In VoisinneG/InteRact: Analysis of affinity purification/tandem MS data

knitr::opts_chunk$set(echo = TRUE)
library(InteRact)
data("proteinGroups_Cbl")

Generate metadata from column names

In the example dataset proteinGroups_Cbl, protein intensities are found in data columns whose names start with "Intensity". We identify such columns using grep()

idx_intensity_columns <- grep("^Intensity.", names(proteinGroups_Cbl))
print(names(proteinGroups_Cbl)[idx_intensity_columns][1:10])

Intensity columns are usually named using a pattern. Here the names of the cell type, of the experimental condition (time of stimulation), of the biological replicate and of the technical replicate are separated by the character _. We can use the function identify_conditions() to map conditions from intensity column names:

condition <- identify_conditions(proteinGroups_Cbl,
                                  Column_intensity_pattern = "^Intensity.",
                                  split = "_", 
                                  bckg_pos = 1,
                                  time_pos = 2,
                                  bio_pos = 3,
                                  tech_pos = 4)
summary(condition)

Import custom metadata

You can also import this metadata from a separate file. The package comes with one such file.

condition_custom <- read.csv( system.file("extdata", "proteinGroups_Cbl_metadata.csv", package = "InteRact") )
summary(condition_custom)

This is the right place to reorder conditions if needed:

levels(condition_custom$Stim.time) <- c("t=0s", "t=030s", "t=120s", "t=300s", "t=600s") 

Finally, metadata column names must be changed to match those obtained by calling identify_conditions().

names(condition_custom) <- c("column", "bait", "bckg", "time", "bio", "tech")
condition_custom$bait <- "Cbl"

As documented for identify_conditions(), column contains intensity column names (as they appear when calling names()), bait is a user-defined name for the bait, bckg contains the identity of the cell type, time contains the name of the experimental condition, and bio and rep contain the names of the biological and technical replicate respectively.

This custom metadata can then be passed to InteRact():

res <- InteRact(proteinGroups_Cbl, 
                bait_gene_name = "Cbl", 
                condition = condition_custom, 
                bckg_bait = "CBL-OST",
                bckg_ctrl = "Wild-type")

Note that we had to change parameters bckg_bait and bckg_ctrl according to the values taken by condition_custom$bckg. Custom names now appears in the interactome and in subsequent plots (as for instance in the title of volcano plots):

print(res$conditions)
print(res$replicates)

plot_volcanos(res, p_val_thresh = 0.005, fold_change_thresh = 3)[[1]]

VoisinneG/InteRact documentation built on May 17, 2022, 11:40 p.m.