Erik Poppleton

Ooh, I would love this!

A day late but STILL IMPORTANT

Trump freezes $200 million in UCLA science and medical research funding, citing antisemitism The Trump administration has frozen roughly $200 million in National Science Foundation, National Institutes of Health and other federal agency-funded research at UCLA, citing allegations the school d...

200 engineering/life science/medical grants, including my lab's NIH MIRA, have been suspended. In case you are wondering, this means immediately stop spending = stop paying salaries for students & postdocs. www.latimes.com/california/s...

This whole situation is enraging! Is the university able to do anything to support the salaries or is everybody suddenly having to scramble?

Screenshot from Google Scholar showing Bryan "Dexter" Holland's publications. The paper has been cited 34 times and the thesis 4 times.

This is just... not true? His thesis was on sequences which resemble miRNA in the HIV genome which affect CD4 expression. It's an impressive thesis, especially from a globetrotting rockstar. But neither it, nor the associated paper is highly cited, let alone by anything to do with vaccines. 🧪

a 3d rendering of a colorful dna molecule with a black background ALT: a 3d rendering of a colorful dna molecule with a black background

Happy #RNA Day everyone!!! 🎉🎉🎉

#AUG1st #RNASky

Identification of conserved RNA regulatory switches in living cells using RNA secondary structure ensemble mapping and covariation analysis - Nature Biotechnology Transcriptome-scale maps of RNA secondary structure ensembles in living cells detect candidate RNA structural switches.

I am so incredibly excited to share our latest work, on the exploration of #RNA secondary structure ensembles and discovery of RNA regulatory structural switches in bacteria and human cells, just out in
@natbiotech.nature.com: nature.com/articles/s41.... A short tread! (1/n)

Engineering Liposomes with Cell Membrane Proteins to Disrupt Melanosome Transfer between Cells Cells communicate by transporting vesicles and organelles, which is essential for maintaining cellular homeostasis. However, dysregulated vesicle transfer between cells can contribute to several disea...

Thrilled to share our latest work in ACS Nano!

We developed a drug-free strategy to reduce melanosome transfer, a key process in hyperpigmentation and melanoma progression, and to broadly modulate vesicle-based cell communication.

Grateful to a fantastic team!
pubs.acs.org/doi/abs/10.1...

Large-scale processing of within-bone nutrients by Neanderthals, 125,000 years ago Neanderthals intensively processed a minimum of 172 large mammals for grease and marrow fat, 125,000 years ago.

So they found a 125,000 yr old Neanderthal fat rendering plant & I have thoughts.

Knowing that they loved saturated fat SO MUCH that they industrialized to get as much of it as humanly possible?

Makes me feel seen, heard, supported, etc www.science.org/doi/10.1126/...

So excited to finally be able to share this manuscript!! Thanks to all collaborators for making this possible! Special thanks to @kgoepfrich.bsky.social and @xaviertrepat.bsky.social for their guidance and supervision in this!

DNA hydrogels as mechanical sensors in cell-culture systems by @towalther.bsky.social 🧬🧬🧬

If you catch him at a conference bar, ask him the story, it’s an adventure and a half!

With Tom Lehrer's passing, I suppose this is a moment to share the story of the prank he played on the National Security Agency, and how it went undiscovered for nearly 60 years.

Unusual protein found in a Loki Asgard Archaea.

Check out this weird protein we found in a Loki Asgard archaea.🧬🧶Only found relatives in other Lokis. 🤔

www.biorxiv.org/content/10.1...

Why do all the rural stations buy NPR programs? Because as part of being public, NPR charges based on listenership! This means it’s affordable to all these tiny stations which only exist thanks to CPB funding because ‘the market’ can’t support them. The affiliates in big media markets will be fine.

👀👀👀

Software is easy to collaborate on across borders. How we can facilitate that:
- better documentation (always the answer)
- good software engineering to make adding a feature relatively self-contained
- Norms around contributing to open-source software

What about lower-investment community building?
- record webinars and experimental protocols. Create a library of what the field is up to with a personal touch
- use the resources like community Slack channels, Bluesky feeds (🧬!) and message boards. Advertise better.
- create norms around their use

Things we would like to see from funding bodies:
- Small-scale grants to facilitate student exchange. Just to cover travel+housing, other sources can cover experiments
- Intermediate grants for workshops. Bring people together to learn
- Long-term collaborations which extend over multiple students

There’s a lot that goes into being a good collaborator and good host for visiting students:
- Be welcoming and have a diverse group (both in expertise and background)
- Focus on outcomes. Use your unique skills and make it easy for others to use your output

These were really interesting conversations with an international group of DNA researchers on how we can improve collaborations. 🧬🧪

Some takeaways:

1/6 One of the key features of functional proteins is their inherent structural flexibility. In our recent work at #ICML, we introduce flexibility to protein structure design! More in a thread below.

Code / Tutorial: github.com/graeter-grou...
Poster: W-109, Thu 17 Jul 11 a.m. PDT — 1:30 p.m. PDT

Rather alarming, but interesting in a “what is this new sensation??” kinda way.

I tried making ‘booch during the pandemic and idk what was in my scoby, but it felt like there were more interesting brain chemicals in there than just ethanol. Felt like I was see-through after a few cups of the stuff.

Can an AI model predict perfectly and still have a terrible world model?

What would that even mean?

Our new ICML paper (poster tomorrow!) formalizes these questions.

One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵

Seriously…

Anybody know if it has the same problem with accent marks in languages like Spanish and French??

I’ve noticed this too and it’s infuriating!

My theory is that because they’re different Unicode characters than the plain letters and much less frequently used in their (I’m assuming multi-lingual) sample data for misspellings, the algorithm sees them as very low-probability substitutions.

What solutions do you see to maintaining private cars as an accessibility option, while not also making them more convenient for everybody? It seems challenging to make the infrastructure and pricing reasonable for people with disabilities, while not also making it optimal for every individual.

Two lil numbers: 12 6

Have you managed to find the scientists yet? Really curious how that cluster looks.