3/3. With MAESTRO we can:. 🔹 Apply controlled picoNewton-scale forces in multiple directions .🔹 Slow Holliday junction dynamics by >5× .🔹 Capture transient, previously unobservable, intermediate states .🔹 Observe interconversion between dynamic kinetic regimes.

2/3. In conjunction with MAESTRO, BNP-FRET, our latest software for analyzing #FRET datasets, is able to reveal hidden dynamics of DNA Holliday junctions. See the companion preprint:.

🚨 New preprint alert! . What if you could pull #DNA (or any other biomolecule) from all directions — and observe how its structure responds in real time under tension?. Meet MAESTRO, our latest work in collaboration with @result_hariadi. #DNAorigami

6/6 . ✅ Simulate complex, non-Markovian biological dynamics .✅ Train conditional normalizing flows to approximate intractable likelihoods .✅ Perform full #Bayesian inference on anything you can simulate. 📄

5/6 . We apply this to yeast expressing GFP under the glc3 promoter. 🌱 At first glance, high fluorescence seems like gene activation. But when you model protein inheritance across divisions. Most cells are actually inactive — just glowing their ancestors GFP.

4/6 . Despite the complexity, these dynamics are easy to simulate — protein production, cell division, fluorescence, all of it. So we flipped the problem: We train neural networks on simulations to learn the likelihood function itself.

3/6 . Because of that clock, division times aren’t memoryless -- they’re not exponential. This breaks standard models of gene expression, that is: .NO Master Equations .NO Fokker-Planck equations . We had rethink how we do inference.

2/6. In this new preprint, we analyze #flowcytometry data of stress regulation in yeast .🧬 We indirectly observe protein levels through fluorescence. But here's the catch: .1 - Proteins live much longer than a single cell cycle .2 - Cell division follows a biological clock.

No likelihood? No problem. The class of stochastic models we can simulate is A LOT larger than the ones we can write likelihoods. What if we could learn the likelihood directly from simulation? See the 🧵👇. #SimulationBasedInference #Neuralnetworks #AI

Overjoyed to be ⁦@EPFL_en⁩ with ⁦@LabRadenovic⁩ . Shall I be friend or foe in my talk tomorrow?.Ps. Talk title “ anomalous diffusion or anomalous state analysis?” 😶. Featuring work of Zachary Hendrix and Weiqing Xu.

Talking #raman deconvolution with ⁦@seungeun⁩ . An idea for how to extract SRS signal without lock-in amplifier, given a calibrated noise model #TellurideScience

Always fun in Telluride with ⁦@MaxPrigozhin⁩ & @CovinoLab. Talking simulation based inference, funding landscape, & more. 😊

Impossible to catch/tag everyone in this improvised selfie but such an amazing time in @telluridesci

Just another science meeting ⁦@TellurideSci . Why do you ask? 😜

Never seen this picture before—but beautiful to see. Rostropovich (the Soviet yo yo ma 😊) inspired both of Shosty’s cello concertos and championed his work across the globe.

Jokes aside this is extremely scary/.

55% planned cuts in NSF suggests no STEM field is critical. Are we spared?.

“US […] administration has ordered US embassies around the world to stop scheduling appointments for student visas…”. If we rely on US students, then dramatic STEM skill ⬆️ needed at HS level. Huge culture change also needed (eg, nerds rewarded).

Just received notification that our lab’s NIH MIRA was cut by 50% for this year. I am @ASU not Harvard…. Maybe more exam grading by lab members & less science … & more piano 🎹 for Steve. It will be a little quieter in the world of data modeling for biophysics…. @ASUResearch.

RT @ERC_Research: Today, a decision to increase ERC support for researchers moving to Europe was adopted. On top of normal grants, the E….