Check here for future ads for positions in our lab, where we'll will focus on dynamical models of omics data, particularly in genetic disease contexts. PhD and postdoc announcements here: ⬇️. PhD: Postdoc:

Contact info at .

Now, we're integrating this dynamical systems approach with spatial transcriptomics data ⬇️to make models of genetic disease 🧬. If you're interested in stuff like that and you want to be a postdoc or PhD student in our lab, send us a CV and short research statement! 🙏 (5/5)

Plus, we can use it on cool videos of developing lizards from @LANEVOL to predict how growing scales will look depending on the lizard's anatomy. (4/5)

That means on test data, TRENDy's reduced-order model can predict what new patterns will grow in unseen parameter conditions (like the mystery box), revealing the patterning landscape better than competing methods.😎 (3/5)

Our solution is TRENDy (, a data-driven dynamical systems model which converts complicated, hidden PDEs (left) into simpler, predictable ODEs (middle) while still preserving info about the underlying patterning dynamics (right). (2/5)

Thread about recent work and new opportunities in our lab! At #ICLR2025, we're presenting work with @mor_nitzan which asks how we can reveal the hidden patterning landscape in the parametric, spatiotemporal PDEs we encounter all over biology. E.g. what's in the mystery box? (1/5)