CRISPR Screen Processing Scripts

R scripts for processing guide-Seq counts. Input expected to be raw counts from mageck count. Scripts should be run in order.

Scripts

PrepareScreenCounts.R

This script should be run in an interactive RStudio environment.
Functions for normalizing counts, checking positive controls, plotting PCA and plotting heatmaps.

Input: mageck count data table; guide library tsv
Output: normalized count (DESeq2), MAUDE compatible; bin statistics matrix, MAUDE compatible

Dependencies

library(DESeq2)
library(ggpubr)
library(ggplot2)
library(pheatmap)
library(RColorBrewer)
library(tidyr)
library(reshape)

MAUDEAnalysis.R

Perform single-sample MAUDE analysis using output from PrepareScreenCounts.R.
All ".tsv" files in "/Counts/" subdirectories are included. binStats.txt path is hard coded currently.

Input: */Counts/.tsv; binStats.txt
Output: MAUDE output at the element _guideLevelStats.tsv and gene level _geneLevelStats.tsv

Example

./scripts/MAUDEAnalysis.R

Dependencies

library(MAUDE)

PlotRankedGenes.R

Plot ranked gene significanceZ scores for all genes from MAUDE analysis. Genes with sigZ > 3 or < -3 are in red. A comma separated list of genes can be added for additional highlighting.

Input: directory to search; tsv suffix pattern; gene names for highlighting
Output: ranked gene .png

Example

./scripts/PlotRankedGenes.R <top_level_directory> <tsv_suffix> <optional_genes_to_highlight> ./scripts/PlotRankedGenes.R ./MAUDE/ _geneLevelStats.tsv Tlr4,Myd88

Dependencies

library(ggplot2)
library(dplyr)
library(ggrepel)

CorrelationPlotting.R

Generates scatterplots by plotting gene significanceZ score pairwise amongst all input files. Only the top "n" (second argument) most significant are plotted and used for correlation calculation. Some genes are hard coded to be included whether or not they are in the top "n" genes.

Input: comma separated list of MAUDE output tsvs; number of genes to plot; output file name
Output: output png with pairwise scatterplots of all inputs and pearson correlations

Example

./scripts/CorrelationPlotting.R <input1.tsv,input2.tsv,...,inputN.tsv> <number_of_genes> <output> ./scripts/CorrelationPlotting.R ./MAUDE/<subdir>/xxxx_geneLevelStats.tsv,./MAUDE/<subdir>/yyyy_geneLevelStats.tsv 25 ./outdir/xxxx_yyyyy_correlations.png

Dependencies

library(ggplot2)
library(ggrepel)
library(dplyr)
library(purrr)
library(tidyr)