I don't consider my own libraries that I've rewritten in Rust to be competing. They are either replacements, or parallel versions. If the rewriter is different from the original dev, it's great if the original dev is involved, but IMO that's not essential to add value.
As an algorithmic bioinfo dev myself, I do feel that Rust provides a clear benefit for many tools in the space (mostly CLI tools). As @lh3lh3.bsky.social points out, there are very real benefits for having the tool written in Rust, rather than just having Rust bindings. lh3.github.io/2026/04/17/t...
Glad I just added support for plex technologies to alevin-fry ;P.
That said, I do have many reservations about the rewrite approach. Specifically, the velocity and, maintainence questions & #of concurrent projects *not* reaching usable status. I think there can be real added value to a true rewrite, but I think we should find a good way to do this. 2/2
The purpose of this project (not mine) is explicitly to avoid bindings. Honestly, I see the value of real ports, and have written some of my own software (and one, for fun example of WFA-lib). Relying on FFI makes the build process harder, opens more safety bugs, & can degrade perf. 1/2
Finally, I also am in favor of disclosure of AI written code (and think the commit messages and other notes are a reasonable way to approach it for now). However, this is just a preference that I have (for several reasons), and the code has the quality it does, regardless of the author... 3/3
Several (though not all) of the repos have caveats about completeness / expectations. On one hand, I absolutely see the potential for bad outcomes, & retain a healthy skepticism. OTOH, I want to give folks the benefit of the doubt, and see openness and discussion moving in the right direction. 2/3
To be clear, I don't think that the README suddenly makes it great, but I think it provides useful information. The person/lab doing all of these rewrites also has a statement on the motivation and method : github.com/henriksson-l...
I think that sustainability & pace are very real questions. 1/3
Nice! Can we get a pic? ;P
Don't port without support!!
“Bioinformatics would be a more enjoyable field if more tools were written in Rust.”
Shout it from the rooftops
QCatch is now published in Bioinformatics (academic.oup.com/bioinformati...)! Great work from Yuan and Dongze for quality control and analysis downstream of simpleaf/alevin-fry (taking advantage of its structured AnnData output). Give it a try: github.com/COMBINE-lab/...
Rust 1.95!
* IF-LET GUARDS; LFG!!!!
* compile-time `cfg_select!`
* a bunch of stabilizations
blog.rust-lang.org/2026/04/16/R...
Ok, so now there's a document that's laying things out a bit more cleanly. I still have a lot of thoughts, but it's useful to understand the motivation and the goals here. Also, the document itself seems to have reasonable first principles: github.com/henriksson-l...
New lab server is accessible! Now, I have to decide on the drive/volume layout. I'm working with ~92TB of NVMe gen5 storage. Any suggestions on setups for such storage that has worked well for folks in the past?
How about calling it "The Best Lab (whose PI is not Dr. Best)".
Computeral Bioinfology is the root discipline. It undergirds both bioinformatics and computational biology, as well as computational genomics and, in fact, reality itself!
@pashadag.bsky.social : the LG is pretty sweet so far!
updates to piscem -> alevin-fry -> simpleaf incoming! Much faster probe based quant with *identical* results to prior versions. 2/2
Massive thanks to @jacksonweir.bsky.social for sharing some 10x Flex v2 data! Not only did it help us solve a latent bug(10x whitelist is reverse complemented), but led to a new special index for probe-based quant. I just quantified 2.3B flex v2 reads on my laptop CPU in 32 minutes! 1/2
And we couldn't even get the name mim...
With the heavy discounting on the first year of .bio domains right now, how can we not? It is decidedly not well-considered, but I will accept debate only in the form of a pull request
compbiovs.bio
Looks very cool, but name collision! link.springer.com/article/10.1...
The resolution is important IMO for sharp text. Contrast maybe less so. However the large physical size on a single monitor is a real win. What is sway? At work I use it on Cosmic Desktop with their built in tilin manager. It works well.
OTOH, I found that B&H has $300 off on the LG, making it $1,499 (pre-tax). That's a steep enough difference for a name brand (with, apparently, an identical panel to the Dell), that I think it's likely worthwhile. Also Dell & LG are (gently) curved, Alogic and INNOCN are not (I want the curve). 4/4
From what I can see though, it has a slightly lower contrast ratio (1200:1) and a slightly slower refresh (100Hz vs 120Hz). It's also marked as "frequently returned" on Amazon. 3/x
I got the dell at work and it is *phenomenal*. Honestly, I'd just go for it at home too, but it gnaws at me that the cheapest I can find is $2,200 and I know that it used to sell ~1,500-1,600. The INNOCN is by far the cheapest option, and the panel seems good (5k2k, IPS, etc.). 2/x
Great question. I don't know much about it. From what I can see there are really 4 options in this space (setting aside clearly wild knock off brands) : The Alogic 5k edge, the INNOCN 40C1U, the Dell 4024QW and the LG 40U990A-W. 1/x
I recently got a Dell 4025QW at work and it is phenomenal. Now I want the same size/res at home. I’m thinking of the LG 40U990A-W; it’s currently about $600 cheaper than the Dell. Does anyone have experience with this monitor? Would it be a mistake given the substantial $ diff?
All it takes to start a flame war with this group is one opinionated, but well-considered, manifesto :).