🧬 2026 Internship: Marioni Group @genentech.bsky.social Seeking PhD intern for Deep Learning on large-scale genetic screens (Perturb-seq/Optical).
Focus: ML for single-cell + imaging to decode phenotypes.
📍 SSF (On-site) | 12-wk paid
⚠️ Apply via portal only: 🔗 bit.ly/45P7TQB
#CompBio #AI #Genetics
Thank you Mike! We did comparisons (not in the ms) with PBMC T cells where we see the "Rest" condition (NTCs) in our data resemble CM/EM CD4+ T cells, while Stim cells are similar to rare proliferating subsets. Didn't compare effector vs memory, mostly due to lack of good annotations.
How well does TWAS estimate a gene’s direction of effect on a trait? We think of this as an important stress-test for the accuracy of TWAS.
In a new pre-print, we find that TWAS gets the sign wrong around 20-30% of the time!
doi.org/10.64898/202...
1/n
Finally, big thanks to Biohub @biohub.org, in partnership with 10x @10xgenomics.bsky.social, Ultima Genomics and Psomagen, for providing tremendous support to make this possible, especially Bailey Marshall and Jonah Cool and the whole Billion Cell project team who helped us launch the effort.
Thanks to Alex and Jonathan for making the ultimate co-mentoring dream team. It's been a privilege to join a long history of successful collaboration between the two labs.
... Ryan Goto @rgoto.bsky.social and Lillian Petersen contributed to several aspects of aspects of Perturb-seq <-> LoF burden and Perturb-seq <-> atlas analysis. Much of our analysis was built on insights and experience from our intellectual guru Mineto Ota @minetoota.bsky.social.
We were supported by an incredible team of scientists across Gladstone @gladstoneinst.bsky.social UCSF and Stanford @stanford.edu. Jun Yan, Justine Reyes Retana and Reese Guitche gave key contributions on data generation...
I've been incredibly lucky to co-lead this project with the brilliant @ronghuizhu.bsky.social. Ron led the massive-scale experimental efforts that made this study possible, and is just as thorough and thoughtful outside the lab, asking all the right questions when analysing the data.
This is only the beginning: with the help from lattice team @biohub.org, we made all the data available. We hope this dataset will be a foundational resource for systems immunology, human genetics, "virtual cell" modeling efforts, and immunotherapy development.
This analysis allows us to (a) group autoimmune genes acting on the same pathways and (b) characterize disease-associated genes with limited functional data in T cells
We also examined regulatory effects on autoimmune disease genes discovered by GWAS. We found enriched regulator clusters and co-regulated (condition-specific) downstream genes.
Extending recent work by @minetoota.bsky.social, we found molecular perturbation effects explain genetic signals on complex traits like lymphocyte counts... but in a condition-specific manner! In fact, even different stimulation timepoints explain the effect of different trait-associated genes.
3) GWAS links genes to traits, but “how” is often a mystery. We think many genes exert their effects through trans-regulation of other genes. Perturb-seq allows us to build an interpretative framework to explain these trait associations.
We tested whether cellular state signatures could be modeled as a linear combination of perturbation signatures, and it worked! We used this model to nominate regulators of Th1/Th2 polarization and T cell aging, and to parse cell atlas signatures into co-regulated programs
2) International efforts like the @humancellatlas.org have generated vast single-cell atlases of human cellular states. But observational data doesn't tell us which regulators and pathways drive a specific transcriptional state. Can we identify these with perturb-seq in primary cells?
Furthermore, Ron explored how perturbation outcomes grouped regulators that act coherently on transcriptional programs. Several known programs emerge, but the real magic is the context specificity: many programs and their regulators change drastically across different stimulation conditions.
1) Breaking target limits of FACS-based screening for immune cells, we explored the potential of Perturb-seq to map regulatory circuitry of many many target genes simultaneously. We focused on immune cytokines as a showcase. Come look up your favorite targets!
We view genome-scale perturb-seq in primary cells as an essential tool linking individual genes → gene circuits → cellular state & human health. We illustrate the power of this approach with a series of applications:
Gene regulatory networks encode the logic of cellular functions, but it has been hard to map these at genome-scale, especially in disease-relevant cell types.
Here we report the first genome-wide perturb-seq in primary human cells, and show what we can learn from it
Together with @ronghuizhu.bsky.social, we are thrilled to present our new perturb-seq study of 22M primary CD4+ T cells, across donors and timepoints – the result of a decade-long collaboration between the Marson @marsonlab.bsky.social and Pritchard @jkpritch.bsky.social labs 🧵 tinyurl.com/gwt2025
GWAS has been an incredible discovery tool for human genetics: it regularly identifies *causal* links from 1000s of SNPs to any given trait. But mechanistic interpretation is usually difficult.
Our latest work on causal models for this is out yesterday:
www.nature.com/articles/s41...
A short🧵:
Thank you Alex! Excited to see our paper published in @nature.com ! Huge thanks to @jeffspence.github.io , @tkyzeng.bsky.social , @emmamarydann.bsky.social, @nikhilmilind.dev, @marsonlab.bsky.social, @jkpritch.bsky.social, and all the members of the Pritchard and Marson labs for your enormous help!
@hakha.bsky.social and I wrote a Research Briefing (with a lay summary + "behind the scenes") of our paper on how genes are prioritized by GWAS and rare variant burden tests. 🧬🧪
www.nature.com/articles/d41...
Excited to see a home for MCP servers in bioinformatics emerge! The community needs a central hub for developing MCP servers for key bioinfo tools. At @scverse.bsky.social we're contributing to this effort - can't wait to share what we're building
I'm recruiting a postdoc for my group (based in beautiful Eugene, OR). Please get in touch if you're interested, esp if you'd like to chat at #ASHG25!
👉 and make sure not to miss the one-and-only Romain Lopez's talk about PhenoBridge, our joint work on ML approaches to integrate perturb-seq and genetic association data (Wed 10.45am)
Come find me or just reach out if you'd like to chat! [2/2]
En route to Boston for #ASHG25 #ASHG2025!
👉 I'll be presenting a poster on our new work on genome-wide perturb-seq screens in primary human T cells (5049W, Wed 2.30pm)
👉 you can hear me talk about it at the Industry Education session presented by Ultima Genomics (Thu 3pm) [1/2]
Last call, scverse community!
🐦 Early-bird closes October 17 for the scverse Conference 2025.
Single-cell, open science, and the people behind it — all in one place.
📝 Register here: www.eventbrite.com/e/scverse-co...
2 days of talks + 1 day of workshops
ℹ️ Info: scverse.org/conference2025
🚀 NEW PANEL ANNOUNCEMENT: Agentic Workflows in Bioinformatics at scverse 2025! 🚀
We're thrilled to introduce an exciting panel discussion at the upcoming scverse Conference 2025!
More information in 🧵
#scverse2025 #Bioinformatics #AgenticAI #Conference
🧬 Tahoe Workshop at scverse conference 2025 🧬
Disease Biology and Therapeutics in the Age of Frontier Datasets
🧵
@tahoetherapeutics.bsky.social
#scverse2025 #ComputationalBiology #DrugDiscovery #TahoeTherapeutics #SingleCell #PerturbationBiology
