Arjun Raj

An excellent introduction (for beginners) to Graph Neural Networks distill.pub/2021/gnn-int...

forget don't meet your heroes, don't live long enough to see your heroes try to engage with your area of expertise

One of my favorite papers from our lab. Not only does it show heterogeneity is important in development, but it also points to division of labor: clones divvy up tasks in the organism based on their relative ability to perform well at a given fate! I like to think of it as "embryo economics".

..and have a look at the highly related paper from @arjunraj.bsky.social biorxiv.org/content/10.1... with complementary experiments that reinforce the point and reveal multicellular #memory@arjunraj.bsky.social

Maybe creativity is just how big a bag of tricks you have; how many different tricks you've seen, applied in different contexts. And perhaps some people have a little bit more skill remembering all those tricks or applying them in different situations. - Arjun Raj, @arjunraj.bsky.social

Side note: If students are using LLMs to help them choose statistical processing pipelines, these skills/plugins may help to synthesize more rigorous choices (via @arjunraj.bsky.social). arjunrajlab.substack.com/p/keeping-yo...

Updated my guide on using LLMs with #Rstats (including #Rstudio, #Positron, and #VSCode options) to be more BYOK friendly and reflect changes in both VS Code Chat and Posit Assistant. github.com/tpavlic/docs...

Scientist goals: get @quantamagazine.bsky.social to write about our work. Check this super nice article about Sophie Dumont's lab poking mitotic spindles:

Why bioRxiv

Thank you. My colleagues and I are pleased that we developed a system that works and that 11 years and over 100 papers #gastruloid papers later, what we felt in the final paragraph of the paper is coming along. There is now a community improving and learning from the system.

I had two undergrads in my office today working on a computational project. My normal advice in research is to use tools to their fullest to get to the cutting edge. For the first time, I was reluctant. I worry that using coding agents would rob them of something. And I’m fully AI-pilled.

🧮 Check things that should add up ⚠️ Never ignore code warnings 📊 Track datapoints like the number of rows 🔍 Investigate outliers 🧠 If something doesn’t make sense, keep digging arjunrajlab.substack.com/p/transition...

AI is pushing computational biology into analytic abundance. The bottleneck is no longer generating results: It’s deciding whether to trust them. 📊 5 habits for sense-checking outputs, via @arjunraj.bsky.social 1/2

A white paper on Stem Cell Based Embryo Models #SCBEM #Innovation #Ethics #Policy, from the community to highlight the potential of this new area of research and its impact in #ReproBiology shorturl.at/dAWL2

napari 0.7.0 is available now! 🚀🎉 It's a BIG release so read the full release notes, with highlights: napari.org/stable/relea... We want to thank everyone who has worked incredibly hard on this release including our 11 new brilliant contributors 🤩 and the community for their support and feedback!

Introducing the Anthropic Science Blog. Increasing the pace of scientific progress is a core part of Anthropic’s mission. The Science Blog will feature new research and stories of how scientists are using AI to accelerate their work. Read the intro:

A question for biologists: we often talk about evolutionary novelty. What does “novelty” mean for you? What’s your favorite definition if any? Or just a rule of thumb for what is and isn’t novel.

Honoured and overwhelmed to receive the 2026 Waddington Medal. Science is a team effort, and I've been fortunate to work alongside exceptional people asking hard questions. Thank you to the BSDB and to everyone who has been part of the journey. bsdb.org/2026/03/24/2...

Untangling mechanisms underlying #CancerDrugResistance www.nature.com/articles/s41... @jess_ljx @arjunraj.bsky.social #Cancer #DrugResistance #TranslationalScience

@mukherja.bsky.social work on RNAPII clusters is in print today: www.nature.com/articles/s41... Congrats! TLDR: RNAPII clusters represent transcriptonally engaged molecules at single genes, not super-stochiometric assemblies. Please check it out and see the the thread below for a summary.

Scientific publishing: Rethinking how research is reviewed and published Review of how the loss of impact factor affected submissions at eLife - uneven drop across countries, but generally holding up remarkably well and shows a new model is possible elifesciences.org/articles/110...

Live imaging of macrophages and neutrophils in a transgenic zebrafish larva. Credit to Prof. Anna Huttenlocher Lab @uwmadisonmmi.bsky.social. #ZebrafishZunday 🧪

Blog post: transitioning to being a PI in the age of AI open.substack.com/pub/arjunraj...

One of my very favorite papers from the lab! Shows that individual cells can learn by forming memories. Amazing work by Jess Li!

A moment of introspection on how much the world has changed during the time between the initial submission and this (hopefully final) revision: "This paper seems so quaint now" "Let's think of it as having contributed an important bit of high-quality training data to Opus 4.7"

Ran an AI coding workshop with the lab. There was a palpable sense of sadness realizing that skills some of us have spent our lives developing (myself included) are a lot less important now. I see the future 100%, but I do think it's important to acknowledge this sense of loss.

The TIG1-high state exists in human lung club and ciliated cells, enriched in IPF patients and tobacco users. Different viruses target different states. Influenza A prefers KRT8-high cells. Arjun Raj @cp-cell.bsky.social doi.org/10.1016/j.ce... 🧪 3/3

Reffsin, Miller, Cherry & Raj et al. discover that intrinsic cell states determine which cells become infected with SARS-CoV-2. Using single-cell clone tracing, they identified a TIG1-high state marking highly susceptible cells. What makes these cells vulnerable? 👇🏻 1/3

Data and experiments move faster, but theory is what gives discovery direction, via @arjunraj.bsky.social arjunrajlab.substack.com/p/chance-fav...

We spend a lot of energy choosing the right research project. But an equally important skill is knowing when to walk away. Quitting isn’t failure: It’s how you make room for better science, via @arjunraj.bsky.social

libertypuzzles.com these are very nice

Anyone have a thought about this weird mol on my hand?

I am happy to share that I’ve been awarded tenure here at Columbia University. Alhamdulillah. I thank my wife, parents, mentors, colleagues, friends, and family. I thank my lab members, present and past, for entrusting their careers in my hands and making this a thrilling journey.

AI is awesome, and is most definitely a transformative technology.

To me, the mean loud position here ranges between ignoring the ai and wishing to fully ban it. The latter won't happen (I don't think even regulation is technically feasible), the former is massively unproductive. If scientists refuse to participate in the discussion,it will just be driven by others

So… Bluesky for science and X for Claude Code?

Blog post: Chance favors the (theoretically) prepared mind Data is big, machines are learning, so what good is theory anyway? Isn't most discovery driven by serendipity anyway, with theory mostly a "post-mortem"? I argue that this view discounts the value of theory. open.substack.com/pub/arjunraj...

Beautiful - recommended! Here, @sasolla.bsky.social recaps her decades-long journey from physics to neural networks (working with LeCun & Hopfield) to motor cortex, & and from industry (including Bell Labs) to academia, all driven by curiosity and awe (which flows from her voice). Inspiring!

i've re-implemented PoPE from scratch as part of a model overhaul (w/ Claude Code). i don't have clean ablations yet but it's working. took about two hours to re-write my entire model i think the PoPE idea is really cool and a great fit for scientific data

2025 was the year we stopped debating whether paying peer reviewers might work and started showing that it does.

Agreed. The sunken cost fallacy is real. What I tend to do is to put projects on the shelf and once in a while look at them again wondering whether new technology or data would allow us to make progress. Emotionally it feels less like quitting and more as postponement

Bad antibodies are such a huge problem. It’s not just the money wasted on buying them. It’s the amount of time spent trying to validate them.

Blog post: Just quit Quitting projects in science is hard, but we should be doing a lot more of it. open.substack.com/pub/arjunraj...

Our preprint "Predictive design of tissue-specific mammalian enhancers that function in vivo in the mouse embryo" is on bioRxiv: www.biorxiv.org/content/10.6... . Amazing collaboration by @shenzhichen1999.bsky.social, Vincent Loubiere (@impvienna.bsky.social,@viennabiocenter.bsky.social),... (1/2)

I just finished my 28th yr of teaching grad compbio. Following the inevitable trend, a 3rd of the course is now Deep Learning (DL). One activity we did was a deep dive into the AlphaGenome pre-print (www.biorxiv.org/content/10.1...). Question is, how do we evaluate it as a science paper? 1/n

We designed SpaceBar to be as easy to use as possible: Cells are labeled via lentivirus transduction, and barcode detection works with any standard imaging-based spatial transcriptomics workflow. w/ Yael Heyman and @arjunraj.bsky.social Check out the thread from the preprint for more info!