A fundamental lesson of modern AI is that scale is essential: training bigger models on bigger datasets unlocks new capabilities. A fundamental lesson of AI engineering is that scaling up isn't trivial: it is not just a matter of spending more money and resources.

Heads up: ignore samtools dot org, similarly minimap2 dot com and likely others. It's owned by a known phishing site and while the binaries they offer look valid currently (but note they may be serving us different binaries to others), that could change. Ie: it's not us (Samtools team)! Be warned

Within 24hrs, our board of trustees closed the local NPR station because transferring to new owners would have cost too much ($17M over 5 yrs) AND they increased the President's pay (total >$20M over next 7 yrs). How does this further PSU's land grant mission to extend knowledge to the public?

New preprint from Alan Brown in our lab, who is co-advised by @llinaslab.bsky.social. In this study, Alan studied chromatin state dynamics during the asexual intraerythrocytic development cycle (IDC) of the human malaria parasite, Plasmodium falciparum. www.biorxiv.org/content/10.1...

New preprint from the lab: Dr. Daniela James led an effort to improve the ChIP-exo protocol for high-res protein-DNA interactions. Particularly focused on optimizing for use in mammalian cells and making the protocol compatible with newer Illumina sequencers. www.biorxiv.org/content/10.1...

Looks like my book is available for pre-order: a.co/d/cTKIwBN #EpigeneticsBook

@jengreitz.bsky.social l & my lab want to co-hire a computational biologist/biostatistician with project management expertise to help map the regulatory code of the human genome and discover genetic mechanisms of disease. Details below careersearch.stanford.edu/jobs/computa... Plz RT

"I want to vibe code the genome" —the illustrious @casey.greenelab.com as quoted by @jperkel.bsky.social @nature.com 🧬🖥️ #MLSky www.nature.com/articles/d41...

The #RegSys @iscb-regsys.bsky.social track is gonna be outstanding again this year with a great selection of talks and keynotes @verapancaldi.bsky.social, Mafalda Dias, Roser Vento-Tormo, and @lucapinello.bsky.social ! www.iscb.org/ismbeccb2025...

Paper alert! I am really pleased to share the final version of the work led by Irène Amblard in the team on a regulatory switch controlling Cdx2 expression during posterior body development! #regulatorylogic #devbio @mrc-lms.bsky.social @imperialmed.bsky.social

This a really exciting leap forward for genomic sequence to activity gene regulation models. It is a genuine improvement over pretty much all SOTA models spanning a wide range of regulatory, transcriptional and post-transcriptional processes. 1/

This has been a wild story to follow. Someone is leading a campaign of pure harassment against @devoevomed.bsky.social 1/ bsky.app/profile/devo...

@saramostafavi.bsky.social (@Genentech) & I (@Stanford) r excited to announce co-advised postdoc positions for candidates with deep expertise in ML for bio (especially sequence to function models, causal perturbational models & single cell models). See details below. Pls RT 1/

A very special first post on Bluesky: I’m excited to share that I’ll be starting my lab at Penn State this coming January in the Dept of Biochemistry & Molecular Biology! I’ll be hiring at all levels soon. Check out updates, thoughts, and positions on the lab account: @hanna-lab.bsky.social

Free in-person registration is open for #MLCB2025! Sept 10-11 at @nygenome.org and online at youtube.com/@mlcbconf. Paper/abstract deadline is June 1, more deets including our fantastic invited speaker lineup at mlcb.org! Please RP.

🧬 Just out in Bioinformatics Advances: "Perspective on recent developments and challenges in regulatory and systems genomics" Find the full paper here: https://doi.org/10.1093/bioadv/vbaf106

Our New perspective explores breakthroughs & challenges in regulatory & systems genomics: from cis-regulatory code & 3D chromatin to spatial omics & AI models. 🔗 doi.org/10.1093/bioa... #Genomics #Bioinformatics #RegulatoryGenomics

ai.stanford.edu/postdoctoral... Postdoc candidates with strong expertise in AI for bio should apply for this new Stanford SAIL fellowship. We encourage applicants who would like to work with multiple SAIL faculty (we are part of SAIL as well). Come join us

our work on the molecular differences between transcription factor isoforms is out now in Molecular Cell! key point: 2/3rds of TF isos differ in properties like DNA binding & transcriptional activity many are "negative regulators" & misexpressed in cancer www.sciencedirect.com/science/arti...

Why AI won’t replace scientists capable of making disruptive conceptual breakthroughs - they pose questions that challenge, rather than align with, the training data set thomwolf.io/blog/scienti...

@austintwang.bsky.social is a fantastic grad student in the lab who will be completing his PhD in the Fall & is looking for research position in AI for bio in industry. www.austintwang.com 1/

By crushing cancer research, what are Elon & Trump making more efficient? Their NIH funding cuts are illegal—so instead, they're suffocating the grant approval process for cancer researchers with so much red tape that labs & clinical trials will be forced to shut down.

🎆Next #FragileNucleosome meeting is less than a week away! We are delighted to host @manucazottes.bsky.social, @robertadolling1.bsky.social & @leandrosboukas.bsky.social ! We will discuss evolution of XIST regulatory network, cancer epigenetics, and MDEMs! 🗓️ us06web.zoom.us/webinar/regi...

🚨PA colleagues: "Senator Fetterman wants to hear from you about how the federal funding freeze is affecting Pennsylvania." "If your project has been impacted, please fill out our constituent impact form:" forms.office.com/g/mFv2JAPxpC Get out your Other Support and share that info!

NIGMS MIRA explicitly asks for the following in investigator biosketches: “Include descriptions of significant service to the scientific community that are beyond normal mentoring and committee duties expected at the PD's/PI's institution.” They can’t then punish a PI based on the type of service

Another school curtails admissions, even offers already made. This is on #Trump, #Musk and #Vought. Time to fight back. www.thedp.com/article/2025...

The University of Pittsburgh confirmed Friday that there would be no new Ph.D. offers of admission while Pitt works to understand how reduced federal aid could impact the institution.

Almost all grant-review meetings under Trump 2.0 remain suspended at the US National Institutes of Health (NIH), preventing the world’s largest public funder of biomedical research from spending much of its US$47 billion annual budget. https://go.nature.com/4gM6oW4

It looks like 8/8 study sections that were scheduled to begin today have been cancelled: www.csr.nih.gov/RevPanelsAnd...

Are other outlets covering this? It's effectively another freeze on NIH funding. By my count, 3/8 study sections were cancelled today, 4/8 are cancelled tomorrow. Will only get worse as we hit those that were not posted before 1/20 @zey.bsky.social ? www.thetransmitter.org/funding/fede...

We think about this Q a lot--as noted above, we used CUT&RUN fragment size to infer that some nuc binding probably occurs, but our thought is that these events are likely transient so won't be efficiently captured via a re-ChIP-style method. FWIW C&R shows ~20% nuc-sized frags for FOXA1 ⬇️

And I’ll point out that in the case of success stories like Bell Labs or many tech research shops, the advances did not come from big vision but anarchic autonomy, researchers told, “Do whatever you want.” The breakthroughs were rarely in anticipated directions. 8/n

JD Vance has expressed a “deep hostility to the entire academic enterprise, not just the so-called woke parts,” writes @michellegoldberg.bsky.social. "He wasn’t talking about making more room for right-wing ideas in universities or even dreaming of taking them over. He wanted to destroy it all."

Bluetorial: The rationale for the “NIH” indirect cost “plan” is based on misunderstood (unintentional or intentional) accounting and math The value of real government transparency

Here's a quantitative illustration of what's happening with NIH funding.

Things that Ted Cruz thinks is woke: Taking middle school kids on a field trip to see the eclipse www.nsf.gov/awardsearch/...

Greenland famously acquired its name as a marketing tool to get viking settlers to move there. I'm fine with the new name if it encourages certain people to move there too

What's the indirect rate at SpaceX? Tough to answer, since they're not reqd to publicize indirect rates like universities. BUT, the cost of a Falcon9 launch is $100M for US gov't and $67M for a private company. So the ADDITIONAL overhead that SpaceX charges for working with the US gov't is 50%.

Many in biotech/private sector are boasting about their ways of funding and doing science But they are staffed with PhDs and postdocs we train in our NIH and NSF funded labs in universities With university labs lacking funds to train, the private sector will be profoundly harmed

"Every $100 million in investment leads to 78 patents and $598 million in further research, according to N.I.H. calculations" Or $100 million will get the taxpayer one Falcon 9 launch with SpaceX www.nytimes.com/2025/02/11/o...

This is an excellent article that accurately describes the situation with NIH funding. Great work @zey.bsky.social ! www.nytimes.com/2025/02/11/o...

Did you all know that SpaceX charges the US Gov't $100M per Falcon 9 launch, even though the rocket costs them as low as $15M? But Elon would have you believe that cancer researchers at non-profit universities are the ones milking the system.

We should talk more about where the US gov't gets huge value for taxpayer money. Take NIH study section: they get 25-30 expert PhD consultants to each spend 40-50 hours scrutinizing grant apps (incl 2 solid days of group discussion) and they pay them ~half of DC minimum hourly wage. And no coffee

It certainly is NOT just a higher ed issue. My company exists today because of the NIH SBIR program. At a 15% indirect rate, that money won't support the research it supposedly funds. My company would not have existed, and would not be supporting drug discovery and development worldwide today.

#BREAKING - 22 states sue to block Trump administration cuts to NIH research payments

This table is useful, but I think it is underestimating the effect. As far as I can tell, the table just multiplies 15% by the received direct costs per institution. But many direct costs do not receive overheads, so the effective overhead rate under the new limit would be much lower than 15%.

If this were just a cut to funding, that would be one thing - we deal with that all the time. What's actually happening is so much more insidious; it's designed to financially hurt universities in the short term and to make it impossible for them to continue scientific research in the long term. 1/

Trump & Musk are making massive cuts to the National Institutes of Health. I know there is a lot going on right now, but this is one of worst things they have done so far, will affect cancer research and trials, the search for cures, innovation and competitiveness, our universities. Your lives.

Why DO universities accept 10-15% indirect rates from foundations? Foundation grants have compelled recipients to subsidize the real cost of overheads. Why do universities subsidize research funded by Bill Gates' tax-exempt foundation at a much higher rate than research funded by the taxpayer?