mzML profile mode?

[rmflight]

From what I can tell, this tool only outputs centroided data, even when the data is acquired in profile mode.

It would be nice to have the option to output either centroided or profile data when samples are acquired in profile mode. Very useful for some metabolomics applications.

[nielshulstaert]

Hi,

Thanks for your input. I decided to output the centroided data when available after a mail conversation with a very helpful expert at thermo. It's important to note that not all profile scans have centroided data available and we don't do any post-processing besides the native thermo library. That's also why we have the "MGF with profile data excluded" option.

I agree that the user should be able to choose, is it ok if I add a flag and keep the default behaviour?

[rmflight]

I think it's great if the default is to output the Thermo peak centroided data. However, that should be documented, and having a flag to get the profile data instead would be fine.

[ypriverol]

@rmflight can you test the new version https://github.com/compomics/ThermoRawFileParser/releases/tag/v1.1.4. We have addressed the following issue.

[timosachsenberg]

I am pretty sure I did not change anything in our call to the TRFP but the spectrum does not get picked anymore.
Using 1.1.5 with outfile option.

[rmflight]

For v112 vs v114, I get file sizes of 1749854 and 1749796, respectively when I specify -f 1, and don't do anything for peak picking.

When I add the -p option in v 1.1.4, my file size jumps to 23974569, or 13X the size of the other two. So from that simple perspective, the -p is definitely reporting profile mode spectra. I will read them into R using xcms hopefully on Friday (currently on vacation), but wanted to download and get this checked ASAP.

[timosachsenberg]

@rmflight does this work for you with the v115 version and using the new outfile= option?
We currently have the strange situation that with peak picking enabled (noPeakPicking flag is not set) we get profile data:
OpenMS/OpenMS#4095
Can anybody confirm that?

[jpfeuffer]

At least they are declared as profile spectra according to the CV term that is created here (without considering the new flag somehow):
https://github.com/compomics/ThermoRawFileParser/blob/a4038c3499435b3aec44b600ef7dc5f0a6e05550/Writer/MzMlSpectrumWriter.cs#L951

[rmflight]

Won't be able to check until tomorrow, unfortunately. Currently on vacation, and driving all day today. Should be able to check tomorrow.

…

On Thu, May 16, 2019, 8:14 AM Timo Sachsenberg ***@***.***> wrote: @rmflight <https://github.com/rmflight> does this work for you with the v115 version and using the new outfile= option? We currently have the strange situation that with peak picking enabled (noPeakPicking flag is not set) we get profile data: OpenMS/OpenMS#4095 <OpenMS/OpenMS#4095> Can anybody confirm that? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#34>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AALQR6V6X25P25YZBUZMJXLPVVFZRANCNFSM4HJRIVDQ> .

[rmflight]

OK, so both v114 and v115 have profile in the output type, there is no centroid to be found, even when thermo peak picking is asked for. This is in there even when I specify the output file directly using -b. So even though peak picking is being performed, the output does not reflect that. I'm guessing there is a small bug, as pointed out by @jpfeuffer

[nielshulstaert]

Hi if you run the parser without the -p flag, the native thermo library centroiding is applied. In the mzML output format, both centroid en profile scans are reported.

Using the -p flag disables the thermo centroiding and results in bigger files. In the mzML output format, both centroid en profile scans are reported.

So using the -p flag results in the same number of spectra as not using it. It's like not selecting the vendor peak picking in msconvert.

Could you provide a link with the RAW file for which you don't get centroided scans?

[timosachsenberg]

Same file. As Julianus linked in this issue We think just the metadata is not properly updated but centroiding is performed.

[rmflight]

Here is a file I've been using, in case it helps to have another one: https://drive.google.com/open?id=1AWSEQV7TtnFtPHwzld5U1KWtz1aM6kUm