fillpeaks creates high background in GC-MS

January 19, 2017, 02:59:16 AM

Hi,
I'm currently analyzing data that I recorded on an Agilent GC-MS (single quad) with XCMS and CAMERA. I use TMS derivatization and a temperature gradient that goes up to 300 C. Because of the temperature I see quite a bit of column bleeding towards the end of the chromatograms, most prominently the masses 207 and 281. Looking at the raw data the peaks that are detected in those two mass traces are quite high above zero, but only slightly above the local baseline. So as long as I do the analysis without peakfill and intb as intensity value, this is not much of an issue.
However if I use peakfill I have the impression that it only fills into. So it adds all those really high intensity signals with the effect that there is a high background signal even in blanks. I tried to get rid of those background signals using the correlation filtering function from CAMERA afterwards, but that did not help much.
I think that ideally peakfill would also do a baseline detection and fill the intb column. Is there a way to do that already?
Do you have any other suggestions how to tackle this issue?
Any hint is more than welcome.
Thanks,
Joerg

Re: fillpeaks creates high background in GC-MS

Reply #1 – January 19, 2017, 09:01:58 AM

Thanks for the question. I think adding other integration methods into fillPeaks would be not entirely trivial,
but might be possible. Patches welcome :-)
Another potential solution would be to pass the missing peak as ROI to centWave with much lower S/N threshold,
and merge these peaks. No guarantee that the second run picks them up, though.
Yours, Steffen