Data processing and analysis scripts for fetal immune atlas (OA link).
src
contains scripts and notebooks used for data processing and analysis.metadata
: contains metadata relevant for sample and cell annotations, pointers to raw and processed data, color palettes and groupings.tutorials
: contains tutorials for model re-use (details here)Browse all processed datasets, models and annotations at https://developmental.cellatlas.io/fetal-immune.
For primary tissue samples, gene expression count matrices from scRNA-seq (10X Genomics 3' and 5') are available as AnnData objects in .h5ad
format for the full dataset (including stromal, immune and low quality cells), and 7 lineage subsets. In all objects adata.X
stores gene counts corrected for background expression using CellBender.
Cell-level metadata, including cell type annotations, are stored in adata.obs
and can be downloaded in .csv
format here (note: here cell type annotations are stored in column anno_lvl_2_final_clean
). Sample-level metadata, including matching between scRNA-seq and scVDJ-seq libraries, can be downloaded in .csv
format here.
Gene expression count matrices in .h5ad
format and high-resolution images in .tiff
format can be downloaded from the data portal (see section Spatial Datasets). Here cell type results for cell type deconvolution analysis with cell2location are stored in adata.obsm
.
Combined gene expression and antigen-receptor sequencing data is available in .h5ad
format (see section VDJ Datasets). Here adata.obs
stores the data from the antigen-receptor libraries using the data structure used by scirpy and dandelion. To match scRNA-seq and scVDJ-seq library IDs, see sample metadata file.
In addition, contig files for all libraries generated by 10X Genomics cell-ranger and dandelion can be downloaded as tarballs (abTCR) (BCR)
For in vitro derived T cells from Artificial Thymic Organoid protocols the gene expression count matrix from scRNA-seq (10X genomics 3') is available in .h5ad
(download). Sample-level metadata can be found here.
Raw sequencing libraries are deposited in ArrayExpress
The imaging data for detection of immune cell progenitors in fetal tissues can be found on the BioImage Archive (S-BIAD515)
Trained models for integrated embedding with scVI and cell type annotation with CellTypist can be downloaded through our data portal. We provide the following tutorials to explain how to use these models for contextualization and fast analysis of new data:
If you use this data or code for your work, please cite
Suo C., Dann E., et al. (2022). Mapping the developing human immune system across organs. Science, 376(6597), https://doi.org/10.1126/science.abo0510
@article{suoMappingDevelopingHuman2022,
title = {Mapping the Developing Human Immune System across Organs},
author = {Suo, Chenqu and Dann, Emma and Goh, Issac and Jardine, Laura and Kleshchevnikov, Vitalii and Park, Jong-Eun and Botting, Rachel A. and Stephenson, Emily and Engelbert, Justin and Tuong, Zewen Kelvin and Polanski, Krzysztof and Yayon, Nadav and Xu, Chuan and Suchanek, Ondrej and Elmentaite, Rasa and Dom{\'i}nguez Conde, Cecilia and He, Peng and Pritchard, Sophie and Miah, Mohi and Moldovan, Corina and Steemers, Alexander S. and Mazin, Pavel and Prete, Martin and Horsfall, Dave and Marioni, John C. and Clatworthy, Menna R. and Haniffa, Muzlifah and Teichmann, Sarah A.},
year = {2022},
journal = {Science},
volume = {376},
number = {6597},
publisher = {{American Association for the Advancement of Science}},
doi = {10.1126/science.abo0510},
}
For any questions, please post an issue in this repository or contact by email ed6<at>sanger.ac.uk
or cs42<at>sanger.ac.uk
.