DESeqAid

Functions to make my DESeq2 life easier

Introduction | Usage | Examples | Reflections

Introduction

During my PhD, I did a fair amount of RNA-Seq analysis, often using DESeq2. I created this library to speed my mRNA-Seq analysis and make it slightly more modular. These are mostly designed to make my life slightly easier, e.g. by bunding my visualization settings into a single function, or processing the DESeq output in the way I like.

I rely on this library for my 16p paper code.

Many thanks to Hilary Parker for enriching the community by providing this easy-to-read guide to writing R packages.

Data exploration

• checkSVs(): explore whether your surrogate variables (from SVA) correlated with known sources of variance
• distPCA(): plots Principal Component Analysis
• plotPCA_pickPCs(): plots Principal Component Analysis using the DESeq2 method, but lets you pick which principal components to visualize
• sampSimilarityHeatmap(): heatmap of sample correlation

Functions for cleaning up the DESeq output a little

• deseq_coef(): DESeq output given a DESeq coefficient
• deseq_contrasts(): DESeq output given a contrast specified in design matrix
• add_lfcShrink(): adds shrunken log FCs to DESeq results table
• addHUGOnames(): if you did DE using ensembl genes, adds a column of corresponding HUGO gene names

Functions for visualizing DESeq output

• DE_heatmap(): makes a heatmap from a count table in the way I wanted at that time
• theoBarPlot(): ordered bar plot of gene expression, named after the advisor who encouraged its development
• fcSortBarPlot_lfcShrink(): The Theo Palmer bar plot, but using shrunken logFCs
• heatmapSexGenes(): heatmap of sex-linked genes
• makeVolcanoPlot(): volcano plot from DESeq output

Usage

Install

First - of course - clone the repo:

git clone https://github.com/kmuench/DESeqAid.git

In order to load this package, you'll need to install the devtools package. In R:

install.packages("devtools")
library("devtools")

Finally, in R, navigate to the same working directory as the DESeqAid dir. In R:

install("DESeqAid")

Great! Now it's installed, and if you want to load it in the future, simply add to your R script:

library("DESeqAid")

Examples

You can find example uses of this library throughout my 16p_resource code - for example,

• makeVolcanoPlot() in line 159 of the DESeq script for Figure 5.
• deseq_contrasts() in line 83 of the DESeq script for Figure 5.
• fcSortBarPlot_lfcShrink() in line 174 of the DESeq script for Figure 5.
• distPCA() throughout the setup script for Figure 6.

To-Do List

• Make the functions more flexible so that they're not so wedded to a particular formatting of data
• Include more flags in arguments that let you use different versions of similar code (e.g., labels or no)
• You call this modular, past me? This could be more modular