New paper! How do RNAs "know" where to go inside a cell? We dug into the sequence elements that route RNAs to the right place. It turns out that, in mammals, they're surprisingly massive (>200 nt), multipartite, and wonderfully complicated. 🧵

🌱 504k public plant RNA-seq samples × 1,200 species → live interactive UMAP atlas. Click any organ (256 PO terms) or evo-devo trait (57 of them) → top gene families upregulated pop up on the right. huggingface.co/spaces/mutwil…

For the first time, most biotech deal value now goes to assets from China:

No scaling laws for single-cell foundation models: when bigger atlases stop teaching the model anything In language and vision, the recipe has been simple: more data, bigger models, better performance. Single-cell biology borrowed that playbook. Foundation models for transcriptomics jumped from 1 million cells to atlases of over 100 million, on the assumption that scale would unlock the same gains. Alan DenAdel and coauthors put that assumption to the test, and the result is sobering. Working from a 22.2-million-cell corpus, they pretrained 400 models across five architectures (from PCA and a variational autoencoder up to the Geneformer transformer) and ran 6,400 evaluation experiments. They varied not just dataset size (1% to 75%) but also diversity, using cell-type re-weighting and geometric sketching to deliberately enrich rare cell types and transcriptional states. The finding: performance saturates almost immediately. On cell-type classification, batch integration, and perturbation prediction, most models hit their ceiling at roughly 1% of the corpus, about 200,000 cells. Beyond that, adding millions more cells changed essentially nothing. More diversity didn't help. Even spiking in genome-scale Perturb-seq data, to give the models perturbed phenotypes rather than just healthy ones, failed to move the needle. Larger models did score better overall, but they too plateaued early on data. Two points stood out. Simple baselines (PCA, logistic regression) often matched or beat the transformers. And the strongest model, SCimilarity, won not because of size but because its contrastive training objective is aligned with the downstream task. For single-cell data, what you train on and how you frame the objective matters far more than how much you collect. This reframes a quiet but expensive habit. In drug discovery, biotech, and any pipeline leaning on cell atlases, the instinct to keep scaling pretraining corpora may be burning compute for no return. The real leverage sits elsewhere: curating high-quality, task-relevant data and matching the training objective to the actual question you're trying to answer. Paper: DenAdel et al., journal license | doi.org/10.1038/s41592-026-0…

its a very small community in ai bio, everyone knows each other. we are working on human health, integrity is a big deal.

Done.

Big funding agencies / VCs / companies etc interested in powering true breakthroughs in AI for biology should fund tech dev efforts that can produce & scale the right types of data to push models out of zones they struggle in. 8/

I like this but the Takifugu rubripes genome is ~1/8 as big as ours, has about the same number of genes, has bones gut neurons etc, maybe start there?

I'm on a mission to make this plot as publicly recognizable as the one with the holes in the WW2 airplane

Another great research piece by @_DimensionCap @bauer_lesavage on Training Data for Bio AI. Models will only be as good as the underlying data, and the biology they learn will be constrained by the limitations of that data. We need to think deeply about scaling the best quality data to solve human bio with AI. Our thesis from day 1 at @NOETIK_ai.

This is such a banger of an ASCO

Animals have expanded the evolutionary legacy of unicellular ancestors in blood cells pnas.org/doi/10.1073/pnas.25…

For those of you who think macrophages do almost everything, rethink “almost.” They are programmable cells, recruited into whatever new function evolution needs. science.org/doi/10.1126/scie…

We built a joint experimental and computational platform for scalable multi-modal single-cell chemical screens — profiling RNA, protein (including phospho-signaling), and chromatin accessibility responses to thousands of small molecule perturbations in parallel. biorxiv.org/content/10.64898…