Problems with CDF files in metaMS_plot

[jsaintvanne]

Hi,

When we use CDF files, it is not possible to obtain automatically their polarity... It's a problem during generations of TIC, BPC and EIC...

[jsaintvanne]

Add a warning in a first time, then correct it!

[jsaintvanne]

Problem for TIC :

A problem with XCMS 3.x for TIC :

on R command :

> xdata <- readMSData(rownames(resGC$xset[[1]]@xcmsSet@phenoData),mode="onDisk")
Polarity can not be extracted from netCDF files, please set manually the polarity with the 'polarity' method.
> chromatogram(xdata,aggregationFun="sum")
Chromatograms with 1 row and 1 column
        0205065.CDF
     <Chromatogram>
[1,]   length: 3060
phenoData with 1 variables
featureData with 1 variables

All is good when we make it with a terminal on R.... But when you run it with Galaxy I have the following error :

Fatal error: Exit code 1 ()
Error in names(vals) <- featureNames(object) : 
  attempt to set an attribute on NULL
Calls: getTIC2s ... chromatogram -> .local -> unique -> polarity -> polarity
Execution halted

==> It looks like this is searching for a polarity whereas we have the warning message when we make the readMSData function ! Strange things : in R command we obtain after readMSData a featureTable with 28 column and one named "polarity" with only NA values, but in Galaxy wwe have only 27 columns and no polarity one... Error associated :

Fatal error: Exit code 1 ()
Error in value[[3L]](cond) : invalid 'row.names' length
  AnnotatedDataFrame 'initialize' could not update varMetadata:
  perhaps pData and varMetadata are inconsistent?
Calls: plotUnknowns ... tryCatch -> tryCatchList -> tryCatchOne -> <Anonymous>
Execution halted

https://github.com/Bioconductor/Biobase/blob/master/R/methods-AnnotatedDataFrame.R#L21

Problem for BPC :

It's the same thing for BPC !

Problem for EICs :

And for EICs it is quite the same problem... When I would like to run plotEICs with a xsAnnotate object (same error in R command) :

Fatal error: Exit code 1 ()
Error in seq.default(minmass, maxmass, by = value) : 
  'from' must be a finite number
Calls: plotUnknowns ... xcmsRaw -> profStep<- -> profStep<- -> seq -> seq.default
Execution halted

When I would like to pass with XCMS 3.x and chromatogram function :

Fatal error: Exit code 1 ()
Error in names(vals) <- featureNames(object) : 
  attempt to set an attribute on NULL
Calls: plotUnknowns ... chromatogram -> .local -> unique -> polarity -> polarity
Execution halted

And that always works well when we are in a terminal on R....

Solution for TIC :

We just need a list with one column for rt and one column for intensities...
So we bind tic result and rt result for a OnDiskMsnExp object when we have a CDF file
See here : jsaintvanne/metaMS@65f2b20

Solution for BPC :

It is the same problem as previous. But we use bpi function to obtain the good intensities values.
See here : jsaintvanne/metaMS@65f2b20

Solution for EICs :

• We tried to pass with a xcmsRaw object, we can't because of error with xcmsRaw (Error in seq.default(minmass, maxmass, by = value) : 'from' must be a finite number)
• We tried with XCMS 3.x but there is the chromatogram error (only with Glaxy)...

We are searching to solve the chromatogram error with Galaxy because it works well when we are on a terminal on R!

Some help ?? @RomainDallet @lecorguille @yguitton

[jsaintvanne]

Sould be good in this PR #40