA Loss of Heterozygosity in Patients With Metastatic Solid Tumors".

I think this flew under the radar & might benefit someone who needs it: . T-Scan added NUT carcinoma as an indication for their TCR-T trial. Patients should enroll in the screening protocol (NCT05812027) to see if their cancer has a target for which T-Scan has TCR-Ts.

Great talk by Andy (Jin Seok) Lee. He’s built a stack of tools to pull potentially coding structural variants from long read DNA seq, mapping their mutational footprint through long read RNA to find extremely non-self mutant proteins. (also, how did my camera get these colors?)

RT @SynBio1: If biotech ever gets faster and more innovative, one reason will be that we've lowered the status of managers and raised the s….

RT @jrkelly: Excited to launch our 1st Antibody Developability AI competition today on @huggingface ! . Show off your model’s ability to p….

“binders per hour”

The people, they exalt Xenium yet abhor Visium.

Very happy to have played a minor role in this paper and awed at Abe's persistence in getting it published. As you might imagine, journals seem a lot more receptive to the original (false but exciting) claims than a careful attempt at replication / deconstruction.

TIL: "Mothers against decapentaplegic".

RT @AbrahamGihawi: Finding microbes in cancer sequencing data has been controversial. Are we seeing real biology… or just contamination + b….

RT @StevenSalzberg1: We're still not finding any good evidence for a microbiome in any cancer type: see our new paper in @ScienceTM led by….

RT @design_proteins: Trying out PXDesign on a target that ColabDesign and RFDiffusion struggled with previuosly for me a while back (CD20)….

TIL: Wellcome Open Research (. Seems like a great model.

The blog post cites two previous studies but mostly to highlight how limited of a sequence space exploration each one can do individually. What they don't say, which I'm curious about, is how much the non-reference sequences are the union of many different different studies.

[*] OpenAI x Retro blog post: Previous work: .

Curious what folks who work on iPSCs think of the @openai x @RetroBio_ [*] collab to design efficient Yamanaka factors using a fine-tuned GPT-4o. I wonder to what degree the designs integrate all the previously published work on sequence properties of improved Sox2/Klf proteins

4yo asking to hear “the song from cake-pop demon hunters”. None on Spotify sounded right bc he’s only heard it when a 5yo friend plays it on piano. Cultural synthesis in the making.

RT @btnaughton: Great blogpost by Nick Boyd and Sam Guns from Escalante Bio (posted on the other site!). They intro….