Skip to main content ↑ ↓
Topic: XCMS fill peaks question/problem (Read 4293 times) previous topic - next topic

XCMS fill peaks question/problem

Hello everyone!



I am trying to apply XCMS together with X13CMS to investigate some 13C-isotope data from an experiment.

The experiment consists of 4 samples, 2 which were treated with the 13C-substance. Another 2 samples are 12C-control samples, which were treated with the 12C-form of the substance

I am then using XCMS to construct the initial peak matrix (centwave, 5ppm [Orbitrap Exactive], 10-50 sec peakwidth)
I then call group and retcor functions to reduce the technical variability across my samples and the data matrix is then used for X13CMS


However, I observed a, for me at the moment, strange behaviour of the fillPeaks.chrom function. It seems as if the function integrates areas that are too far away from chromatographic peaks of interest.



I attached an example to that post. In the figure, the red dots represent the isotope of interest (very low abundant, but it seems like a chromatographic peak). This chromatographic peak is observed in only the 2 13C-substance treated samples, but not in any of the other 6 samples. Therefore, after xcmsSet, group, retcor and group calls I get the following output for the shown area of the samples

## R-Code
xset.aligned@peaks[xset.aligned@peaks[,"mz"]<=166 & xset.aligned@peaks[,"mz"]>165 & xset.aligned@peaks[,"rt"]<=1330 & xset.aligned@peaks[,"rt"]>=1310,c("mz","mzmin","mzmax","rt","rtmin","rtmax","into","intb","maxo","sn","egauss","mu")]
## Output
                mz     mzmin     mzmax           rt    rtmin     rtmax           into            intb   maxo   sn    egauss       mu
[1,] 165.0957 165.0951 165.0963 1321.84 1317.7 1327.00 39349.33 39341.07 5898 5897 0.1779203 1162.468
[2,] 165.0955 165.0952 165.0957 1323.90 1317.7 1328.04 35296.62 35071.81 6842   19 0.2430947 1164.754



When I then call fillPeaks.chrom to integrate the area in other files, I get the following information
## R-Code
xset.aligned@peaks[xset.aligned@peaks[,"mz"]<=166 & xset.aligned@peaks[,"mz"]>165 & xset.aligned@peaks[,"rt"]<=1330 & xset.aligned@peaks[,"rt"]>=1310,c("mz","mzmin","mzmax","rt","rtmin","rtmax","into","intb","maxo","sn","egauss","mu")]
## Output
                mz     mzmin     mzmax             rt    rtmin     rtmax           into            intb   maxo   sn    egauss       mu
[1,] 165.0957 165.0951 165.0963 1321.840 1317.7 1327.00  39349.33   39341.07   5898 5897 0.1779203 1162.468
[2,] 165.0955 165.0952 165.0957 1323.900 1317.7 1328.04  35296.62   35071.81   6842     19 0.2430947 1164.754
[3,] 165.1000 165.0951 165.0960 1326.686 1317.7 1327.52  613841.34           NA 71092     NA        NA       NA
[4,] 165.1000 165.0951 165.0960 1318.230 1317.7 1327.52  648458.75           NA 75408     NA        NA       NA
[5,] 165.1000 165.0951 165.0960 1317.702 1317.7 1327.52  722256.92           NA 75776     NA        NA       NA
[6,] 165.1000 165.0951 165.0960 1320.853 1317.7 1327.52  776177.15           NA 79314     NA        NA       NA
[7,] 165.1000 165.0951 165.0960 1320.370 1317.7 1327.52  790287.20           NA 81644     NA        NA       NA
[8,] 165.1000 165.0951 165.0960 1327.204 1317.7 1327.52  879661.13           NA 96006     NA        NA       NA



It seems as if fillPeaks.chrom has integrated the area in the 6 files (rows 2-8), which do not contain the chromatographic peak of interest (red). However, I now have into-values for that chromatographic peak in the remaining 6 samples.

I don't understand why these values are there and why fillPeaks.chrom has integrated them. There is an m/z difference of 125 ppm between the chromatographic peak of interest (red) and the noise (black), which is present in all 8 samples. In the above table the mzmax and mzmin values are also saying, that the black signals should not have been integrated as the m/z values are out of the mzmax and mzmin area.



What am I missing here? Why are there into-values for that chroamtographic peak after calling fillPeaks.chrom?



Thank you for your help in advance!
Best regards,
Peter


[attachment deleted by admin]