Rob Patro

Would be a cool list to start. The criteria would be e.g. widely used, not actively maintained, C/C++…

I mean, no existing model is going to do anything close to “one shotting” it, ofc. But this is what I mean by a small team driving it. Also, the ability to generate essentially unlimited test data by running the real STAR on data would be hugely helpful.

This is why I think a big push with a small team largely driving some state-of-the-art models for most of the legwork might be a great project. Huge community benefit, and a massive concrete success story for the agent (if it works).

Yes & no. No, in that we have a reference impl to work from, metrics for success are clear, & the original is completely open (obv.). Yes in that setting aside dedicated time for a rewrite is hard & one can’t easily justify much PhD student time on it (but cool research once the new version exists).

Or malicious incompetence!

$ jj describe -m "Hello, world!"

Well, Rust is already in the Linux kernel, so we're getting there. There's also RedoxOS (www.redox-os.org).

I wonder if this is the kind of project that @anthropic.com might be willing to sponsor? If so, please let me know how to get in touch!

GitHub - RagnarGrootKoerkamp/oxford-bioinformatics-template: Improved version of the Oxford Bioinformatics paper template Improved version of the Oxford Bioinformatics paper template - RagnarGrootKoerkamp/oxford-bioinformatics-template

Community service:
Updated the oxford-bioinformatics template repo because apparently the template changed again, and it's still the top google hit, and figuring out the right parameters remains messy, and they are still inconsistent.

github.com/RagnarGrootK...

Is it 1%? How large are the signatures? I guess that matters a lot.

Alternatively, you could dedicate one such bit for contiguous runs of keys of various sizes to trade off overhead for extra lookup rate.

What if you stored an extra bit per stored signature to signal if a key was discarded at this spot? That would increase size by 1 bit/slot (which relatively is non-trivial), but may allow you to often avoid the discard table lookup.

crates.io: Rust Package Registry

@mikelove.bsky.social : There was PyDESeq2, and now there is rust_deseq2! crates.io/crates/rust_...

This was not, e.g. the 2010 snowmageddon. The ice certainly made things more difficult, but they've also had *plenty* of time to deal with this situation. I think they need a clear strategy for better planning and response, and a promise to execute better in the future.

I think late last week was questionable, but I'm willing to give them that, even. But today, into the second week, I just can't see it. This is a failure of the county govt. to properly clear sidewalks, roadways and parking lots. Yet, universities & businesses have shown it's certainly feasible.

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.

Oh hey, your schools opened! That's a minor win. MCPS is continuing to shirk their duties & remain closed despite every business, university and residential facility being open, and despite transport busses running as normal. Nope, just schools closed over a week after a storm that left 7" of snow.

I think that all of the decision makers for @montgomerycountymd.bsky.social should be sent (at tax payer expense) to spend the rest of the winter in and around Boston, with a duty to report back what the shape of competence looks like, and a concrete (snowcrete?) plan to improve their operations!

Cool paper on representing a collection of sets via a spanning tree of their differences. This builds upon work by Bookstein ('91 )! as well as work we did in using this representation to compress color sets in Mantis MST. I think this repr. has many important applications! arxiv.org/pdf/2601.23240

No @mcdotnow.bsky.social schools; it is not reasonable to cancel school again tomorrow. You've and an *entire* week to perform snow and ice removal. Businesses are open, Universities are open, recreational activities & museums are open. Children need to return to school tomorrow! This is wrong.

I'm looking to update the SRA growth curve for the intro lecture of my computational genomics course, and I noticed that the data traces provided by SRA end in 2024 ... Does anyone have an idea what's up here or how to get more up-to-date data? www.ncbi.nlm.nih.gov/sra/docs/sra...

I cannot begin to convey how frustrating it is that different systems choose subtly different default installation paths (e.g. `lib` vs `lib64`, even though they are both 64-bit systems) when building the *same* project (in this case zlib-ng) from source. WHY?!?!

Really, looks like it has arm too, right?

E.g. here anaconda.org/channels/con...

Ok a new bioconda recipe question. Any idea why the build
system seems unable to find the package ‘zlib-ng-compat’, despite the fact that this package clearly exists on conda forge?

MCPS Explains Why Snow Days Are Not Virtual Learning Days - The MoCo Show Montgomery County Public Schools has  clarified why it is not shifting to virtual learning during weather related school closures, despite ongoing questions from families. According to MCPS, virtual i...

I have a better idea, why don't @montgomerycountymd.bsky.social schools explain why they've cancelled school for an entire week for ~7" of snow, and why it's not a deriliction of their duties? I think it's time for some new leadership, some legal action, or both... mocoshow.com/2026/01/29/m...

So the standard crate for this ik rust just requires replacing the archive struct with the multi-archive struct, so it's a trivial fix. But one needs to know what they are looking for!

Yes, not all code handles multi-archives out of the box directly. We had to be careful in our mim work to properly handle bgzf www.biorxiv.org/content/10.1... !

What nonsense. You do what you want, Carl, but I think silencing yourself for being called out for such imagined "transgressions" is counterproductive. Personally, I feel people going around policing such speech calling out fascism, and insufficient responses to it, are "problematic".

Frankly, it's probably time that @montgomerycountymd.bsky.social scools face some kind of legal action for their deriliction of duty. It's come to be expected, but it's so incredibly disappointing.