T-cell-effectorness

This repository contains all the codes used for the downstream data analysis steps in the paper "Single-cell transcriptomics identifies an effectorness gradient shaping the response of CD4+ T cells to cytokines" For a detailed explanation of the aims and theoretical bases of these codes, please refer to our study in Nature Communications:

https://www.nature.com/articles/s41467-020-15543-y

All codes are written in R (either as R files or R markdowns) and have been organised by data type (see below for a detailed description of the data analysis steps contained in each directory). For further information about these codes, please either raise an issue or email [email protected]

Bulk RNA-sequencing

Contains all the codes used for analysing bulk RNA expression in T cells upon stimualtion in the presence of cytokines. The analysis steps contained in the repository include:

1. Integration of the counts table and metadata information into a single SummarizedExperiment object
2. Exploratory data analysis (i.e. normalisation and log-transformation, batch regression, pca and visualisation)
3. Differential expression analysis between resting and activated T cells, as well as between T cells activated in the presence or absence of cytokines

Proteomics

Contains all the codes used for analysing bulk protein expression (i.e. LC-MS/MS data) in T cells upon stimulation in the presence of cytokines. The analysis steps contained in the repository include:

1. Integration of the relative protein abundances table and metadata information into a single SummarizedExperiment object
2. Exploratory data analysis (i.e. batch regression, pca and visualisation)
3. Differential expression analysis between resting and activated T cells, as well as between T cells activated in the presence or absence of cytokines

Multiomic analysis of RNA and protein expression

Contains all the codes used for integrating RNA-seq and quantitative proteomics data. The analysis steps contained in this repository include:

1. Integration of RNA counts and relative protein abundances into a single "multiomics" expression marix
2. A multiomic differential expression analysis which takes into account both RNA and protein expression values for each gene. This analysis is based on the f-divergence cutoff index (fCI) method.
3. Identification of cell type specific "proteogenomic" signatures. That is, genes which are expressed specifically upon stimultion with a given cytokine, both at the RNA and protein level. The functions for this analysis are available in the R package 'proteogenomic' (https://github.com/eddiecg/proteogenomic)

Single-cell RNA-sequencing

Contains all the codes used for analysing single-cell gene expression (i.e. 3' 10X data) in T cells upon stimulation in the presence or absence of cytokines. The analysis steps in this repository include:

1. Deconvolution of cells of different individuals based on natural genetic variation. This step of the analysis is based on the Cardelino (now Vireo) algorithm (https://github.com/single-cell-genetics/vireo).
2. Integration of single-cell RNA counts and cell annotations from different sample into a single expression matrix
3. Exploratory data analysis of over 40,000 single T cells from different conditions (e.g. normalisation, log-transformation, dimensionality reduction and visualisation)
4. Unsupervised clustering of resting T cells
5. Unsupervised clustering of T cells stimulated in the presence or different cytokines and identification of cytokine-specific cell states
6. Comparison of T cell response to different cytokines at the level of cell populations. This analysis is based on the single-cell implementation of the UniFrac method (https://github.com/liuqivandy/scUnifrac)
7. Ordering of T cells into pseudotime trajectories, which reveals the existence of a "T cell effectorness" gradient
8. Analysis of the relationship between T cell effectorness and TCR clonality. This analysis is based on publicly available single-cell paired TCR-sequences
9. Comparison of T cell effectorness between resting and stimulated cells based on canonical correlation analysis
10. Modelling of changes in gene expression as a function of T cell effectorness and cytokines