Skip to main content
Topic: Error in grouping Br and Cl isotopes (Read 760 times) previous topic - next topic

Error in grouping Br and Cl isotopes

Hi

In looking at some features in a data set that I had identified in other software as containing Cl and Br, I found that MS-DIAL sometimes identified the isotopic peaks as separate features, even with "Consider Cl and Br elements" checked. Some examples are below in screenshots. There is a feature with two Br where the M+4 is not included in the monoisotopic alignment spot, and is a separate alignment spot. There is a feature with one Cl where the M+2 is not included in the monoisotopic alignment spot, and is a separate alignment spot. Finally, there is an example of the Cl feature in MS-FINDER, where because of this the M+2 is not found and the molecluar formula cannot be fit to the correct isotope pattern.

Does anybody know any tricks for improving the grouping of Cl and Br isotopes with the monoisotopic alignment spot?

Thanks,
Lisa

Re: Error in grouping Br and Cl isotopes

Reply #1
Hi Lisa,

from the last week, I checked my source code, and modified them to deal with your issue.
I completely modified isotope trackers. Especially, the previous msdial did not correctly reflect the isotope estimation results from each file's result. I hope the following developmental version is fine for you. Please let me know if it's good for you or not.
https://briefcase.riken.jp/public/f9rMQA4nJYhAN3I

Thanks,

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #2
Hi Hiroshi,

Thank you for getting back to me about this. :)

I ran a subset of the files with the new version of MS-DIAL (4.27) as a quick test. Unfortunately, it still picks the M+2, M+4, etc. isotope peaks from halogens as separate alignment spots, as shown in the screenshots attached.

 I am also quite confused where the isotope pattern exported to MS-DIAL for the Br2-containing feature came from. Looking manually at the heights of these spots, the ratio between M and M+2 should be approximately 1:2 as expected for the Br2 isotope pattern. For some reason, it is approximately 1:3 in the exported isotopic abundances.

In your reply you mentioned the isotope trackers. This is not an isotope labelling experiment and everything except the internal standards should be at natural isotopic abundances, which for Cl and Br includes 35Cl, 37Cl, 79Br, and 81Br.

Kind regards,
Lisa


Re: Error in grouping Br and Cl isotopes

Reply #3
Hi Lisa

sorry, OK, could you please send me the files that you now process?
Because I personally do not have an experience for profiling environmental chemicals, I would like to check my source codes by checking the real raw data.

Thanks,

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #4
Hi Lisa,

thanks for sending several files to me. I modified the source codes again, and I would like you to evaluate the modified version from the below link.
https://briefcase.riken.jp/public/u90sQABnkEhAJOU

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #5
Hi Hiroshi

Thank you for the new version.

I tried this version and still found the same problems with the isotope groupings. I am particularly confused by the isotopic MS1 mass spectra that are exported to MS-FINDER. They seem not to reflect the actual data in MS1 in the samples.

An example is attached. From the peak spot from the file with the largest peak for this feature and zooming in in MS1, the M+2 is 310.8383 and M is 308.8403. The difference is 1.9980, which is consistent with the difference between 79Br and 81Br. Then, when looking at the M alignment spot, the M+2 is 310.8471, which is not the actual mass that was observed and the difference is 2.0068 consistent with 2x13C. In addition, the relative intensity of the M+2 is higher for the alignment spot than the peak spot from a sample. It is not clear to me where this isotopic MS1 spectrum is coming from.

Thanks,
Lisa

Re: Error in grouping Br and Cl isotopes

Reply #6
Hi Lisa,

at least, for the peaks that you are now showing were fixed on my side. Could you please check the parameters again? For example, did you tick "halogen option" of data correction tab?
Or, did I have a mistake when uploading new version...?

Thanks,

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #7
Hi Hiroshi,

Thanks for your response. Yes, the "Consider Cl and Br elements" is checked in the Data collection tab. For some reason, this does not appear in the parameters text file that is exported.

Lisa

Re: Error in grouping Br and Cl isotopes

Reply #8
Hi Lisa,

I checked your parameter file. At least, at this moment, can you try to use the following parameters?
#Centroid parameters
MS1 tolerance   0.01
MS2 tolerance   0.025

#Peak spotting parameters
Mass slice width   0.1

Other parameters are same as you use.
Thanks, maybe, a critical thing is the tolerance for centroid MS1. As you said, the current program tries to recognize isotopes by the mass shift based on 13C-12C difference. Therefore, the Br difference can be out of the range of the isotope scope.

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #9
Hi Hiroshi,

Thanks you for this suggestion.

This change does result in the isotopes grouping together. However, the isotope pattern that exports to MS-FINDER is still not accurate in terms of the relative intensities, even to the peak heights in the sample that is used to get the representative spectra.

 Also, the mass assigned to the isotopic peaks remains the theoretical 13C peak mass and not the one that is actually observed. This is problematic for high resolution data where the 13C and Cl or Br isotope peaks are resolved and that difference can tell use which isotope it is.  In the example below, the MS1 isotopes export as: 308.84042:1313347 309.84377:0 310.84713:3151799. The peak height in the SRM file used for the spectra is 2726809. It's unclear where 3151799 comes from. Also, the actual M+2 mass is 310.83826.

It would be really good to have an export file of the MS1 isotopic spectrum that takes the heights and accurate masses from all the isotopic peaks assigned in a sample with a good signal, including beyond M+2. This would give a possibility to interpret the full isotope pattern and use it in formula assignment.

Thanks,
Lisa



Re: Error in grouping Br and Cl isotopes

Reply #10
Hi Lisa,

I forgot to answer about this. Yes, you are right. The current alignment result does not have the proper isotopic patterns.
At this moment, if you need the real isotopic information for msfinder, you have to export the peak information from each analysis file result (in peak viewer). In next year, a completely new architecture of msdial will be launched in which you can export the isotopic patterns from the alignment result data.

However, I will try to implement this in the current/old msdial code for you. Please give me one more week.
Thanks,

Hiroshi


Re: Error in grouping Br and Cl isotopes

Reply #11
Hi Hiroshi

Thank you for helping with this. I think it will be very valuable whenever the isotope pattern export is improved.

Lisa

Re: Error in grouping Br and Cl isotopes

Reply #12
Hi Hiroshi,

In looking at our data set after adjusting the centroiding parameter, we found a concerning issue with one of a few features that we'd tentatively identified previously. With the centroiding parameter at 0.005 Da or lower, we identified dichlorophen, which was a very large feature in a number of samples. The three chlorine isotopes show up as separate alignment spots, as shown in the attached image of the alignment spots. The M+0 peak is at m/z 266.9979 (Figure 1).

Using the centroiding parameter of 0.01 Da, the dichlorophen alignment spot disappears from the final alignment and the nearest spot is at m/z  266.9205 and appears to be a satellite peak (a Fourier transform artifact, Figure 2).

Looking at a sample to determine how it disappeared, shows that the M+0 has been classified as the M+2 of a peak 2 orders of magnitude smaller than the dichlorophen peak that happens to be close enough to the mass difference for a C13 isotope peak (Figure 3). Because of this issue, we are not going to be using data processed with a centroid parameter of 0.01.

Cheers,
Lisa

Re: Error in grouping Br and Cl isotopes

Reply #13
Hi Lisa,

long time no see. If you are still interested in using msdial program, I improved the program to solve this feedback.
https://briefcase.riken.jp/public/t9IIQABn4ohAQ_w
I appreciate you if you can evaluate this upcoming version which will be uploaded in public soon.
Thanks,

Hiroshi

Re: Error in grouping Br and Cl isotopes

Reply #14
Hi Hiroshi,

Based on a quick test, this did not seem to solve the issue, unfortunately. I am not actively working on this project as I was before.

Lisa