Roby Bhattacharyya

😬 🤨🫥

The gall to use genAI (which hallucinated citations, including fabricated ones attributed to the submitted manuscript's authors) to review a manuscript with a public peer review trail...

ah, the blissful days of Very Real Solicitations of a new preprint, from Very Real Not Predatory Journals.

Screenshot of a spam email from a made-up address with subject line "Reg: We Want Your Brilliance"

Get in line, "Celi..."

The “V” Word In early 2025, a child died from measles—the first measles death in more than 20 years. Thirteen months later, a public health official finally said the “V” word?

open.substack.com/pub/pauloffi...
More than one year passed before public health officials said the "V" word during a recent measles epidemic. Why?

🤯

Started, going, etc

But at least it's still V1 🤨

This is the warning on the blood pressure monitor my dad just got.

Exceptional work. Anybody know what the fuck it’s supposed to mean?

Last line is especially ominous

I read this as Joan Baez and... had some questions

A line graph showing the success rates for fiscal years 2015-2025 as a function of percentile score. The curve for fiscal year 2025 is substantially lower than for previous fiscal years.

With these results in hand, we can now reproduce the graph that @joshuasweitz.bsky.social posted.

It is still shocking, but we can now understand why it looks the way it does: The number of R01 and R56 awards were down while the number of applications were up.

16/20

A line graph showing the number of NIH R01-equivalent awards for fiscal years 2015-2025 as a function of percentile score. The curve for fiscal year 2025 lies substantially below the other curves.

Here are the results for the numbers of "R01 equivalent (R01 and R37 (Merit)) awards.

The number of awards for fiscal year 2025 falls off substantially faster than for earlier fiscal years.

10/20

a cartoon says hey everybody an old man 's talking while bart simpson looks on ALT: a cartoon says hey everybody an old man 's talking while bart simpson looks on

A short Bluetorial on plots of NIH application success rates as a function of percentile.

1/20

Not personally, but I know Jim Collins' lab at the Broad Institute has used it quite a bit. @jasonyanglab.org, any tips on a Seahorse for bacteria?

Let’s just say it was a slightly wonky but obviously bad-faith argument, and I thought too hard about pointing this out, then let it go with few regrets.

The reason I now regret that? Yeah, they’re in the administration now 🫠🙃

My whole life there was only one Twitter beef I regret abandoning.

move slow and repair things

I’ve been wondering this myself, from the sidelines

a penguin with a bandana on his head says " don 't worry i 'm a doctor " in white letters ALT: a penguin with a bandana on his head says " don 't worry i 'm a doctor " in white letters

For the next 24h, drop any infectious disease and/or vaccine-related questions in response to this post, and I will do my best to answer them. Answers are for information only, NOT medical advice.
#AskAnIDDoc March 2026 edition.

Yeah, it’s a wild Discourse atm

Yup, this is actively happening to every PI I know.

Below 10% paylines, you could apply 15 times - once per cycle, every cycle before your R01 lapses - and have <80% chance of getting 1 grant funded.

That's for 100% funding probablity at 10th percentile. 2025 was 50% probability at 10th percentile.

Unsustainable.

Lake Wobegone-ass funding lines.

Despite this, preprint submitted!

Filename ending "Paper_FinalV1_rpb"

Sure enough, tale as old as time:

filename ending in "Paper_FinalV1.docx"

Postdoc in my lab, planning ahead with "FinalV1" 😅... still, preprint ahoy!

Today I Will Look The Burrito Man In The Eye and other stories of manifesting success

If there's a good, accepted way to think through & "correct for" this problem in transcriptional time-series analyses, I haven't found it. We've built analyses to account for it, but only for specific (diagnostics related) applications; I haven't seen a good general solution.

And if a drug inhibits growth, while control samples keep growing, then what looks like "drug response" might just be the lack of transcriptional changes of growth, thus not specific to your drug, & often not interesting.

See this a lot in bacteria / antibiotic response, within hours. Non-trivial.

🥳🎉

The moon going from full moon to a red moon and starting to go back to a normal full moon, then my camera battery door fell off.

Here's my first edit on the March 3rd Lunar Eclipse. 15 moon shots, 1 foreground. Taken over a 3 hourish period.

My favorite from the Covid lit was ‘Cleanup on IL-6’