Hi,
just a question about the peaks that will be presented in the tsv generated by the diffreport. I am comparing two groups (each having 3 samples). When I manually checked for the presence of a specific peak in one of the groups, I found it to be present in all samples of this group. Also the peak shape looked ok: peakwidth approximately 15 seconds and the retention time deviation between the samples was max 10 sec (determined at the apex of the peak). It's not a very abundant peak but clearly present (>10 scans, intensity > 1000 cpm). This is the script I used:
xset<-xcmsSet(method="centWave", ppm=2.5, peakwidth=c(10,90), snthresh=25,mzdiff=0.01,prefilter=c(3,1000))
xset<-group(xset, bw=10, mzwid=0.015,minfrac=0.5,minsamp=1)
xset2<-retcor(xset,method="obiwarp",profStep=0.1)
xset2<-group(xset2, bw=5, mzwid=0.015,minfrac=0.5,minsamp=1)
xset3<-fillPeaks(xset2)
Although the difference between the two groups was not statistically significant for this peak, I was expecting to find it at least back in the results.tsv file. Or am I wrong here?
Koen
p.s. I know that the peakwidth preferably should be something like c(15,90), but this also did not work out for this peak.
Yes, the diffreport contains all features, not just the dysregulated ones.