CAR T Cell STudy

In this study, we analyze the effect of a three-shot vaccination regime against SARS-CoV-2 on the human immune response repertoire. This repository contains the single-cell analysis of T cell receptor (TCR), Gene expression, Surface Proteins (Antibody Captured), and dextramer staining across 14 donors and 7 time points.

Data

• The raw sequencing data can be downloaded from GEO TODO (pre-sample) and TODO (post-sample).
• The cellranger output can be downloaded from the same repository and should be stored as ./data/infusion/sample{1-2}/sample{1-2}_feature_bc_matrix.h5, ./data/infusion/sample{1-2}/sample{1-2}_contig_annotations.csv, and ./data/patient/sample{1-2}/sample{1-2}_feature_bc_matrix.h5, ./data/patient/sample{1-2}/sample{1-2}_contig_annotations.csv.
• The processed and annotated data can be downloaded from Zenodo (TODO) and stored as ./data/02_annotated_cd8.h5ad (entry point notebooks 03)

Additionally, we use the following external resources (not provided by us, store in ./data/scores/):

• Genes for various T cell scores by Szabo et al. (Nature Communications, 2019) (https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-019-12464-3/MediaObjects/41467_2019_12464_MOESM7_ESM.xlsx)
• Genes for Cell Scores by Meckiff et al. (Cell, 2020) (https://ars.els-cdn.com/content/image/1-s2.0-S0092867420313076-mmc3.xlsx)

Reproducibility

To recreate the results of the paper:

git clone https://github.com/SchubertLab/CAR_tcell_study.git
cd CAR_tcell_study
conda env create -f environment.yaml

Following, the notebooks must be run (ideally in this order). Note, that there are issues with reproducing UMAPs across different machines, even when the same seeds and package versions are used. Results might therefore look slightly different. To fully reproduce the paper results, use the annotated data.

To separate multiple sequencing runs from the cellranger output in the folder ./analysis/:

• 01_01_preprocessing_pre_infusion.ipynb
• 01_02_preprocessing_post_infusion.ipynb
• 02_annotation.ipynb: Note, we do not provide the donors' genotype. You can skip the related cells, or recreate the genotype following the scSplit pipeline.
• 03_visualization_CAR.ipynb: If you use the annotated data object from Zenodo, you can directly start at this point
• 04_phenotyping_tables.ipynb

Citation

If you refer to this work, please consider citing the following paper:

TODO