Now that we figured out how to distribute AVX2-capable sassy binaries, it's time for the next question: what about AVX-512? Sassy2 is up to 2x faster with AVX512, but it's a bit sad if this is only available via building from source :/

This is a big part of why Claude-written code comments are so uniformly terrible. Just the most superficial imaginable explanation of what’s going on.

This is the scary one. The earth’s energy imbalance has never been so out of whack. And it’s increasing. The only way it can balance again without stopping emissions is for the globe to get hotter. (BTW: It’s also the empirical evidence some climate sceptics bang on about that it doesn’t exist)

New blog post, a short follow-up to the previous one: smallcultfollowing.com/babysteps/bl...

Surely all LLMs have a larger context window than this... One would at least hope for more than a HMM

"Everything is vague to a degree you do not realize till you have tried to make it precise" - Bertrand Russell - from https:// stevekrouse.com/precision

"Those who cannot remember the past computations are condemned to repeat them." -- George Santayana (1905), explaining memoization to his computer science students

Lessons for software development: UX, service, and know your customer

Den Haag vertelt ons dat de aanstormende energiecrisis geen reden is tot maatregelen in Nederland. "COVID blijft in Brabant". Hier waarom we absoluut wel aan de bak moeten want anders lopen we deze winter een enorm risico. En het is duidelijk genoeg wat we moeten doen: berthub.eu/articles/pos...

Niceee, was waiting for this one, and already used it in a few places 😆 Congrats @imartayan.bsky.social et al :)

5m tonnes of CO2 emitted in just 14 days of US war on Iran, analysis finds Exclusive: War in the Middle East is draining the global carbon budget faster than 84 countries combined www.theguardian.com/world/2026/m...

New blog post: Maximally Minimal View Types smallcultfollowing.com/babysteps/bl...

PanXpress: Gene expression quantification with a pan-transcriptomic gapped k-mer index https://www.biorxiv.org/content/10.64898/2026.03.19.712873v1

Hierarchical genomic feature annotation with variable-length queries https://www.biorxiv.org/content/10.64898/2026.03.15.711907v1

Russia's main TV channel is airing songs about how great life is without the internet. Getting even closer to North Korea.

Super Bloom: Fast and precise filter for streaming k-mer queries https://www.biorxiv.org/content/10.64898/2026.03.17.712354v1

Fixed my first (very simple) issue using copilot CLI! It works well, but also isn't very fast. The ~30 line diff was probably faster to do by myself, but being lazy is nice I suppose

Strong quote! "The process of organization is one of the strongest metamemory strategies. Cut that out, and you've cut out learning."

🚀 Hiring a Postdoc in Computational Genomics @ ETH Zurich Join the RADIANT-Dx project—a pioneering effort with UZH & University Hospital Zurich (with Ashley Rooney & Petra Dittrich)—to develop next-gen diagnostics using long-read sequencing & real-time analysis. 👉 www.jobs.ethz.ch/job/view/JOP...

Cute new idea in here: Existing ABB+ (green) minimizers prefer kmers starting with 1 and then as many 0 as possible (in the binary case). "Spacers" (orange) instead prefer kmers starting with 10 that have maximum distance to the next occurrence of 10. (For DNA, A maps to 1 and CTG map to 0.)

10-minimizers: a promising class of constant-space minimizers https://www.biorxiv.org/content/10.64898/2026.03.16.712052v1

PSA: The PDF you upload to biorxiv does not have to be the exact same as the one you submit.

Time, throughput, memory usage, and input size plot axes should nearly always be logarithmic! eg: 'closer to the asymptote at 0s is better' is bad because the best tools will all look the same. _If_ you do linear axes, then please 🥺 at least have them start at 0...

New Space-Time Tradeoffs for Subset Rank and k-mer Lookup https://www.biorxiv.org/content/10.64898/2026.03.16.712042v1

Eagerly awaiting the stabilization of ByteStr & ByteString! Great for printing ASCII-encoded DNA as actual ASCII, and transparent deref to &[u8]. (And no more need for String::from_utf8_lossy at each print site.) doc.rust-lang.org/std/bstr/str...

In jj, you have: - jj abandon: delete a commit - jj split: split a commit up - jj squash: combine commits - jj duplicate: copy a commit - jj rebase: move commits around - jj absorb: move contents of a commit to other commits - jj arrange: interactively modify the commit graph

WHAT?!? Turns out that rust's BTreeMap nodes store not only keys and values, but also both parent and (owned) child pointers! I always assumed that they are based on a flat array, like static search trees, but instead they're more like c++ map's red-black trees but with a higher branching factor.

When you deploy heavily LLM text, you currently have no way to prove that you actually read it and therefore cannot convince me to read it

Level up your #pangenomics game! Workshop, conference & biohackathon in #Memphis, May 11-15, 2026. Join us and get your hands dirty with sequence alignment, explicit and implicit graphs, cutting-edge methods, and AI. Register now! pangenome.github.io/MemPanG26/ #Bioinformatics #MemPanG26

Kamila Szewczyk, Sven Rahmann: Hecate: A Modular Genomic Compressor https://arxiv.org/abs/2603.15390 https://arxiv.org/pdf/2603.15390 https://arxiv.org/html/2603.15390

Sensitive and scalable metagenomic classification using spaced metamers, reduced alphabets, and syncmers https://www.biorxiv.org/content/10.64898/2026.03.13.711249v1

This is kinda bonkers. In addition, 20 amino acids were found on Ryugu, along with uracil and vitamin B3. These have been found in other instances, but since the samples from Ryugu were collected directly from the asteroid and delivered in sealed capsules, contamination on Earth could be ruled out.

I finally wrote up a paper for perbase! The process with @joss-openjournals.bsky.social was really great! github.com/sstadick/per...

I just discovered #snakesee from Nils Homer: github.com/nh13/snakesee. This is a nice approach to monitor all running snakemake workflows.

𝗣𝗼𝘀𝘁𝗱𝗼𝗰 𝗮𝗻𝗱 𝗣𝗵𝗗 𝗽𝗼𝘀𝗶𝘁𝗶𝗼𝗻𝘀 𝗶𝗻 𝗖𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗚𝗲𝗻𝗼𝗺𝗶𝗰𝘀 / 𝗔𝗹𝗴𝗼𝗿𝗶𝘁𝗵𝗺𝗶𝗰 𝗕𝗶𝗼𝗶𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗰𝘀 I am currently recruiting for both: 🔹 Postdoc position su.varbi.com/what:job/job... 🔹 PhD position su.varbi.com/en/what:job/... Please share with anyone who might be interested!

Sassy2: Batch Searching of Short DNA Patterns www.biorxiv.org/content/10.6... 🧬💻🧪 github.com/RagnarGrootK...

This was a truly gross event. Centrist intellectuals featuring as audience to legitimate an anti-democratic grifter in the service of entertaining the ultra-rich reactionary class. Can’t capture the “heterodoxy” movement better than this. One big grift, ultimately only helping the far right!

We have short read and long read, I'm waiting for tl;dr sequencing.

👀

These are the first asexual vertebrate species known to science go.nature.com/4s2N0dM

In the LP relaxation we drop the integrality constraint for our variables, and immediately we see tons of fractionality. Using twelve 80% fractional wall pieces around the grid boundary and two 20% wall pieces right next to the horse, the LP can make almost the entire field be 80% enclosed

enclose.horse is a game in which you enclose a horse. You start with a map like this with water, fields, and horse. dynomight.substack.com/p/horse

“There are people who are like, ‘I've done it. I'm going to put it out there (bioRxiv), get some feedback and then I’ll go to a journal,' says Sever." Def 💯 me and many people I know! Either I get feedback or not there is 1 week to a few gap between preprint and submission, priority is preprint!

It's a good day when the first item in your feed is your own work :) @rickbitloo.bsky.social was annoyed that scanning reads for all 96 rapid kit barcodes is bottleneck in Barbell, so he made Sassy2: 13x (150bp) to 4.6x (8kbp) faster than v1 by batch-searching patterns, and >100Gbp/s on 16 threads!

What's the point of a preprint server if they keep rejecting preprints and make it hard to update versions after revision. All we need is PDF-hosting with auto-assigned DOIs; how hard can it be. (Also the submission UX is a mess, unfortunately.)

⏰ Last chance! Register your submission with a placeholder abstract for #RECOMBSeq TODAY. Final paper due March 15, 2026, 23:59 AoE Submit now 👉 recomb-seq.github.io/seq2026/call...

Wow! What a tool: half million downloads from Bioconda - Congrats @robp.bsky.social and Combine-Lab! 👏 Happy to see SSHash plugged here.

The state of (rust) allocators in 2026 by @cetra3.bsky.social cetra3.github.io/blog/state-o...

Open question for anyone that develops/maintains of bioinformatics tools, especially those used for public health - what do you think is the right way to fund this long term? Grants are great for new research directions, but aren't really appropriate for most tool dev/maintenance.

1/ Our paper on Multi-Context Seeds is now out, with @tolyan.bsky.social spearheading the work and contributions from Nicolas and @marcelm.net. We introduce a new seeding concept that improves read alignment accuracy while maintaining speed. link.springer.com/article/10.1...