Hey, has anyone checked in with the Harpers letter guys? They doing OK?
Leading the bioimage analysis facility at Human Technopole Milan. Formerly cell and developmental biologist at the NIH. Serial immigrant 🇦🇺➡️🇳🇿➡️🇺🇸➡️🇮🇹🇪🇺. Fan of dogs, open source tools, and gin.
1,079 followers 980 following 346 posts
view profile on Bluesky
Hey, has anyone checked in with the Harpers letter guys? They doing OK?
Come and work with us 5-year position available for a University Researcher in Bioimaging 1.11.2025-31.12.2030 abo.rekrytointi.com/paikat/index...
On the politics, that is. On the policy, I think he vastly underestimates how bad the situation is within the executive branch. It's by no means a settled question whether the government doing nothing is better than doing what it's currently doing.
This piece isn't bad, necessarily, it just falls into the pundit trap of "we can easily extrapolate from what came before to what will happen now". Nothing about the last 6 months leads me to believe that is a useful analytical framework in 2025.
From #ome-ngff-validator you can open open #OME-Zarr images in a bunch of viewers, and neuroglancer's rendering of multi-channel images just got a lot nicer: ome.github.io/ome-ngff-val... - See github.com/google/neuro...
If this isn't GOTY, there's no justice in this world.
Truly awful for sure
Here is the "pro res zoom" shot. Like... How?
Not going to get into the "what is a photo" debate, but I will say the zoom on the new pixel is really something. Here's a normal shot from my backyard. If you zoom way in on the top of the mountain you'll see a small village there.
3D script is such a great tool
-50 DKP
I guide each photon to the aperture by hand
Honestly I can't disagree this is phenomenal
God damn the environmental art team on this went hard
Ok let's see what all the hype is about
🧬🔬🧪 🎉 Our team at @czbiohub is thrilled to share TWO companion papers out today in @NatureMethods! 📦 Ultrack — robust, scalable nD cell tracking 🌐 inTRACKtive — a beautiful, open-source web viewer for lineage exploration Let’s dive in! 👇 www.nature.com/articles/s41... www.nature.com/articles/s41...
Abolish ICE not epigenetics.
ReSCU-Nets: Recurrent U-Nets for segmentation of three-dimensional #microscopy data. New tools developed by Raymond Hawkins, Rodrigo Fernandez-Gonzalez (@rodrigofg.bsky.social) and colleagues @utoronto.ca: rupress.org/jcb/article/... #Cytoskeleton #SystemsBiology #ComputationalBiology #Development
Saddened to hear about the death of Jim Hudspeth. What a great loss.
Reading a book on the history of the late 19th/early 20th century and I'm beginning to suspect this Kaiser fellow might be up to no good.
Hiking the western end of the Gothic line in the Alpi Apuani above Massa/Carrara. Taking a quick water break in an old German bunker
This is an amazing thread
It's typically served either with ligurian pesto or with a butter parmesan sauce. We're going to eat it with a pesto I made earlier today from the garden
Then you boil it in water for around 4-5 minutes
Testaroli is a pancake-like sheet cooked in a special pan (called a Testo). Originally this pasta is thought to come from the Etruscans, who pre-dates the romans. Step one is to cut it into pieces.
Cooking one of Italy's oldest (and least known) pastas, Testaroli di Pontremoli
Oh awesome we're going to put the dumbest motherfuckers in the country in charge of overseeing grants. www.whitehouse.gov/presidential...
Potentially related photos
So true. I never really got that show. And then I moved to Texas and lived there for 7 or 8 years. Then it was like "oooohhhhhhhhh..."
Small brushes Small brushes *Bangs table* SMALL BRUSHES
I felt a disturbance in the force
Small brush
Incidentally the sheer amount of willful and uncharitable misreading of this piece going on around here is, while not surprising, very disheartening.
Congratulations! You'll be missed at nature methods, best of luck with the new gig!
Like most of the things Ian writes, this is very, very good.
This week, we have Jörg Bewersdorf 🔬 visit us at @humantechnopole.bsky.social. He gave a fantastic talk on Monday and a very useful ad-hoc microscopy workshop today. The future of Pan-labeled Expansion Microscopy will be amazing! 🌟 Thanks, Jörg, for visiting & sharing! 🙏
📢Release alert! BoneJ 7.1.4 is out now, bringing 💥100x faster💥 GPU-accelerated 3D Skeletonization, part of the new BoneJ+ extensions. You'll need a token to try it out, from bonej.org/shop Drop us a line bonej@bonej.org to get a coupon for a 100% discount on your first token.
Looks like a useful new tool
Starting up OnlyLabs, where people can pay to watch researchers who used to have federal funding do Hot Uncensored Cutting Edge science experiments.
Beautiful work
Thanks again to the great #bioinfocore folks for having me, and then letting me crash their panel and discussion groups. I think there's a bright future for image analysis and bioinformatics cores to work together.
Super fun to be at ISCB/ECCB in Liverpool talking image analysis, AI, and core facilities. Great discussions with everyone here!
Btw, and in the spirit of #OpenScience, all my #Napari #bioimage analysis #tutorials are freely available — from basics to #segmentation, #tracking, and #clustering. Feel free to explore and share ✌️ 🌍 www.fabriziomusacchio.com/teaching/tea...
That whole sequence, from the cigar in his mouth as soon as he wakes up to his glorious day in the corps speech to the "look into my eye", is just a master class in character introduction.
Great to see @alexhego.bsky.social share instructions on a DL pixel‑classification problem we worked on at image.sc. If you’re working on a challenging 3D DL semantic‑segmentation problem and need more options than most tools provide, check out this repository. github.com/AlexHego/DL-...
Abolish the death penalty, except for people who watch TikTok videos on trains without headphones.
Yeah this is extremely bad data analysis.
⛴️ BiaPy just reached version 3.6.1 💫 (latest release)! github.com/BiaPyX/BiaPy... ➡️ v3.6.1 came right after (the more dense) v3.6.0 to fix some small bugs 🐞 Let me tell you a bit about what's new 👇
Some very interesting stuff here 👇🏻👇🏻👇🏻
🚀The napari team is excited to announce our brand new development roadmap — a community-informed vision for the future of napari! 🔍Read the blog to learn about our process and priorities: napari.org/island-dispa... 🗺️OR check out the roadmap itself on napari.org: napari.org/stable/roadm...
The last #GloBIAS seminar 'Fiji with Python power: the best of both worlds' by Curtis Rueden is now available on our YouTube channel: youtu.be/dJrn0qelBhg?... Curtis has also created a follow-up thread on image.sc: tinyurl.com/57w29ust a good place for further Q&A!
Key numbers in cell biology Having a sense of scale helps to think more rigorously and realistically about biological systems.
🔬API-first feature extraction for image-based profiling workflows If you need to obtain interpretable features from your segmented microscopy images, but want to do it in a fully automated way, we know the struggle. 1/6
Gast rula The mic rula The old schoola You wanna trip? I'll bring it to ya.
It's Reason so as always you have to ignore the petty bullshit political hits (why is it important that a social movement occurred "during the Barack Obama administration"? 🤷🏻♂️), but otherwise this is something that we all need to be having a more nuanced discussion about than we currently do.
I always knew there was something off about PCR. Take one transcript and shamelessly copy it so you can distribute it to the masses? Downright un-american if you ask me.
🔬👨💻📰 #SReD is out! Automated structural detection for #ImageJ & #FIJI, from nano to macro ✨🐘. No training data, no bias - texture analysis with GPU acceleration!⚡️ Brainchild of @afonsomendes92.bsky.social and adventure w @christlet.bsky.social lab + friends. Check: www.nature.com/articles/s41...
🐾 The BioImage Model Zoo just got a major upgrade! It has been fully redesigned, offering a new backend, model uploader, community curation tools, and integrated browser-based test runs powered by BioEngine. 🧪 Explore now 👉 bioimage.io
Look I'm just trying to reassure the future American refugees
Good news I've fixed the worst thing about Europe
Pritzker 2028: Yes We Kahn.
The great Kahn has proclaimed a Kurultai.
Skill issue, 8th army didn't need it.
Slightly darker black would also be acceptable.
Hey does anyone have access to a submarine and some black turtlenecks? Because I've got a great idea to never have to write a grant again. www.mining.com/germany-ital...
Follow up: How are you thinking about the fact that most non-trivial phenotypes are then likely to be out of distribution with respect to the training data for the virtual labeling?
Huh. I just saw it pop up on the "what's new" splash screen of Firefox so I assumed it was new. At least I'd never seen it before. Now I don't even have to pretend to keep my tabs in order 😅
I feel like there hasn't been much innovation in web browsers lately, but this is a game changer for those of us with, um, tab hoarding disorder.
I would argue though, if you're investing in a massive screen anyway, you're back in the territory of "it's probably better to just train your own model to predict mitochondria directly from your label free image".
And this from Beth
Re: last thread, I've thought about it a bit more and I think this is a place where it could be helpful
Yeah I guess you could think of it like that, but then most models are gonna output a probability map anyway. For this I guess it would be interesting to see how it fails (ie what do generated florescence images where the model isn't sure look like).
Yeah see this is a good one. If you already have a trained classifier that operates on, say, mitochondrial images and you wanna screen a bunch of stuff, then getting a fake mito image for free is useful
But, as a bioimage analyst, I know the first thing people are gonna ask me to do is segment their fluorescence images so... Why not just predict the segmentation and save me the work. 😿😹
BUT I see a use case argument like the following. We have tons of pipelines and tools for dealing with fluorescence images. Few labs can train a model to extract whatever they want from their label free images, but any lab could pipeline a fluorescence image. So it opens up the downstream workflows
No I totally get that, my question is more a downstream one. What do you do with those images once you have them,. If you're predicting FL from label free, then necessarily all the downstream information you would get from your FL Images is already there in the label free. Just predict directly.
Want to be clear that I'm not criticizing the work here at all. It's super cool stuff. But it's good to have an open discussion about this before people use it irresponsibly.
But otherwise if you're just gonna segment and quantify the fluorescence images anyway, what's the point of the intermediate images? Just train to segment the label free!
But what are you doing in your screening? Like I said in another comment, if the point is to plug I to an existing screening pipeline that relies on florescence, I guess that's a use case.
But why not train directly to segment the label free? Why generate the intermediate image at all? Is the point that you can then use standard segmentation tools on the virtual image? I could kinda see that.
Look, fine, this is a techically cool accomplishment, but I am still struggling to think of a use case for generating virtual fluorescent images. OK, you've got your fake fluorescence image. What are you gonna do with it? www.nature.com/articles/s42...
Are you looking for a postdoc? I will be looking for 2-4 people after the summer to join the lab !! Do not hesitate to reach out and disseminate !!
I'm an experienced hiker. Last summer I was out in the Alps with two other experienced hikers on a three-day and regular, non-AI AllTrails fucked us so badly that we had a serious discussion about whether/when we should call for rescue. This is a terrible idea.
This is a great little conference tip
All hail the machine god
Paper alert: "Rotated Fourier transform (RFT) enables the quantification of anisotropic structure in high-moisture plant-protein extrudates" by @drnjitram.bsky.social et al. With fancy #ImageJ/#FIJI plugin to analyze anisotropy in ANY image! #FoodMicrosopy, doi.org/10.1016/j.fo...
Becoming more and more convinced that the great Khan is the only democrat who understands how to play the game.
Proud to see some of my former NIH colleagues on the list of signatories here. www.standupforscience.net/bethesda-dec...
For a lot of reasons, I think this is the way the world is moving, and so what our french friends here are doing will be a useful template.
Passionfruit vine seems to have survived transplant and is doing well, making fruit
Big weekend of garden work at the old Italian village house. Garden fence built and orto planting finished.
Oh no
I am once again asking Cxcr4 to stop doing new things. Leave some jobs for the rest of the genes.
Happy Killing Nazis at the Beach Day to all who celebrate!
Of some interest: Advanced methods in bioimage analysis www.embl.org/about/info/c...
Hey bioimaging folks, join us for the next OME-NGFF Community Call on June 24 at 9:00 CEST and 17:00 CEST! Details and agenda on the one and only image.sc here: forum.image.sc/t/join-us-fo...
