Valuable #immunology cell atlas: #scRNAseq + paired B & T cell receptor seq for 330,000 tissue-resident immune cells across 16 human tissues.
CellTypist: new & robust immune cells annotation algorithm, finding 101 immune cell types in 1,000,000 cells‼️
Why this is important👇🧵
CellTypist: new & robust immune cells annotation algorithm, finding 101 immune cell types in 1,000,000 cells‼️
Why this is important👇🧵
Assessing cell types in healthy human tissues is hard. That's why most human immune studies so far profiled immune cells circulating in the blood.
But we know that the multitude of immune cells residing in tissues play distinct roles in development & disease than blood cells 2/7
But we know that the multitude of immune cells residing in tissues play distinct roles in development & disease than blood cells 2/7
This study is the first to characterize in depth the single cell expression landscape of the immune system as an integrated cross-tissue machinery.
The resulting cell-level complexity is disentangled with CellTypist, a newly-introduced logistic regression-based framework. 3/7
The resulting cell-level complexity is disentangled with CellTypist, a newly-introduced logistic regression-based framework. 3/7
CellTypist identified 43 distinct immune cell types/states, with impressive resolution in classifying T cell states (e.g. effector and functional), B cells, or the myeloid compartment. 4/7
Identifying shared & unique phenotypes in tissue-resident immune cells hints to signatures distinguishing between conserved and specialized evolutionary pathways across tissues.
This has implications for disease & aging therapeutics enhancing both innate or adaptive immunity 5/7
This has implications for disease & aging therapeutics enhancing both innate or adaptive immunity 5/7
CellTypist is fast,well-documented & easy to run. It offers two annotation modes: high-hierarchy (low resolution) and detailed (high res).
We used it for annotating immune cells from both healthy and cancerous tissues. It performed remarkably well 6/7
celltypist.org
We used it for annotating immune cells from both healthy and cancerous tissues. It performed remarkably well 6/7
celltypist.org
Beyond the analyses included in this study,the data resource it generated is its most valuable contribution for the community moving forward.
Raw & preprocessed data can be downloaded and interactively explored here for cross-tissue immune phenotypes 7/7
tissueimmunecellatlas.org
Raw & preprocessed data can be downloaded and interactively explored here for cross-tissue immune phenotypes 7/7
tissueimmunecellatlas.org
Thanks to @teichlab & to all authors involved for providing these resources to the community, and for making CellTypist both well performing and a breeze to run.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
