Lecture notes. Written from the material of StatOmics Sga21, licensed CC BY-NC-SA 4.0. These are notes, not a transcript: the material has been reorganised and rewritten. This adaptation carries the same licence, and the original is linked above.

40. Proteomics Software Setup

What this covers

This short chapter is not a lecture on statistics — it is the course's software setup page for the Proteomics Data Analysis 2021 (PDA21) module, telling a participant what to install before the sessions on quantitative proteomics data analysis can be followed hands-on. It assumes nothing beyond having a computer to install software on; it does not introduce any statistical concept.

R, RStudio, and the analysis package

The course's computational work is done in R, through RStudio, using a Bioconductor package built for the course: msqrob2. Three things need to be in place before a session:

1. R, version 4.1 or higher, from CRAN.
2. RStudio, the desktop interface for R.
3. msqrob2, installed through Bioconductor's package manager rather than install.packages, since it is a Bioconductor package:

if(!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
BiocManager::install("msqrob2")

msqrob2 is the package that fits the models used later in the course to test for differential protein and peptide abundance from mass-spectrometry-based proteomics data; installing it here is what makes those later analyses runnable.

Optional: a graphical interface

Participants who prefer a point-and-click interface over writing R code can additionally install msqrob2gui, a Shiny app wrapped around msqrob2. It depends on the remotes package to pull the app itself from GitHub rather than from Bioconductor's own repository:

if(!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
if(!requireNamespace("remotes", quietly = TRUE)) {
  BiocManager::install("remotes")
}
BiocManager::install("statomics/msqrob2gui")

This is an alternative front end to the same package, not a separate analysis tool — anything done through the GUI is msqrob2 underneath.

Checking the installation

The course supplies a short worked check: load a built-in example dataset (pe), collapse its peptide-level measurements into protein-level ones, fit an msqrob2 model of protein abundance against experimental condition, and print the fitted coefficient for the first protein.

library(msqrob2)
data(pe)
pe <- aggregateFeatures(pe, i = "peptide", fcol = "Proteins", name = "protein")
pe <- msqrob(pe, i = "protein", formula = ~condition, modelColumnName = "rlm")
getCoef(rowData(pe[["protein"]])$rlm[[1]])

Running this end to end — feature aggregation, then model fitting, then coefficient extraction — and getting output rather than an error is the intended sanity check that the local installation is ready for the course.

Avoiding installation altogether

For anyone who would rather not install R, RStudio and the packages locally, the course offers a pre-built alternative: an RStudio interface running inside an R Docker container that already has the Bioconductor proteomics packages installed, launched through a hosted Binder instance built from one of the course's GitHub repositories. This trades local setup for a browser-based session with the same environment already configured.

Sources

• docs/omics-statistics/statomics/sga21/software.md — the course's "2. Software for Proteomics Data Analysis 2021 (PDA21)" page, converted from software.Rmd in the statOmics/SGA21 repository (commit 0ad787d), licensed CC BY-NC-SA 4.0.
• No slides, transcript or exercises were supplied for this chapter; it covers only what that software-setup page contains.

---

← 39. Variance-Stabilizing Transformation for Poisson Counts · Contents · 41. Stage-wise Omnibus and Post-hoc Testing →