Skip to main content ↑ ↓
Topic: fillPeaks produces high intensities (Read 4940 times) previous topic - next topic

fillPeaks produces high intensities

Hello!

I was wondering if fillPeaks always calculates the right intensities. Here an example:

[attachment=0:1pmk1zva]findPeaks.png[/attachment:1pmk1zva]

[attachment=1:1pmk1zva]fillPeaks.png[/attachment:1pmk1zva]

The peak is correctly identified by fillPeaks but the intensity is to high compared to the peak from findPeaks.

sessionInfo:
R version 2.14.1 (2011-12-22)
Platform: x86_64-pc-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=de_AT.UTF-8      LC_NUMERIC=C             
 [3] LC_TIME=de_AT.UTF-8        LC_COLLATE=de_AT.UTF-8   
 [5] LC_MONETARY=de_AT.UTF-8    LC_MESSAGES=de_AT.UTF-8 
 [7] LC_PAPER=C                LC_NAME=C               
 [9] LC_ADDRESS=C              LC_TELEPHONE=C           
[11] LC_MEASUREMENT=de_AT.UTF-8 LC_IDENTIFICATION=C     

attached base packages:
[1] stats    graphics  grDevices utils    datasets  methods  base

best regards
Gunnar

[attachment deleted by admin]

Re: fillPeaks produces high intensities

Reply #1 –
Are you sure this intensity is a result of fillPeaks(), and not caused by a retention time correction/alignment problem ?
 A peak with an intensity of 2e6 should have been detected at the first place !

Re: fillPeaks produces high intensities

Reply #2 –
Hello Ralf,

we had a closer look into the raw data and may have found the origin of the problem:

Measurements were made with a LC/FTICR-MS system (LC coupled to  Thermo LTQ FT Ultra) which produces data with quiet a high resolution. In the raw-files with the high intensities is a very small peak, therefore not found by findPeaks. Unfortunately nearby (~60ppm and same RT) is a high peak which is later filled in by fillPeaks. That led to the clue that the step-width of fillPeaks was to wide. Since step is taken from the xcms-object we added
profparam = list("step" = 0.01)
to the xcmsSet-parameters.
Now fillPeaks looks in a more narrow mz-slice and, as far as we can say, chooses the right peak.

best regard

PS:
that also benefits EIC-diagrams which are more accurate now :-)