Open cbroeckl opened 3 years ago
For the first error: yes, indeed, that looks like a memory problem. What might also help is that you could do the centroiding and export the results as an mzML file again. The advantage will be that any further data processing will be faster. You could eventually have a look here, specifically the centroiding.R script which I use to do the centroiding on all of our (profile) mode data. What I do there is: load each single profile mode file, perform the centroiding and export that file again.
And also note (maybe relevant for point 2)) that I use a different centroiding for MS1 and MS2 data - since I did run into the same issues. In MS2 data (on our instrument) we have much less peaks and thus I had to change the smoothing settings (actually drop that). You can perform the centroiding for MS1 and MS2 data separately by specifying the MS level on which to apply the processing with msLevel. = 1L
or msLevel. = 2L
(also here, have a look at the centroiding.R script).
@jorainer - thanks. i didn't try removing smoothing and will give that a go. I guess i was just puzzled why i couldn't find any settings that worked without an error. The sparseness is certainly the most likely explanation.
Hello all,
Trying to use continuum mode LC-TOF data in XCMS. i have implemented smoothing and peak picking, and visual examination suggests the following parameters are pretty suitable for my data (LC, Waters Xevo G2 XS QTOF - ie. i see centroids that look plausible):
All goes well through XCMS centWave peak detection, correspondance, etc. no error messages, peak detection looks plausible at first glance, object called 'xs' created.
Problem 1: I am having trouble with fillPeaks, seems to be related to BioCParallel.
When i switch to SerialParam() instead, i get a new error:
This certainly could point to e memory issue - i am working on a computer with less than ideal memory availability. Windows does report more than 2.2 Gb of available memory, but i anticipate this is a memory issue and i am not sure there is a way to resolve it without moving to a bigger computer.
Problem 2: MS/MS data processing issues.
I am trying to use the MS/MS aggregation functionality, as this is DDA data:
This is odd to me, becauase I used the same halfWindowSize values in the feature finding data on MS1. Both are continuum data collected at the same MS resolving power. Moreover, if i try to reduce the halfWindowSize value, i run into new problems: