Diane Adjavon

Newest version of Cellpose can spot cell boundaries even in cloudy conditions Picking out individual cells in distorted microscopy images is now as easy as clicking a button. A new version of Cellpose – the popular tool that maps the boundaries of diverse cells in microscopy

🖱️ Picking out individual cells in distorted microscopy images is now as easy as clicking a button, thanks to Cellpose3: www.janelia.org/news/newest-...

Super excited to share our new method: Context-dependent Identification of Spatial Motifs (CISM)! 🚀 A two-step approach for uncovering fine-scale intercellular modules associated with human disease states from single-cell spatial data.

doi.org/10.1101/2025...

🧵

1/n

We are proudly hosting

𝐃𝐫. 𝐓𝐚𝐥𝐥𝐞𝐲 𝐋𝐚𝐦𝐛𝐞𝐫𝐭 @talley.codes (CITE, Harvard Medical School):

𝐒𝐢𝐦𝐮𝐥𝐚𝐭𝐢𝐧𝐠 𝐫𝐞𝐚𝐥𝐢𝐬𝐭𝐢𝐜 𝐥𝐢𝐠𝐡𝐭 𝐦𝐢𝐜𝐫𝐨𝐬𝐜𝐨𝐩𝐲 𝐢𝐦𝐚𝐠𝐞𝐬 𝐮𝐬𝐢𝐧𝐠 𝐦𝐢𝐜𝐫𝐨𝐬𝐢𝐦

When: 𝐌𝐚𝐲 𝟐𝟗𝐭𝐡 @ 𝟏𝟏 𝐚𝐦 𝐄𝐃𝐓 (Boston time)

Join us on Zoom! harvard.zoom.us/j/9762343464...

🚀🔬🦠 Releasing 🤖Cellpose-SAM🤖, a cellular segmentation algorithm with superhuman generalization 🦸‍♀️. Try it now on 🤗 huggingface.co/spaces/mouse...

paper: www.biorxiv.org/content/10.1...
w/ @computingnature.bsky.social 1/n

🥳 Happy #DNADay!

🧬 What better way to celebrate than with cycleHCR: one of our newest tools, developed by the Liu Lab, that uses a novel DNA barcode system to track hundreds of RNA & protein molecules in single cells within thick biological samples ➡️ www.janelia.org/news/new-met...

The tilt illusion arises from an efficient reallocation of neural coding resources at the contextual boundary | PNAS The tilt illusion—a bias in the perceived orientation of a center stimulus induced by an oriented surround—illustrates how context shapes visual pe...

Very happy to see our work finally in print!
www.pnas.org/doi/10.1073/...

TLDR: Tilt illusion is not a bug, but a feature of a well-designed visual system that maximizes information capacity adaptively based on spatial context. (1/6)

AI@MBL: Machine Learning for Microscopy Image Analysis | Marine Biological Laboratory The goal of this course is to familiarize researchers in the life sciences with state-of-the-art deep learning techniques for microscopy image analysis and to introduce them to tools and frameworks th...

How can AI help you to analyze your microscopy images?

Find out at the "Machine Learning for Microscopy Image Analysis" course at the @mblscience.bsky.social in Woods Hole from Aug 22-Sep 6!

Applications are due 🚨 April 30 🚨

Read on for more! 👇

www.mbl.edu/education/ad...

Igor Siwanowicz of @hhmijanelia.bsky.social captured these photos of pollen grains on the sigma of a morning glory flower. This is part of the early stage in creating a seed, it might also be the reason why you can’t stop sneezing!

I was wrong before, *this* was actually tailor-made for me.

Yes, the paper is cool...
BUT if you’re a fan of fly vision and neuroanatomy, the Supplemental Materials are the real treat 🎉
We made a catalog of all the cells and it's ✨beautiful✨
🧭 Interactive explorer: reiserlab.github.io/male-drosoph...

Connectome-driven neural inventory of a complete visual system - Nature A connectome of the right optic lobe from a male fruitfly is presented together with an extensive collection of genetic drivers matched to a comprehensive neuron-type catalogue.

I know there's a 𝒍𝒐𝒕 going on right now, but I couldn’t be prouder to share this long-incubated labor of love: the complete connectome of the male 𝐷𝑟𝑜𝑠𝑜𝑝ℎ𝑖𝑙𝑎 optic lobe 🧠🪰
🔗 www.nature.com/articles/s41...

tiny light orange ant cricket amongst many small black ants, as found beneath a flipped stone

in the crowd ... an ant cricket (possibly a desert ant cricket, Myrmecophilus manni) cruises amongst 100s of ants beneath a flipped stone

oak woodland near Warner Springs CA

the thomas sanders x dropout combo feels tailor made for me

I will be, next week :)

When/how long is this course? I might stop by and say hi!

Join Janelia Senior Group Leader & Head of Computation & Theory Stephan Saalfeld @herrsaalfeld.bsky.social this Thursday at #ScienceOnTap at Old Ox Brewery, hosted by @loudounlibrary.bsky.social

🤖 Learn how #AI aids in understanding the complexity of biological systems & their interactions.

YouTube video by Simon Clark How much progress have we made on climate change?

It can feel like we have made no progress in the fight against climate change.

That's not the case.

youtu.be/h1jOqyjcO4g?...

Large Concept Models: Language Modeling in a Sentence Representation Space – Janelia CVML Summary of and thoughts on Meta’s recent paper

We read "Large Concept Models" in our reading group. It models language at the sentence resolution - "concepts". We had lots of thoughts about it! You can read them in our new Janelia CVML blog post:
janelia-cvml.github.io/blog/posts/L...
#ML #AI #Janelia @adjavon.bsky.social @bogovicj.bsky.social

The preprint is now available! Check it out... and... if you want to repost, why not 👇the post down here👇, so others get the bioRxiv link right away? Oh, and... 𝐭𝐡𝐚𝐧𝐤𝐬! 🙏

doi.org/10.1101/2025...

bsky.app/profile/flor...

Schematic overview on how MicroSpit operates. It can split superimposed structures in fluorescent image channels by using a suitably trained AI.

Concrete example for how MicroSplit unmixes four superimposed structures.

Immagine you could image two cellular structures in the same fluorescent channel and still reliably get them separated afterwards…
What would you do with this?

Now… what would you do if that also worked with 4 structures at once? 👇 #MicroSplit #preview🧵

📢 Submissions are now open for the CellMap Segmentation Challenge.

Build the best method for segmenting cellular organelles in vEM using our 289 annotated volumes, 40 organelle classes, and 22 diverse cell and tissue types.

🔗 hhmi.news/410evJQ

Different year, different conference, same vibe.

If you want a narrative version of this project, check out our blog post on the newly re-minted Funke Lab website!
funkelab.github.io/blog/2024/qu...

Thanks to @bobqubit.bsky.social, @asbates.bsky.social, and @jefferis.bsky.social, & @janfunkey.bsky.social for this 🦆ing awesome collaboration!

6/6

A query synapse, of class GABA, is on the left with its classification. It is converted to a counterfactual synapse of class Acetylcholine. The changes are minimal and highlighted with a contour.

Using QuAC, we can translate a GABA synapse into an ACh synapse by making minimal changes (highlighted). The images are much easier to compare: now we see that the post-synaptic density got darker.

Now that we know what to look for, we can actually see this feature in the examples above!

5/6

Two sets of images of synapses in EM. The synapses on the left are GABA, those on the right are Acetylcholine. They are grouped in a 2x2 matrix each

We developed QuAC specifically for biomedical applications.

Here is our favorite example: it is really difficult to notice any obvious differences between EM images of synapses releasing different neurotransmitters (e.g., GABA vs. ACh). There is just too much going on!

4/6

Quantitative Attributions with Counterfactuals We address the problem of explaining the decision process of deep neural network classifiers on images, which is of particular importance in biomedical datasets where class-relevant differences are no...

QuAC (Quantitative Attributions with Counterfactuals) is an #XAI method to do exactly that: visualize the sometimes subtle differences between images of different classes by generating counterfactual examples with minimal changes.

www.biorxiv.org/content/10.1...

3/6

An illustration of a cell, labelled "real", that points to another cell with very slight differences labelled "counterfactual". The differences are highlighted by a contour.

Wouldn’t it be easier if we could make paired images that only differ in class-relevant features?

2/6

An illustration of two plates of cells, with a slight difference in the size of one of the organelles.

Can you spot the difference between the cells on the left and right?

Read on for the solution (and a preprint announcement)… 👇

1/6

📢📢📢
The 12th iteration of the workshop on Fine-Grained Visual Categorization will be taking place at #CVPR2025 this year:
sites.google.com/view/fgvc12/...

There is a separate call for 4 page non-archival and 8 page official CVPR proceedings papers.

PLEASE SHARE: Postdoc Opportunity in Montpellier France. 18 months starting May 2025: Generative AI for Studying the Influence of Habitats on the Diversification of Bird Color Patterns. Applications and information here: tinyurl.com/2368teak
@evoldir.bsky.social