NAs after protein group comparison

[froehlic...@gmail.com]

Dear MSstats developer team,

I have to do a power analysis on a pilot study in order to plan the real study.

It is DIA data analyzed in DIA-NN and I manually brought the data in 10 column format.

When I run the proteingroupcomparison I only get NAs for FCs pVals etc.

After dataprocess, the Abundance values were calculated, so I am not sure where the problem lies. 

I had the same problem before and Mateusz found that my annotation was wrong, so here I already checked the annotation twice for double annotated bioreplicates and even changed the bioreplicates in the data to make sure, there is no problem with the assignment of 

Condition <-> BioReplicate

I set up a really small reproducible sample, which I will attach.

Best, Klemens

[froehlic...@gmail.com]

I cannot attach the R script. Here is the script for the example:

#packages

library(MSstats)

#set working directory to current folder of R script

setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

# input

test2 <-  readRDS(file = "test.rds")

##### msstats analysis ######

msstats_dataprocess <- MSstats::dataProcess(raw = test2,

  normalization = F,

  MBimpute = F)

#investigate this

msstats_dataprocess$ProcessedData %>% View()

#contrast.matrix for comparison

comparison <-  data.frame(V1 = 1, V2 = -1)

colnames(comparison) <- levels(msstats_dataprocess$ProcessedData$GROUP_ORIGINAL)

rownames(comparison) <- c("Group.A vs Group.B")

msstats_comparison <- MSstats::groupComparison(data = msstats_dataprocess, contrast.matrix = comparison)

msstats_comparison$ComparisonResult %>% View()

[Mateusz Staniak]

Hi,

thanks for the example data, I'll see where the problem is

Kind regards

Mateusz

[froehlic...@gmail.com]

Hi Mateusz,

Thank you very much for looking into this! 

Please let me know, if you find anything.
Is there any way, that msstats could give an error message when and why NAs are returned in this step? I realize there can probably be a plethora of different reasons, but any pointer would be super helpful there.

Best, Klemens

[Mateusz Staniak]

Hi,

if you change the way you create comparison matrix to

> matrix(c(1, -1), nrow = 1, ncol = 2)

will it solve the problem? I  just tested that on older R version and it worked.

I can add a check for class of contrast.matrix parameter. It's hard to catch all possible errors, though, and error message are not always helpful. (That was the case here - it's way it took me some time).

I'll see what I can do.

Kind regards,
Mateusz

[froehlic...@gmail.com]

Hi Mateusz

as.matrix() fixed the problem too.....

Thank you very much for looking into this! I am sorry to have bothered you with something so trivial ....

Best, Klemens