Error in tibble package during Data Processing

[Laure Bastide]

Hello,

I have a new issu when I want to use MSstats as an external tools in Skyline, when I want to start the Data Processing, i have this warning message :

"** Loading the required statistical software packages in R .....
 
Registered S3 methods overwritten by 'tibble':
  method     from  
  format.tbl pillar
  print.tbl  pillar
Error : package or namespace load failed for 'MSstats' in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]):
 namespace 'tibble' 2.1.3 is already loaded, but >= 3.0.0 is required

 Can't finish analysis."

I desintall and reinstall MSstats and R files and I try to install directly the "tibble" package in R but I can't because the package 3.0.0 is not available for R version 3.6.1...

What can I do ? Thanks a lot for your answer !

[Mateusz Staniak]

Hi,

looks like Skyline developers need to make an update of R/packages/MSstats, please contact them.

Alternatively, you can use it outside of Skyline. MSstats will run as an R package on newest version of R without problems.

Kind regards,
Mateusz

[nic...@proteinms.net]

Laure,

I have been working on a new version of the Skyline MSstats External tool which uses R version 4.1.

You can test it out by using the Skyline menu item:

Tools > External Tool

and then push the "Add > From File" button and paste the following big URL into the File name textbox on the Open dialog:

https://github.com/nickshulman/pwiz/raw/Skyline/work/20210525_MSstatsRVersion41/pwiz_tools/Skyline/Executables/Tools/MSstats/MSstats.zip

If you (or anyone else) feel like it, it would be helpful if you could test this out and let me know if you encounter any problems.

If this new version of the tool seems to work, then I will update the tool in the Skyline tool store.

-- Nick Shulman

MacCoss Lab, University of Washington

[Laure Bastide]

Hey  Mateusz,

I did use MSstats in R as a package but I obtain a strange QCplot...

I used your script :

#Installing MSstats
if (!requireNamespace("BiocManager", quietly = TRUE))
  install.packages("BiocManager")

BiocManager::install("MSstats")

## Library MSstats and load MSstats input from Skyline in R

library('MSstats', warn.conflicts = F, quietly = T, verbose = F)
Dataframe <- read.csv("MSstats Input MS4MS_LibInitMethods.csv")
str(Dataframe)

## Delete iRT
DF <- Dataframe[-which(Dataframe$Protein.Name=="MS4MS_iRT"),]

## reformating and pre-processing for Skyline output.
quant <- SkylinetoMSstatsFormat(DF)
head(quant)

### Default option use
DataframeProcessed <- dataProcess(raw = quant,
                                 normalization = 'equalizeMedians',
                                 summaryMethod = 'TMP',
                                 censoredInt = "NA",
                                 MBimpute = TRUE,
                                 maxQuantileforCensored=0.999)

#Visualisation of processed data
#Quality control and normalization effects
### use type="QCplot" with all proteins
## change the upper limit of y-axis=35
## set up the size of pdf
dataProcessPlots(data = DataframeProcessed, type="QCplot", ylimUp=35,
                 width=5, height=5)

And I obtain this kind of QCplots with a Log10 of intensity (normally I should have Log2 and the median have strange behavior despite the normalization) (in attachment) and when I did it in Skyline after the help of Nick I obtained a more "normal" QCplots...

I do not understand where is the problem... If you can help me I will be grateful !

Laure.

[Mateusz Staniak]

Hi,

please use censoredInt="0" option with Skyline inputs. Let me know if there are other issues.

Kind regards,

Mateusz Staniak