[SOLVED] Proteomics: Create a MSnSet class file with MSnbase

Proteomics: Create a MSnSet class file with MSnbase

I want to create a MSset file (proteomics data, data corresponds to spectral counts) but I get error messages and I am stuck (after reading manuals, helps, forums, etc).

You can get my files here: https://www.dropbox.com/sh/dw7zfgiku6cteba/AADP3U2yxB5LgXy5ykJYFf0ga?dl=0

Here is the code I have tried:

setwd("~/Desktop/analyse")

## The spectral counts data:

data <- as.character(read.delim("sc.txt", header=TRUE,sep="\t", row.names=1, as.is=TRUE))

## Feature meta-data:
fdata <- as.character(read.delim("fdata.txt", header=TRUE,sep="\t", row.names=1, as.is=TRUE))

## Pheno data:
pdata <- as.character(read.delim("pheno.txt", header=TRUE,sep="\t", row.names=1, as.is=TRUE)) 


library("MSnbase")

readMSnSet(exprsFile = data,
                  phenoDataFile = pdata,
                  featureDataFile = fdata,
                  header=TRUE)

This last command returns an error message

Error in file(file, "rt") : invalid 'description' argument

I have verified the following:

class(data)
[1] "character"
class(fdata)
[1] "character"
class(pdata)
[1] "character"

dim(data)
NULL

dim(fdata) 
NULL

dim(pdata) 
NULL

str(data)
 chr [1:15] "c(4, 6, 11, 4, 3, 6, 2, 9, 8, 14, 15, 2, 8, 16, 5, 0, 0, 0, 0, 1, 2, 0, 0, 2, 1, 1, 0, 13, 11, 5, 0, 4, 6, 116,"| __truncated__ ...

str(pdata)
 chr [1:2] "c(\"treatmentA\", \"treatmentA\", \"treatmentA\", \"treatmentA\", \"treatmentA\", \"treatmentB\", \"treatmentB\"| __truncated__ ...

str(fdata)
chr [1:9] "c(222, 273.06, 335.8638, 413.112474, 508.128343, 624.9978619, 768.7473702, 945.5592653, 1163.037896, 1430.53661"| __truncated__ ...

I have also tried, instead of "as.character()", to use "as.matrix()" for "data", and "as.data.frame()" for "fdata" and "pdata".

Dimensions match correctly and are not "NULL" in this case but it does not solve the problem because i get the following message:

Error in (function (file, header = FALSE, sep = "", quote = "\"'", dec = ".",      : 
  'file' must be a character string or connection

If I try:

all(rownames(pdata)==colnames(data))
   TRUE

I have tried to create my MSnSet file with the following (with initial reading as.character...):

MSnSet(data, fdata, pdata)
Error in (function (storage.mode = c("lockedEnvironment", "environment",  : 
  'AssayData' elements with invalid dimensions: 'exprs'

If i read the files "as.matrix" for "data" and "as.data.frame" for "fdata" and "pdata":

> MSnSet(data, fdata, pdata)
Error in validObject(.Object) : 
  invalid class “MSnSet” object: 1: feature numbers differ between assayData     and featureData
invalid class “MSnSet” object: 2: featureNames differ between assayData and     featureData

> row.names(pdata)[1:5]
[1] "sample1" "sample2" "sample3" "sample4" "sample5"
> colnames(data)[1:5]
[1] "sample1" "sample2" "sample3" "sample4" "sample5"
> row.names(data)[1:5]
[1] "prot1" "prot2" "prot3" "prot4" "prot5"
> row.names(fdata)[1:5]
[1] "prot1" "prot2" "prot3" "prot4" "prot5"

So I have no clue where the problem comes from. Any idea on how to create properly my MSnSet file ??

Many thanks in advance for your help.

SkyR

Solution

From the argument name and from the help page, phenoDataFile should be the path to the file, not the contents of the file. From your question, I would guess the argument should be phenoDataFile = "~/Desktop/analyse/pheno.txt"