xcms 3.0 roiList about tools-metabolomics

[lecorguille]

According to the xcms3 documentation: findChromPeaks-centWaveWithPredIsoROIs

roiList

An optional list of regions-of-interest (ROI) representing detected mass traces. If ROIs are submitted the first analysis step is omitted and chromatographic peak detection is performed on the submitted ROIs. Each ROI is expected to have the following elements specified: scmin (start scan index), scmax (end scan index), mzmin (minimum m/z), mzmax (maximum m/z), length (number of scans), intensity (summed intensity). Each ROI should be represented by a list of elements or a single row data.frame.

I tested with this file: roiList.txt

scmin	scmax	mzmin	mzmax	length	intensity
200	300	300	350	100	5000

You gave it the IPO output. Sorry for this naïve question, but is it designed for this purpose?
I can't see the expected field but I can modify the IPO output if needed.

[yguitton]

@lecorguille
for that peakpicking findChromPeaks-centWaveWithPredIsoROIs function it will be more efficiente to pass that ROIlist as an option (see help saying "If ROIs are submitted the first analysis step is omitted", the IF is essential that's mean that the ROIlist is not mandadory and can be automatically built by centwave.

that new centwave is one cool thing for xcms 3.0 as it is capable of automatically improving the peaklist using the peaks found by the first analysis step of centwave.

so please can you change a bit the .xml and put that ROIlist as FALSE by default and only experts will us roilists as input

see default parameters for that function centWaveWithPredIsoROIs
p <- CentWavePredIsoParam()
Object of class: CentWavePredIsoParam
Parameters:
ppm: 25
peakwidth: 20, 50
snthresh: 10
prefilter: 3, 100
mzCenterFun: wMean
integrate: 1
mzdiff: -0.001
fitgauss: FALSE
noise: 0
verboseColumns: FALSE
roiList length: 0
firstBaselineCheck TRUE
roiScales length: 0
snthreshIsoROIs: 6.25
maxCharge: 3
maxIso: 5
mzIntervalExtension: TRUE
polarity: unknown

Regards
Yann

[lecorguille]

@yguitton See #107
I was waiting for your approval

[lecorguille]

@yguitton Can I close this issue?