xcms fillPeaks.MSW about xcms HOT 8 CLOSED

Hi, let's clarify what you mean by direct infusion. Is it a 30sec measurement of some syringe-injected sample ? Resulting in a raw file with, say, 30 scans ? MSW was designed for input files with exactly one spectrum. Yours, Steffen

from xcms.

Hi Steffen

It is a 40 second run of the syringe injected sample with 30 scans per sample.

Thanks for responding so quickly

Jon

from xcms.

Hi Steffen

It is a 40 second run of the syringe injected sample with 30 scans per
sample.

Thanks for responding so quickly

Jon

On 3 August 2016 at 11:56, sneumann [email protected] wrote:

Hi, let's clarify what you mean by direct infusion. Is it a 30sec
measurement of some syringe-injected sample ? Resulting in a raw file with,
say, 30 scans ? MSW was designed for input files with exactly one
spectrum. Yours, Steffen

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#45 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/AT0IbuMVo8SwiEFo3TWPLWkoaus4gvSWks5qcHPjgaJpZM4JbbrK
.

from xcms.

So the actual fix is to add a more meaningful error message if n>1 spectra
are present in the file, which fillPeaks.MSW is not designed to handle,
since it does not know from which scan to take the intensity.

from xcms.

Okay is that the same for the xcmsSet.MSW method as well? Is there a way of obtaining an averaged spectra of the scans before using the fillPeaks.MSW function in R

from xcms.

Hi Steffen

I have been looking into the original issue that I was having with the fillPeaks.MSW function and because the program doesn't know which intensity value to extract I have added a mean function in the code which appears to fix the problems. Instead of just taking the first scan intensity it will average all the intensities found in that mass range, but I was wondering if the same is true for the xcmsSet (method=MSW) if this also only takes the first scan and if so could the same fix solve the problem?

nppos <- nppos + 1
mzpos <- which(abs(lcraw@env$mz - groupmat[g,"mzmed"]) == min(abs(lcraw@env$mz - groupmat[g,"mzmed"])))
mmzpos <- mzpos[which(lcraw@env$intensity[mzpos] == max(lcraw@env$intensity[mzpos]))]
mmzpos<-mean(mmzpos)
mmzr <- seq((mmzpos-mrange[1]),(mmzpos+mrange[2]))
maxo <- max(lcraw@env$intensity[mmzr])
medMZmin <- median(peakmat[groupindex[[g]], "mzmin"])
medMZmax <- median(peakmat[groupindex[[g]],"mzmax"])
minMzpos <- min(which(abs(lcraw@env$mz - medMZmin) == min(abs(lcraw@env$mz - medMZmin))))
maxMzpos <- max(which(abs(lcraw@env$mz - medMZmax) == min(abs(lcraw@env$mz - medMZmax))))
into = sum(lcraw@env$intensity[minMzpos:maxMzpos])

Also if the peak intensity is only found in a fraction of the scans could this also be attributed to noise? and if so is it worth building into the function a scan number and fraction of scans the peak needs to be present to be counted as a real peak?

Thanks

Jon

from xcms.

The findPeaks.MSW was also just uses one spectrum,
our previous FT-ICR did essentially a single-spectrum.

Instead of trying to tweak xcms, I'd rather check if some other
feature extraction tool was developed for direct infusion data.
And only if that fails to modify xcms. RMassBank was designed
to extract data-dependent MS/MS spectra, and it also had
the multiplicity filter against noise that you just described.

Yours, Steffen

from xcms.

This is now also fixed with the new user functions.

from xcms.