search

sneumann / xcms

This is the git repository matching the Bioconductor package xcms: LC/MS and GC/MS Data Analysis
Other
178 stars 81 forks source link

Error: No MS level 2 spectra present #603

Open 13479776 opened 2 years ago

13479776 commented 2 years ago

Hi, all,

I use the xcms3 to get the XCMSnExp object with dataset including full scan or DDA mode. The table(XCMSnExp) function can give me the MS1 and MS2 scan information.

However, when i use the chromPeakSpectra() function, the error "Error: No MS level 2 spectra present" was happened. Is it possible to process and got the (MS2) spectra associated with the features the dataset with mixture scan mode? such as mzML with full scan and mzML with DDA mode.

bes wishes, hees

jorainer commented 2 years ago

If I got you right, you have an experiment with some files that contain only MS1 data and some files with MS1 and MS2 data? In such cases xcms will throw this error when you call chromPeakSpectra or featureSpectra. I will try to get that fixed.

The workaround at the moment is to make a subset of the data containing the files with MS1 and MS2 spectra and then call the functions on these. In the example code below I have a similar data set with some MS1-only files and some with MS1 and MS2 (with two different collision energies). Running the featureSpectra on the full data set results in the same error:

> table(msLevel(data_WP))

   1    2 
8361 1226 
> featureSpectra(data_WP)
Error: No MS level 2 spectra present.

The solution is to subset the data to the files with the MS1 and MS2 spectra. In my case the information on the type of aquisition (MS1 only or MS1 and MS2) is encoded in a column named "mode" in the pData of the xcms object. Thus, what I'm doing below is getting the index of the files with the MS1 and MS2 data and using filterFile

> data_WP$mode
[1] "FS"   "FS"   "FS"   "FS"   "FS"   "FS"   "CE20" "CE30"
> data_ce <- filterFile(data_WP, grep("CE", data_WP$mode), keepFeatures = TRUE)

Importantly, you need to use keepFeatures = TRUE, otherwise feature definitions will be dropped in the returned object. With that you can now call the function and you get the MS2 spectra for the features.

> feature_ms2 <- featureSpectra(data_ce)
> feature_ms2
MSpectra with 751 spectra and 2 metadata column(s):
                              msLevel     rtime peaksCount |  feature_id
                            <integer> <numeric>  <integer> | <character>
   FT0007.CP36983.F08.S0384         2  161.7605          3 |      FT0007
  FT0015.CPM42198.F07.S0200         2   83.6344         13 |      FT0015
                        ...       ...       ...        ... .         ...
   FT7040.CP37143.F08.S0418         2   175.959        156 |      FT7040
   FT7041.CP68730.F08.S0378         2   158.887         43 |      FT7041
                                peak_id
                            <character>
   FT0007.CP36983.F08.S0384     CP36983
  FT0015.CPM42198.F07.S0200    CPM42198
                        ...         ...
   FT7040.CP37143.F08.S0418     CP37143
   FT7041.CP68730.F08.S0378     CP68730

Works obviously also for chromPeakSpectra:

> cpeak_ms2 <- chromPeakSpectra(data_ce)
> cpeak_ms2
MSpectra with 751 spectra and 1 metadata column(s):
                      msLevel     rtime peaksCount |     peak_id
                    <integer> <numeric>  <integer> | <character>
  CP34717.F07.S0119         2   41.3219         60 |     CP34717
  CP34719.F07.S0095         2   37.5071         13 |     CP34719
                ...       ...       ...        ... .         ...
  CP68688.F08.S0459         2   180.698          9 |     CP68688
  CP68730.F08.S0378         2   158.887         43 |     CP68730

As said, I will fix that. But the workaround should also work for you.

jorainer commented 2 years ago

Note: a second solution would be to use return.type = "Spectra" or return.type = "List". In both cases the function should work on the full data set.