ICYMI

My MPG primer talk from last week at the Broad Institute. Primers are weekly lectures on genetics topics related to human populations and disease. #RVAT #Stat #NERINE #Genetics #CAD #Parkinson's #StatGen Link to the talk: youtu.be/KO-6M5lXTlI?...

In my latest column, I explain some of my reasons for being deeply skeptical about AI models that claim to understand DNA, genes, and genomes: stevensalzberg.substack.com/p/ai-is-star...

Multiple protein structure alignment at scale with FoldMason www.science.org/doi/10.1126/... 🧬💻🧪

Our take-aways are that GWAS can be broadly linked to eQTLs—just not in humans (yet). This is probably due to selection and population history. We expect that as we find human eQTLs under stronger selection, colocalizations will increase.

Check out this exciting preprint from @noahconnally.bsky.social to see how studying other species may be useful for searching trait-associated eQTLs in humans.

"… I realized that an environment where people would feel free to share their thoughts with me and think creatively, regardless of the power hierarchy, was exactly what I wanted in my lab." #ScienceWorkingLife https://scim.ag/4adQwdQ

I have to believe in the many worlds hypothesis because, somehow, we ended up in the worst of all possible worlds. I have to believe worlds that evaded this horrific fate must exist.

European Scientists who said NO to Epstein-funding are the heroines of this hour🦸‍♀️ They show: It was your choice to SAY NO to Epstein's $ as scientist- you just had to have ethics & a spine Kudos to Physicist IVETTE FUENTES who said NO She was working at @quantumunivie.bsky.social @univie.ac.at

At a time when the UK should be capitalising on its global reputation as a great place to do science, this is an enormous own goal. I am so sorry for all those who jobs are going to be impacted.

We are thrilled that our study on the evolution of gene regulation in mammalian cerebellum development – led by @ioansarr.bsky.social, @marisepp.bsky.social and @tyamadat.bsky.social, in collaboration with @steinaerts.bsky.social – is now out in @ScienceMagazine! www.science.org/doi/10.1126/...

With all the excitement about AI giving us answers, we often forget that the real trick is to figure out what is the question! Thank you Tim Ferriss for recommending "What is the question" on the Night Science creative process in your newsletter! link.springer.com/article/10.1...

In a rapidly changing AI landscape, how do you stay relevant? Focus on timeless skills: 1. the ability to learn and have the curiosity to learn.

Check out the latest preprint from @j-e-mitchel.bsky.social on a new single-cell genetic colocalization tool, scJLIM, which shows significant power gain over bulk colocalization. It has been a pleasure being part of the effort!

We are excited to share GPN-Star, a cost-effective, biologically grounded genomic language modeling framework that achieves state-of-the-art performance across a wide range of variant effect prediction tasks relevant to human genetics. www.biorxiv.org/content/10.1... (1/n)

Do you know ~60% of human SVs fall in ~1% of GRCh38? See our new preprint: arxiv.org/abs/2509.23057 and the companion blog post on how we started this project and longdust: lh3.github.io/2025/09/29/o.... Work with Alvin Qin

We are deeply saddened to hear of the passing of Jane Goodall. Jane was a much-loved member of the conservation community in Cambridge and worldwide. She completed her PhD at @darwincollegecam.bsky.social and was an Honorary Fellow at @newnhamcollege.bsky.social. www.youtube.com/watch?v=ZIzR...

Yup. This was me 10 years ago.

The 2025 Ig Nobel-winning research was as wacky as ever. But wars, visa restrictions, Trump's border and research policies, and even an all-too-frequent airplane incident kept nearly half the winning teams from traveling to Boston @science.org www.science.org/content/arti...

In a new preprint from @brianhie.bsky.social's lab, the team reports the first generative design of viable bacteriophage genomes. Leveraging Evo 1 & Evo 2, they generated whole genome sequences, resulting in 16 viable phages with distinct genomic architectures.

Brilliant paper by Visscher et al. Populations differ in traits/disease burden. Are these differences due to genetics? Comparing single variants or polygenic scores between populations is biased due to environmental confounders correlated with the variants. 1/3 www.medrxiv.org/content/10.1...

More of "The AI Coming For Us". Well, one way it's going to take us out is if the next generation of folks relies on this slop for anything serious ....

Whether we like it or not, AI is here and not going anywhere. The excellent colleagues behind @callingbullshit.bsky.social (@jevinwest.bsky.social & @carlbergstrom.com) now have a great resource to help us understand AI and how to deal with it. Highly recommended! thebullshitmachines.com/index.html

❗ I clearly consider this result as THE most important result achieved over this last decade for exploiting and democratizing genomic data. I think there will be a "before" and an "after" logan and logan-search github.com/IndexThePlan... logan-search.org Have a look at this thread

New York Times story with profiles of researchers whose grants were terminated. [Gift Link] www.nytimes.com/2025/08/24/o...

Even superintelligent AI cannot simply replace humans for most of what we do, nor can it perfect or ruin our world unless we let it, AI Snake Oil’s @randomwalker.bsky.social tells EFF’s Cindy Cohn and @thejasonkelley.com on the new “How to Fix the Internet.”

Academia may not give you job security, flexibility, or wealth, but it will let you unexpectedly connect to eduroam in foreign cities

This is also quite good. The "doing more with less" definition of intelligence (and "doing less with more" definition of stupidity) has an appeal. Reminds me of an old philosophy professor I had. Brilliant guy. Barely was on email. But his work never suffered for it. youtu.be/jXa8dHzgV8U?...

Dijkstra’s is not optimal anymore for finding shortest paths!? Breaking the Sorting Barrier for Directed Single-Source Shortest Paths arxiv.org/abs/2504.17033

At the same time, we made thousands of synonymous mutations in endogenous yeast genes and measured their growth. We used careful statistics and controls. Only 3%, 204 of 6874, had a fitness effect! This goes against a controversial recent result that most synonymous mutations had fitness effects.

‘Seek Funding’ Step Added To Scientific Method

thought of this episode from Recess www.youtube.com/watch?v=1mLo...

AlphaFold is great, but contrary to public opinion it has not completely solved the protein folding problem. Much work remains to be done. clauswilke.substack.com/p/no-alphafo...

“There is a cult of ignorance in the United States and there always has been.” — Isaac Asimov.

No you cannot determine which drugs are safe and effective with machine learning. You cannot. It will not work. People will die.

Pleased to share our ICML Spotlight with @eberleoliver.bsky.social, Thomas McGee, Hamza Giaffar, @taylorwwebb.bsky.social. Position: We Need An Algorithmic Understanding of Generative AI What algorithms do LLMs actually learn and use to solve problems?🧵1/n openreview.net/forum?id=eax...

1/27 We have a new paper out! Turns out that snowflake yeast have been hiding a secret from us - they've evolved a (very!) crude circulatory system. Not with blood vessels or a heart, but through spontaneous fluid flows powered by their metabolism. 🧪🔬 www.science.org/doi/full/10....

The NIH has not issued a single new call for proposals since the end of January.

Eid mubarak to all who are celebrating!

More from @holdenthorp.bsky.social in @science.org "A world where the United States is no longer leading the scientific enterprise will still benefit from science [...] The global enterprise will adapt to the lack of American leadership, but the steep loss for the country itself is unambiguous."

A case study on the challenges of evaluating AI predictions in biology and the implications for published results. 1/2 rachel.fast.ai/posts/2025-0...

“The group estimated that reducing the operating budget by $900 million, as the Trump administration wants to do, would require closing 350 of the 433 parks, monuments, historic sites and other locations overseen by the Park Service… we would be witnessing the dismantling of a century-old system”

HEY! The countdown is on—abstract submissions for #ASHG25 close in 11 days! 🔬 Share your research. 🤝 Build your network. ✨ Inspire your field. 🗓️ Submit now: www.ashg.org/meetings/202... #ASHG #HumanGenetics

An emerging narrative on the right is that international students take seats away from deserving domestic students - a zero sum thinking about what universities do and their function 1/

Now published! Congrats Yon Ho on this huge lift conducting dozens of GWAS in Korea and performing meta-analyses with other Asian biobanks reaching a sample size comparable to UK Biobank. The fine-mapping, novel loci, and ancestry-enriched discoveries are very cool. www.nature.com/articles/s41...

My father worked in a tire factory in Appalachia. Friends and family in the military. Used to vote Republican. Got into college on a diversity program for the underprivileged. Ended up a Harvard professor. And now my funding is terminated because we're trying to stick it to the elites.

NIH funding supporting the HMMER and Infernal software projects has been terminated. NIH states that our work, as well as all other federally funded research at Harvard, is of no benefit to the US.

Genuinely... quite excited about this. I think I might have to install it just for old skool x new school kicks. (plus - great for controlled cloud environments!)

AI is not changing or improving work for the vast majority of early adopters - most of whom are adopting out of fear of being left behind. The fear is being directly driven by the incessant hype produced by the same corporations benefitting from the hype cycle. Corporate Centipede.

Corporate administrators who have only experienced jobs whose main benefit was pay cannot fathom that the very best find intrinsic value in excellence and are best left to do it without undue interference or burden