Traver Hart

More evidence of this being the worst/dumbest timeline.

bsky.app/profile/domw...

Images of cells showing protein mislocalization caused by a Noonan syndrome mutation, with the idea to screen drugs to find any that can reverse the unhealthy mislocalization

Have $10M and want to cure some diseases?

It costs ~$1B to invent a new medicine…but $10M buys you Project Encore, to see if ANY existing drug might be repurposed for ~100 diseases that have no treatments.

Contact me - be a hero for desperate patients!
docs.google.com/document/d/1...

This is so, so important to show. Thanks for your efforts.

Beeswarm plot of the prediction error across different methods of double perturbations showing that all methods (scGPT, scFoundation, UCE, scBERT, Geneformer, GEARS, and CPA) perform worse than the additive baseline.

Line plot of the true positive rate against the false discovery proportion showing that none of the methods is better at finding non additive interactions than simply predicting no change.

Our paper benchmarking foundation models for perturbation effect prediction is finally published 🎉🥳🎉

www.nature.com/articles/s41...

We show that none of the available* models outperform simple linear baselines. Since the original preprint, we added more methods, metrics, and prettier figures!

🧵

The Sydney Sweeney debate is resolved by the 🧵 below.

Photo of a sketch of a Neanderthal. The headline reads: Scientists Growing Neanderthal 'Minibrains', Will Connect Them to Crab Robots

Go for it. I don't even care any more.

But it *does* mean it's among the former. Heh.

It is! if you know exactly what you want, it seems to be able to generate it for you.

What would have been an intern's summer job is now two days of iterative LLM work.

New prototype of Pickles now available for your review. Complete rewrite into node.js and JavaScript. Much faster for basic tasks. And entirely written by Claude/Sonnet, with ~8 hours of interactive prompts.

pickles.hart-lab.org

name of the grant: Testing if and how 3-factor learning rules operate at cortico-basal ganglia synapses to drive motor skill learning

Amidst all the chaos, I'm thrilled to have received a notice of award for my K99/R00! I've been waiting on pins and needles since this was scored in April (impact score of 14!!). I am officially on the job market - reach out if you are hiring or know of interesting positions anywhere!

same

What are your essential books on complex adaptive systems? I am making a list of my favorites and want to be sure I don't miss something. #complexity

For 9£/$12, this would sell like crazy here

We are LeGuin.

Contrasting photographs of the night-time skylines of Manhattan (left) and Nijmegen (right), with matching genome-wide association plots underneath each.

Not sure who came up with "Manhattan Plot", but in 2014 I coined the alternative term "Nijmegen Plot" (inspired by the Dutch town where I live) to describe underwhelming results from our earliest genome-wide association scans of language/reading traits.

Biologist folk (especially in evolutionary biology and/or ecology, but it don’t matter):

Can you give me your favorite examples of trade offs in biology? Organism or system don’t matter. Primary literature or reviews preferred.

I am looking to connect with my plant bio assistant professor cohort. If you know someone (or if it’s you!) please let me know.

#PlantScience

This is one of those awesome "I always wanted to do that experiment" papers. And a great 🧵.

Herr Goebbels Angered By Anti-Nazi Wit Comedians Banned For Humorous Remarks on Leaders --The Montreal Star, 4 Feb 1939

1939

Looking to combine ML and functional genomics to tackle cancer drug resistance? Exciting job opening at the Sanger!

sanger.wd103.myworkdayjobs.com/en-US/Wellco...

Full disclosure: I have been sitting on a 3% for six months now, afraid to step on cracks or walk under ladders on my way to work. I avoid mirrors entirely lest I accidentally crack one. It is not trivial to me.

But in the big picture, absolutely.

Oh I see. Disregard last skeet ;) And thanks!

If half the grants are not discussed, but your grant is discussed, and is the worst of the top half/discussed group, is your impact score 50% or 100%? I honestly don't know. I've never seen a percentile > 50%, and I don't know whether to pat myself on the back or not.

News worth celebrating!

why does time spent fasting feel like it ticks by so slowly

Just learned expit. apparently social media not always bad. who knew.

No:

bsky.app/profile/jwoo...

Things I learned this month: after breaking a rib (dumb story; please don't ask), atrophy of one side of your body after three weeks of trying desperately not to use it makes getting back into running astonishingly difficult. Like, whole physio systems just don't seem to work anymore.

National Cancer Institute projection of a 4% pay line in FY2026.

Good grief, the US National Cancer Institute projects it will fund only one in 25 RO1 grant applications in 2026. This is massively defunding cancer research. Despicable vandalism. www.cancer.gov/grants-train...