Sign in

Lucas van Dijk

@lucasvandijk.nl
205 followers 386 following 67 posts

Bioinformatics Software Engineer @ @pacbio.bsky.social 🧬🖥️. Previously at the data sciences platform and bacterial genomics group @ @broadinstitute.org. Interested in algorithm development, human genetics, and the microbiome.

PostsRepliesMedia
Lucas van Dijk @lucasvandijk.nl · 29/08/2026
This is awesome news!
000
Lucas van Dijk @lucasvandijk.nl · 23/07/2026
What a huge achievement! 450 human haplotypes. PacBio Kinnex transcriptomic data, Fiber-Seq (nucleosome occupancy), and so much more!
042
Lucas van Dijk @lucasvandijk.nl · 11/06/2026
It’s really ridiculous how quickly it flags anything biology or bioinformatics
030
Reposted by Lucas van Dijk
PacBio @pacbio.bsky.social · 26/05/2026
SPRQ-Nx chemistry is now shipping worldwide, bringing sub-$300 HiFi genomes to the Revio platform! This global rollout cuts sequencing costs by 30%, making it easier than ever to scale high-throughput workflows Full press release here: bit.ly/49HwmcL #SPRQNx #PacBio #HiFisequencing #Revio
0147
Reposted by Lucas van Dijk
Alex Crits-Christoph @acritschristoph.bsky.social · 27/04/2026
oh that's very very cool:
13410
Lucas van Dijk @lucasvandijk.nl · 22/04/2026
Also, yes, consensus generation is implemented!
110
Lucas van Dijk @lucasvandijk.nl · 22/04/2026
I'm still working on semi-global alignment, though it's currently a "personal time" project so I can't give a timeline for that yet. Claude has been helpful though recently :)
010
Lucas van Dijk @lucasvandijk.nl · 22/04/2026
Is global alignment sufficient? I am working on a new POASTA version with a Ukkonen-style band doubling strategy here: github.com/broadinstitu... Global alignment seems to work pretty well, but not extensively tested yet.
github.com
GitHub - broadinstitute/poasta at banded-poa
Fast and exact gap-affine partial order alignment. Contribute to broadinstitute/poasta development by creating an account on GitHub.
200
Lucas van Dijk @lucasvandijk.nl · 13/04/2026
Mood
010
Lucas van Dijk @lucasvandijk.nl · 11/04/2026
Independent of “coding being largely solved” (which I think is definitely hype), coding is only one part of a software engineer’s job. I’d argue that system architecture and design is not yet solved, and in my experience current LLMs are still not great at coming up with the right data abstractions.
290
Lucas van Dijk @lucasvandijk.nl · 18/03/2026
HiFi sequencing of many, many metagenomes!
031
Reposted by Lucas van Dijk
Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · 11/03/2026
So what's the most interesting, or weirdest, application of hidden markov models (HMMs) that you've seen? I once built an HMM for sage grouse mating behavior, fit to data collected by a robotic grouse (pictured doi.org/10.1086/703518). Looking for other examples for a new lecture on HMMs.
The robotic female sage-grouse, shown here, was used to experimentally manipulate sage-grouse males’ courtship behavior via remote control.
[Photo credit: Anna C. Perry]
1714421
Lucas van Dijk @lucasvandijk.nl · 09/12/2025
Congrats! Seems like a great fit for you!
010
Lucas van Dijk @lucasvandijk.nl · 30/11/2025
That x86 ABI mug is going to be powering my mornings :)
140
Lucas van Dijk @lucasvandijk.nl · 28/11/2025
I am very excited to have started a new role as a senior software engineer at @pacbio.bsky.social! I’ll be joining the instrument analysis team developing high-performance software on instrument delivering those beautiful HiFi reads to you 😁
250
Reposted by Lucas van Dijk
Mike Schatz @mikeschatz.bsky.social · 14/10/2025
Getting excited for #ASHG25. The must see talk is tomorrow at 1:30 by Qiuhui (Iris) Li presenting the first large-scale integration of long-read genome sequencing with electronic health record data across >1000 All of Us participants meetings.ashg.org/event/ASHG25...
meetings.ashg.org
PheedLoop
PheedLoop: Hybrid, In-Person & Virtual Event Software
1439
Lucas van Dijk @lucasvandijk.nl · 07/10/2025
>1,000 human genomes sequenced with PacBio HiFi, offering many new insights into human structural variation and its associations with disease. www.medrxiv.org/content/10.1...
medrxiv.org
Population-scale Long-read Sequencing in the All of Us Research Program
The All of Us Research Program (AoU) is a national biobank seeking to enroll one million individuals in the United States to link genomic and biomedical data, including short- and long-read whole-geno...
010
Lucas van Dijk @lucasvandijk.nl · 27/07/2025
Easier might be to use “basedpyright” which has built-in uv environment detection support.
030
Reposted by Lucas van Dijk
Erik Garrison @thinks.lol · 17/07/2025
Postdoc position opening in my group! Research projects: pangenomes for diverse organisms, genome evolution, biocomputing, language models. Please reach out if interested!
12830
Reposted by Lucas van Dijk
Alexandria Ocasio-Cortez @aoc.bsky.social · 03/07/2025
I am grieving the barbarism that is going to unfurl from all this. People are going to die. Livelihoods gone. All to feed a corrupt kleptocracy. I see every day up close how different it is from the first time around. There are no guardrails. A disaster. I’m sorry we have to live through this.
34849983121746
Lucas van Dijk @lucasvandijk.nl · 09/04/2025
Exciting! Good luck tomorrow 🙌
020
Lucas van Dijk @lucasvandijk.nl · 17/02/2025
Spain spent 57 bln euros to build its entire high speed rail network. www.theguardian.com/commentisfre...
theguardian.com
Spain’s high-speed trains aren’t just efficient, they have transformed people’s lives | María Ramírez
A fiasco like HS2 could never happen here – our fast train network is so popular that no Spanish government would dare give up on it, says journalist María Ramírez
010
Lucas van Dijk @lucasvandijk.nl · 17/02/2025
The efficiency of the market
110
Lucas van Dijk @lucasvandijk.nl · 17/02/2025
What the actual fuck is this
030
Lucas van Dijk @lucasvandijk.nl · 11/02/2025
And now officially out in Oxford Bioinformatics :) academic.oup.com/bioinformati...
academic.oup.com
Fast and exact gap-affine partial order alignment with POASTA
AbstractMotivation. Partial order alignment is a widely used method for computing multiple sequence alignments, with applications in genome assembly and pa
1131
Lucas van Dijk @lucasvandijk.nl · 09/02/2025
Thanks to @jdidion.bsky.social for writing 99% of the tree-sitter grammar at the heart of this extension. I made some modifications to ensure compatibility with Zed's WASM plugin system and fixed a few bugs in the C lexer extension.
000
Lucas van Dijk @lucasvandijk.nl · 09/02/2025
For those using WDL for their bioinformatic workflows: I wrote an extension for the @zed.dev code editor enabling syntax highlighting and integration with wdl-lsp. It also supports bash highlighting in command blocks. Installable through the Zed extensions pane! github.com/broadinstitu...
Screenshot of the Zed editor showcasing syntax highlighting for the Workflow Description Language.
150
Lucas van Dijk @lucasvandijk.nl · 27/01/2025
A printed thesis! Did not expect it would feel so good to have the final version physically in your hands 🙌
120
Reposted by Lucas van Dijk
Nature Reviews Genetics @natrevgenet.nature.com · 27/01/2025
Diversity and consequences of structural variation in the human genome go.nature.com/4awEtaY #Review by Ryan L. Collins & Michael E. Talkowski @broadinstitute.org @harvard.edu @harvardmed.bsky.social
go.nature.com
Diversity and consequences of structural variation in the human genome - Nature Reviews Genetics
Collins and Talkowski provide a broad overview of structural variation in the human genome that covers their mutational properties, the dynamics of population genetics and functional consequences in d...
2267
Lucas van Dijk @lucasvandijk.nl · 23/01/2025
I feel this one 🤣
110
Reposted by Lucas van Dijk
qntm @qntm.org · 17/01/2025
COMPUTER SCIENTISTS: "k" means one thousand and twenty-four STORAGE MANUFACTURERS: "k" means one thousand DISPLAY MANUFACTURERS [thinking hard]: "K" means nine hundred and sixty
621811299
Lucas van Dijk @lucasvandijk.nl · 10/01/2025
That looks super interesting and I would read that in a heartbeat!
010
Reposted by Lucas van Dijk
Travis Wheeler @wheelerlab.org · 07/01/2025
Still in beta, so possible minor API changes ... but our Suffix Array library (in Rust) is ready for business. Builds on lovely ideas in CaPS-SA from @robp.bsky.social. Low mem construction. Spaced seeds. Optional crazy-low-mem search. More info on the horizon.
0115
Reposted by Lucas van Dijk
Ming Tommy Tang @tommytang.bsky.social · 07/01/2025
5/ In my experience, most of the time, you don’t need deep learning. Often, simple methods like linear regression or random forests work well — especially for moderate-sized datasets. Deep learning excels in image analysis & protein structure prediction.
3192
Lucas van Dijk @lucasvandijk.nl · 12/12/2024
I remember this blog: kobzol.github.io/rust/cargo/2... But it was released earlier this year, so not super recent.
kobzol.github.io
Making Rust binaries smaller by default
Have you ever tried to compile a helloworld Rust program in --release mode? If yes, have you seen its binary size? Suffice to say, it’s not exactly small. Or at least it wasn’t small until recently. T...
110
Lucas van Dijk @lucasvandijk.nl · 25/11/2024
Anyone insight how I can configure pylsp such that it finds and autocompletes packages installed with uv?
000
Lucas van Dijk @lucasvandijk.nl · 19/11/2024
ChatGPT when you ask it to write something clearly and succinctly: "Salutations and felicitations, sir, and a most jocund welcome to the sink. I am your electronic valet and household central processor. May I be of service, sir?"
010
Lucas van Dijk @lucasvandijk.nl · 19/11/2024
Found ChatGPT in Fallout New Vegas: "Are you some kind of AI?" "Regrettably not, sir. All modules in this habitat are synthetic personalities atop a mundane operating system. There is no intelligence here, sir." www.youtube.com/watch?v=4XJw...
A view of the Central Intelligence Unit in Fallout New Vegas: Old World Blues.
110
Reposted by Lucas van Dijk
Bluesky @bsky.app · 15/11/2024
💡 Tip: You can set your website as your username on Bluesky! This is one form of self-verification, and we especially encourage official organizations and brands to do this! Examples include @npr.org @react.dev @duolingobrasil.com.br. Here's our guide for how to do this:
bsky.social
How to set your domain as your handle - Bluesky
Using a domain as your handle helps with account identity, verification, and portability. Here's how to set your domain as your handle.
1445366396770
Lucas van Dijk @lucasvandijk.nl · 16/11/2024
“Liberal democracy,” he says, “offers moral constraints without problem-solving” — a lot of rules, not a lot of change — while “populism offers problem-solving without moral constraints.”
000
Lucas van Dijk @lucasvandijk.nl · 16/11/2024
This piece that includes a conversation with a professor of autocracy who’s an Hungarian exile has some strong points in my opinion (gift article) www.nytimes.com/2024/11/15/o...
nytimes.com
Opinion | This Is the Dark, Unspoken Promise of Trump’s Return
Liberal democracy offers moral constraints without problem-solving. Populism offers problem-solving without moral constraints.
210
Lucas van Dijk @lucasvandijk.nl · 15/11/2024
Every time I watch a HBO doc on a cult (The Vow, Love Has Won, Breath of Fire, Synanon, …), I’m taking notes 📝
020
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
Tagging for 🧬🖥️
000
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
It took me nearly a week to figure out why the BAM generated by a coworker gave different results than one that I generated myself 😱 To prevent this, we manually specified the fragment length mean with BWA's -I flag in future analyses.
100
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
If a downstream task filters improperly paired alignments (which is pretty common), it might filter accurate alignments that just happened to be in a batch with few other reads that aligned.
100
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
By increasing the batch size, successfully estimating the fragment length mean and std will be more likely! Thus, it is more likely to have fewer improperly paired reads.
100
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
Now, what happens when you increase the number of threads in BWA? It will multiply the read batch size by the number of threads, distributing read alignment over each thread and processing more reads in parallel!
100
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
In that case, -all- aligning reads in that batch will be marked as improperly paired, even if the alignments are accurate.
100
Lucas van Dijk @lucasvandijk.nl · 14/11/2024
Here's the kicker: if the species of interest in a metagenomic sample is at low abundance, occasionally so few reads will successfully align to the reference that BWA cannot estimate the fragment size mean and std.
100