Congratulations to Chi-Jane who led our new paper on graph coarsening approaches for single-cell data. We show that by shrinking the graph in a principled way, you can achieve similar performance in downstream bioinf tasks as if using the full graph. biorxiv.org/content/10.110…

Do you want to get into scRNA-seq but are lost in the sea of never-ending papers? No worries, I got you covered. Here's a list of the best review that have help me understand the core concepts of the technique, analysis and interpretation:🧵

Scientists spend too much time taking “professional selfies”. A paper from PNAS describes what it takes to be a scientist today. Key points (that I always emphasize myself): #AcademicTwitter Academic Chatter™ #research

How to stay on top of the literature? Here are 5 tips to stay current, just follow along ⬇️

Interested in "integrating" multimodal #scRNAseq data? W/ Maria Carilli, Gennady Gorin, Yongin Choi & Tara Chari we introduce biVI, which combines the scVI variational autoencoder with biophysically motivated bivariate models for RNA distributions. 🧵 1/ biorxiv.org/content/10.110…

Just found out that python's default behavior is to prioritize loading packages in your base installation over your active conda environment... that was a fun bug to fix 🥲

So you just used a fancy deep-learning method (e.g., U-Net) to segment your cells. Do you trust your results? Should theorists model your data? Beware of “What You Put Is What You Get (WYPIWYG).” Your results will reflect your bias in your training dataset. Short thread. 1/

Congratulations Alec Plotkin for winning a best poster award at data science day for work predicting CD8 T-cell fates with Justin Milner :) Truthfully, this a picture from last week, but is the same poster. Great job Alec.

Very evocative figure. "We present an alternative to Graph Neural Networks (GNNs) termed Graph Structured Neural Networks (GSNN), which uses cell signaling knowledge, encoded as a graph data structure, to add inductive biases to deep learning."

.Dylan Schaff et al. show that resistant cancer cell clones are different from one another, yet the cells within a clone remain similar during treatment. Transcriptional differences between clones result in different cell fates when treated sequentially with different agents or

In this preprint biorxiv.org/content/10.110… with @sindri_e we compared seven widely used methods for batch correction of single cell RNA-seq data. We found that all but one of the methods introduce batch effects when there are none. 1/N

Are you puzzled about your batch correction practices after reading: biorxiv.org/content/10.110… . We were too and provide some insights here. Pall Melsted Adam Gayoso Yosef Lab Martin colab.research.google.com/drive/15k5kahl…

And now congrats to Dr. Haidong Yi for great contributions to featurizing single cell data to predict outcomes :) congratulations haidyi 🎉🍾🥳

.Jolene Ranek Natalie Stanley Purvis Lab @CompMedUNC present DELVE, an unsupervised feature selection method to identify core molecular features driving cell fate decisions. #BiotechNatureComms nature.com/articles/s4146…

I've been reading up on flow matching recently, especially wrt. single-cell data. It seems like it's basically fitting a generative model on top of optimal transport--can anybody tell me if my intuition is right?

Guidelines for T cell nomenclature bit.ly/48258LX

Guess it's time to rename the subsets in my manuscript for the 9th time

Online now: IL-9 as a naturally orthogonal cytokine with optimal JAK/STAT signaling for engineered T cell therapy dlvr.it/TPPPm1

A major advance in cancer using multimodal AI to do virtual proteomics of H&E slides, unraveling the tumor microenvironment, key proteins, biomarkers, and informing prognosis Cell Hoifung Poon Microsoft Research cell.com/cell/fulltext/…

🚀 New preprint! The Human Cytokine Dictionary - 9.7M cells, 12 donors, 90 cytokines - our largest human single-cell perturbation atlas yet. Users can map cytokine activity in their own data. Huge thanks to Lukas, Sören, Larsen, Parse Bio & Seelig lab! 🔗 biorxiv.org/content/10.648…