A mutation in Themis contributes to peanut-induced oral anaphylaxis in CC027 mice

This document describes how to reproduce heritability and QTL analyses of various peanut allergy-related phenotypes in a backcross between CC0027 (reacts orally to peanut) and C3H (does not react orally to peanut).

Environment prep

This project uses a Docker container to produce an environment similar to that used in the original analysis (e.g. R v4.2.1 and R package versions available on August 1, 2023). In order to run this container you will need Docker installed.

Build the docker container:

docker build . -t pnut 

Run the docker container, opening a terminal session within the container:

docker run -e PASSWORD=pw123 --rm -v $(pwd):/home/rstudio/work -p 8787:8787 -it pnut /bin/bash

Navigate to the working directory:

cd home/rstudio/work 

Prepare R/qtl file

Run the following code to produce a file in the format required by R/qtl.

Rscript rqtl-file-gen-pnutbc.R

This script will generate the following .csv file, which can be imported into R and analyzed using the Rqtl package: derived_data/Rqtl_CC27xC3H_BCv2.csv

Inbred analysis

Analyze inbred parent data:

Rscript parent-analysis.R

This script will generate:

• figs/parent-temps.png: plot of temperature trajectories in inbred CC027 and C3H mice (Fig. 1A in manuscript)
• results/inbred-parent-data-summary.csv: means and ranges for each phenotype by strain (Supp. Table 1 in manuscript)
• results/h2-from-inbred-parents.csv: heritability estimates from inbred parent data (Supp. Table 1 in manuscript)

Backcross analysis

Note: Add steps to produce genomic relationship matrix for backcross mice (will be used in heritability calculations). For now, plink files are included in source_data directory (.map and .bed produced by plink_file_setup.R, everything else produced by commands in plink.sh).

Generate a report of analysis performed on backcross mice, including heritability estimation and QTL mapping.

Rscript -e 'library(rmarkdown); rmarkdown::render("qtl-analysis.Rmd", "html_document")'

This script will generate a report, qtl-analysis.html, as well as:

• several figures in the figs directory (including figures that compose Figures 2-4 and Supplemental Figures 1-5 in manuscript)
• results/h2-from-bc-mice.csv: heritability estimates from backcross mice (Supp. Table 1 in manuscript)
• results/QTLsummary.csv: summary of QTL mapping results (Table 1 in manuscript)

Candidate gene analysis

First, unzip the compressed directory of VCF files for each QTL region:

unzip source_data/VCFs.zip -d source_data/

Then unzip the GTF file representing all genes in the mouse genome:

gunzip source_data/Mus_musculus.GRCm38.97.gtf.gz --keep

Create a map between Ensembl IDs and gene names to be used by later scripts. This will produce a csv file called GRCm38-gtf-genemap.csv which maps ensembl IDs to gene symbols.

Rscript make_ensembl_genemap.R

Count total genes in each QTL region:

Rscript count_genes.R

Count candidate genes in each QTL region. This loops over all QTL and runs the cand_gene_analysis.R script to count variants segregating between C3H/HeJ and CC027 (protein-coding and regulatory) within each gene.

bash cand_gene.sh

This analysis can take several hours. To run the above command on a SLURM-based high-performance computing cluster, run:

bash cand_gene.sh -m slurm 

This will generate two directories:

• results/cand-gene-analysis/summary: summary of candidate gene analysis for the region
• results/cand-gene-analysis/vardata: data on candidate genes in the region, including the number of total variants, regulatory variants and protein-modifying variants

Create a map between mouse gene symbols and human homologs:

Rscript make_mouse2human_genemap.R

Summarize candidate gene data in table:

Rscript summarize_cand_genes.R

This script will produce a summary of the candidate gene analysis results, in results/cand-gene-summary.csv (Table 2 in manuscript).

Themis analysis

This script will generate a report of analysis (linear regression, Tukey's honest significant difference, and PCA) on flow phenotypes from Themis-variant CC strains.

Rscript -e 'library(rmarkdown); rmarkdown::render("themis-analysis.Rmd", "html_document")'

This script will generate a report, themis-analysis.html, along with several figures in figs/fig5 (Fig. 5 in manuscript) and figs/supplemental/tukey/ (Supp. Fig. 6 in manuscript).