Lightnews — Scholar-powered news

Liana Lareau

@lianafaye.bsky.social

1.6K followers 240 following 54 posts

Assistant Professor of Bioengineering, Berkeley. Living in an RNA world. lareaulab.org

Posts Media Videos Starter Packs

Pinned

Liana Lareau @lianafaye.bsky.social · Aug 7

This preprint from Helen Sakharova is one of the coolest things to come out of my lab: “Protein language models reveal evolutionary constraints on synonymous codon choice.” Codon choice is a big puzzle in how information is encoded in genomes, and we have a new angle. www.biorxiv.org/content/10.1...

Protein language models reveal evolutionary constraints on synonymous codon choice

Evolution has shaped the genetic code, with subtle pressures leading to preferences for some synonymous codons over others. Codons are translated at different speeds by the ribosome, imposing constrai...

www.biorxiv.org

6 83 210

Reposted by Liana Lareau

Liz Neeley @lizneeley.bsky.social · 6d

Some problems have simple solutions. This new “Compact for Academic Excellence in Higher Education” is one. The answer is simply saying no.

Simple does not mean easy. But Brown, Dartmouth, MIT, Vanderbilt, Univ of AZ, UPenn, USC, UT Austin & UVA must understand what they’ll lose if they sign on.

2 190 490

Reposted by Liana Lareau

Andy Sellars @sellars.bsky.social · 8d

AAUP v. Rubio is out, and look at how it starts. storage.courtlistener.com/recap/gov.us...

Ahead of the case caption, a handwritten note: "Trump has pardons and tanks . . . . what do you have?" Judge Young's reply: "Dear Mr. or Ms. Anonymous,
Alone, I have nothing but my
sense of duty.
Together, We the People of the
United States –- you and me --
have our magnificent Constitution.
Here’s how that works out in a
specific case –- "

42 540 1.7K

Reposted by Liana Lareau

Atul Gawande @agawande.bsky.social · Sep 1

9 CDC Directors going back to 1977 speak out. What RFK Jr has done to our nation’s public health system "should alarm every American."

It "is unlike anything we have ever seen at the agency, and unlike anything our country has ever experienced." www.nytimes.com/2025/09/01/o...

Opinion | We Ran the C.D.C.: Kennedy Is Endangering Every American’s Health

www.nytimes.com

330 5.2K 12K

Reposted by Liana Lareau

Andrea Love, PhD | Biomedical Scientist @drandrealove.bsky.social · Aug 31

mRNA vaccines are safe, have been in development for decades, & have far broader applications than older vaccine technologies.

New vaccine technologies evolve when scientists have tools & knowledge to make something more elegant or complex that wasn’t possible before.

5/

1 36 86

Reposted by Liana Lareau

Carl Zimmer @carlzimmer.com · Aug 29

On the new acting CDC director. www.theguardian.com/us-news/2025...

O’Neill, unlike Monarez, has no training in medicine or infectious disease science. He is a former speechwriter for the health department, during the Bush years, who went on to work for the tech investor and conservative megadonor Peter Thiel.

During the Covid pandemic, O’Neill voiced public support for unproven treatments that were not supported by scientific evidence, including ivermectin and hydroxychloroquine, as well as vitamin D as a supposed “prophylaxis”.

He also posted a number of conspiratorial theories on social media, including the baseless claim that “the name #COVID was chosen to conceal the origin of the virus. This name made it harder to study and probably slowed the response.”

83 770 1.7K

Reposted by Liana Lareau

Carolyn Bertozzi @carolynbertozzi.bskyverified.social · Aug 27

Spread the word - we at @stanford-chemh.bsky.social are searching to fill a new junior faculty at the interface of molecular and computational science. See link below!

Sarafan ChEM-H @stanford-chemh.bsky.social · Aug 26

Open faculty position!
We're seeking applicants for a tenure-track faculty position at the junior level (Assistant or untenured Associate Professor) with research programs that exist at the interface between molecular science and computation. Apply here: stanford.io/45MF3Qa

Open Faculty Position: Assistant or Associate (Untenured) Professor

Application deadline: 11:55 PM on Wednesday, October 15, 2025

Sarafan ChEM-H is seeking candidates with research programs that exist at the interface between molecular science and computation.

Click the link above to learn more & apply!

78 77

Reposted by Liana Lareau

berkeleyMCB @berkeleymcb.bsky.social · Aug 20

We're Hiring! Assistant Professor of Immunology and Molecular Medicine in MCB. Learn more and apply online:
aprecruit.berkeley.edu/JPF05096

31 21

Reposted by Liana Lareau

berkeleyMCB @berkeleymcb.bsky.social · Aug 20

We're Hiring! Assistant Professor of Molecular Therapeutics in MCB. Learn more and apply online:
aprecruit.berkeley.edu/JPF05098

7 4

Liana Lareau @lianafaye.bsky.social · Aug 20

Or rather, the study section was still giving lots of good scores but the grants were just never funded. It was such a waste of so many scientists’ time on both sides.

Liana Lareau @lianafaye.bsky.social · Aug 20

Was this perhaps an NIGMS-heavy study section implementing their mostly-unannounced policy to pretty much stop funding R01s in favor of R35s (which go to different study sections)?

Reposted by Liana Lareau

Marla Feller @mfeller.bsky.social · Aug 20

Come join our new Department of Neuroscience @ucberkeleyofficial.bsky.social as an Assistant Professor! aprecruit.berkeley.edu/JPF05041

Assistant Professor - Molecular, Cellular, and Developmental Neuroscience - Department of Neuroscience

University of California, Berkeley is hiring. Apply now!

aprecruit.berkeley.edu

1 82 77

Liana Lareau @lianafaye.bsky.social · Aug 9

Yes, it's a discrete counting thing: think about the most extreme log2 fold change numbers you could observe if you start with 100 or 10 or 2 items and end up with 1 item.

Reposted by Liana Lareau

Brendan Nyhan @brendannyhan.bsky.social · Aug 9

Would you want to invest in a country where the regime expropriates assets from its opponents?

In related news, "personalist regimes are characterized by lower total factor productivity and ... low private investment, poor public-goods provision, and conflict" www.nber.org/system/files...

10 74 220

Reposted by Liana Lareau

Joé McKen @joemcken.net · Aug 8

Science is effectively dead in the US for at least the next 3½ years, and will then take another several years to even get started again. Canceled research doesn’t just uncancel itself, and scientists who find opportunity elsewhere won’t just come flocking back.

The damage is generational.

Mark Lemley @marklemley.bsky.social · Aug 8

All scientific grant funding must now be approved by a political appointee and "demonstrably advance the President's policy priorities."

I wonder where innovation will happen in the future? It won't be in the US

arstechnica.com/science/2025...

New executive order puts all grants under political control

All new funding on hold until Trump administration can cancel any previously funded grants.

arstechnica.com

32 590 1.4K

Liana Lareau @lianafaye.bsky.social · Aug 8

That's really interesting - incidentally, we also have some work on how the speed of elongation might feed back on initiation (not in the context of membrane proteins, in this case, and no mechanism yet!). Definitely some interesting possibilities. www.biorxiv.org/content/10.1...

Translation elongation as a rate limiting step of protein production

The impact of synonymous codon choice on protein output has important implications for understanding endogenous gene expression and design of synthetic mRNAs. Synonymous codons are decoded at differen...

www.biorxiv.org

Reposted by Liana Lareau

Lisa Jarvis @lisamjarvis.bsky.social · Aug 7

"Preparedness is a deterrent," Nuzzo says. Walking away from an investment in mRNA also "sends a message to our adversaries that the United States is uncommitted to preparing for future health emergencies - that the United States is now more vulnerable than it has ever been because we are systematically taking off the table multiple approaches to protect ourselves."

1 22 68

Reposted by Liana Lareau

Elizabeth Jacobs, PhD @elizabethjacobs.bsky.social · Aug 7

We are on a bullet train to the end of US science, and possibly the country as a whole, when political appointees are reviewing every scientific grant submitted to NIH to determine if The Regime finds the science acceptable.

29 280 930

Liana Lareau @lianafaye.bsky.social · Aug 7

In all, how much constraint on synonymous positions? From our model, the equivalent of 9% are predictable (sum of smaller effects on more positions) but that excludes strong overall preferences for fast codons. And 3% is just what we can see in a short competition. So: more than zero, less than all!

Liana Lareau @lianafaye.bsky.social · Aug 7

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.

Scatterplot showing fitness effect of ~7000 synonymous mutations in yeast: read count at start vs log2 fold change. Most data points are not significant but 204 points are significant outliers, either advantageous or deleterious.

1 9

Liana Lareau @lianafaye.bsky.social · Aug 7

The experiment and our model aligned to highlight localization, especially to the mitochondria. There is interesting biology to pursue there. More broadly, our model works -- and that means that evolutionary forces on proteins shape codon choice in nuanced but real ways.

3 2 9

Liana Lareau @lianafaye.bsky.social · Aug 7

3 29 80

Liana Lareau @lianafaye.bsky.social · Aug 7

So. Yes, information about codons can be pulled out of a model that only saw protein sequence! It predicts slow codons in signal peptides and transit peptides for localization. Fast codons in protein domains and active sites for accuracy. But, no big signal about co-translational protein folding.

A schematic showing that synonymous codon choice affects translation speed, which in turn affects protein output, cotranslational localization, and translation accuracy. A fourth category, cotranslational folding, is greyed out to show it was not a strong finding.

1 11

Liana Lareau @lianafaye.bsky.social · Aug 7

Before we get there, we have to factor out some legit but trivial things. A model can learn that certain types of genes tend to have fast codons or high GC content. We did a lot of work to keep that out of our predictions.

1 4

Liana Lareau @lianafaye.bsky.social · Aug 7

There’s been a lot of exciting work recently using large language models to capture the language of biological sequence. The Evolutionary Scale Model takes every protein ever seen and learns what works and what doesn’t. What if it could predict codon choice without ever seeing a codon?

1 6

Liana Lareau @lianafaye.bsky.social · Aug 7

Like, it might be important to slow down the ribosome so the protein folds properly and doesn’t tangle up. We were looking for signals like that, directly related to translation. (There are also many other ways that sequence matters: mRNA structure, binding sites, etc.)

1 8

Liana Lareau @lianafaye.bsky.social · Aug 7

Synonyms are not identical, in language or in biology. Some synonymous codons are preferred over others, and the preferred ones are decoded more quickly, but they all work, more or less. Where does codon choice really matter?

1 5