Cell biologists! This is a fantastic statistics primer custom made for you. Very practical, must-read! From Pollard lab five years ago: molbiolcell.org/doi/10.1091/mb…

Happy Friday! I want to elaborate on a thought provoking discussion from last night’s Kendall Square panel AI in drug discovery featuring Derek Lowe, @JenNwankwo, Alex Snyder, and me (thx Rhie Lim/CIC for organizing!) It should interest techbio/biotech VCs and founders-to-be: 🧵

For all the fancy single cell profiling methods out there, cell morphology is surprisingly informative… and ~free.

Derek Lowe has been sharing his opinions on biomedical research to a global audience for over 20 years. Here, with grace and good humour, he takes a tour some of his most spectacular misses. chemistryworld.com/opinion/being-…

Don't miss the next TriRhena Gene Regulation Club, a half-day symposium on gene regulation and related topics. Plus, you’ll get to visit our new building on the Novartis campus in Basel! 👇🏽

A great panel setting many things into a realistic (and still very exciting) perspective. AI can't generalize outside of the distribution it was trained on. I guess it all comes down to AI-assisted (versus AI-driven) and creating high quality relevant data for training.

Thrilled to share our „PRedictor Of PHEnoTypes“ model Prophet! Led by Alejandro Tejada Lapuerta & Yuge Ji, Prophet is a transformer-based model that predicts outcomes for unseen experiments. It aims to understands biology by learning across assays and phenotypes over 4.7M+ experiments.

Congratulations luke isbel ! Best wishes

🚀In our latest work we explored Parkinson’s disease molecular pathology. We found that certain small molecules can rescue multiple morphological features, induce mitochondrial uncoupling & decrease αSyn protein levels. Huge thanks to our amazing team! biorxiv.org/content/10.110…

1/ I'd like to share a short thread about our project to build a nationwide, distributed core-facility for bioimage analysis. We discuss this in our recent preprint: arxiv.org/abs/2409.15009

Just returned from a short, but productive trip to Chicago for the Chan Zuckerberg Biohub Network Chicago Bioengineering conference. I wanted to list some take home points I gathered from the meeting! (A thread).

Nobel Prize in Physiology or Medicine: Biology Nobel Prize in Chemistry: Biology Nobel Prize in Physics: Sounds like Biology #NobelPrize

Love this!

Great overview about some key innovations in science during the last 20 years. One of my favorites: ImageJ/FIJI. Possibly the software most responsible for democratizing image analysis for non-experts. nature.com/articles/s4159…

Interesting that the focus is on better data, rather than novel computational approaches/algorithms.

Excited to share our latest work on an image analysis paradigm that combines cell state and morphology analyses of patient biopsies at single cell resolution to uncover clinically relevant head-and-neck #cancer phenotypes! Cell cell.com/cell/fulltext/… tweetorial👇

Make sure you know which fibrils you are using...

This kind of naive thought is attractive because it makes the problem amenable to fashionable technology (AI). The real problem is we barely understand what many diseases even are And that’ll require basic science, which will seem esoteric or irrelevant until it really isn’t

CellPainting offers powerful profiling alongside Omics. The JUMP consortium tested >15K genes by overexpression, KO, or both. The data is shared as the genetic subset of the >100K small molecules tested by JUMP before. Congrats Anne Carpenter, PhD and team & my Ksilink colleagues!