Skip to main content
Topic: Spectral Similarities of in-house measured reference compounds quite low (Read 2830 times) previous topic - next topic

Spectral Similarities of in-house measured reference compounds quite low

Hi all,
We have performed MSMS matching in a measurement file for certain reference substances, from which the reference MSMS is taken. It would be expected that the spectral similarities would result in 1000, since identical spectra are compared with each other. However, we obtain results that differ significantly from this (see appendix: results for dpc/rdpc: 750 or 880). How can this be?

The detailed procedure was as follows:
1) Export of the respective MS2 spectra from the ABF files of single standard runs (as single txt-files)
2) MS2 spectral data from exported txt-files are merged in in-house spectral library (MSP file)
3.) In order to control, if MSP-file based annotation is working in MS-Dial, this updated MSP-file is regularly used to annotate the reference standards in the standard runs

Thank you very much for your help!
With kind regards
FalcoB

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #1
Yes, if you export spectra to a library and then use that to annotate the source file you exported from you should get a score of 1.00.  I see that you got that in your third screenshot. What version of MS-DIAL are you using? 

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #2
Hi drchrispook,

we are currently using the version 4.90 (since we started building up our library using this version several months ago).

Do you known if this problem may have been solved in the versions 4.9.221218 and 5.1.230517, respectively?

Thanks.
FalcoB

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #3
I don't use MS-DIAL v5 because I haven't found it to be stable. I'm using 4.9.221218.

In your first post you mention that you acquired the library spectra from single standard runs, extracted the spectra and compiled them into a single .msp. Can you clarify if your question is whether the same standard solution should give you a dot product of 1.00 every time you run it (it won't), or whether if you analyse a file from which the exact library spectra was extracted you should get a dot product of 1.00 (you should).

There's a few caveats here around what your acquisition settings are. For example, if you acquired multiple DDA scans for a peak, then the one that you've extracted for your library might not necessarily be the one MS-DIAL extracts and uses. They can be quite different. Especially if the peak was particularly large, or you extracted the first scan on the rising edge of the peak with quite a low threshold.

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #4
Hi,

Regarding your first paragraph:
It´s the case that "whether if you analyse a file from which the exact library spectra was extracted you should get a dot product of 1.00 (you should)".

Regarding your second paragraph:
In fact, we measure DDA. Four build-up of our in-house MSMS database we are currently using MS-Dial following the instructions described here: http://www.metabolomics-forum.com/index.php?topic=1627.msg4889#msg4889

We do it step-wise, i.e. processing approx. 10 standards in parallel
step 1: converting the raw data files (wiff->abf) of the standard runs by ABF converter
step 2: processing abf files by MS-Dial
step 3: extracting the appropriate MS2 spectrum from MS-Dial GUI (as a single MSP-file)

Note: Per standard only one single MS2-spectrum (presumably that one with the highest absolute intensity, but I am not sure) is displayed in the MS-Dial GUI when selecting the corresponding ion of the reference compound. That means, we have no option to choose between different MS2 spectra for the same peak. The raw data of this MS2 spectrum we export to single MSP-File.

step 4: We then copy the data from the single MSP-files in our database-MSP-file (we copy the original MS2-data without making any changes in the mz/intensity matrix).

step 5: We re-process the standard runs in MS-Dial, however, in this case using the updated database-MSP-file (including the before added reference compound data) for automated identification.

Usually, when you have a look at the MS/MS spectra alignment figure, both spectra look identical. Nevertheless, sometimes (regularly) we get score values (dpc, rdpc) < 1000 (the lowest ever observed value was 750 I think). And we do not know why.

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #5
Hi,

next to version 4.90, we still have tested version 4.92, but the problem remains the same.
So, it seems not to be a problem of the version.

However, we would be very happy if someone could help us.

Best regards
FalcoB

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #6
Hi, this is a really odd one. Do you have an example file and an .msp file you can share for me to test?

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #7
Hi,

I have send you a personal message.

Regards
FalcoB

Re: Spectral Similarities of in-house measured reference compounds quite low

Reply #8
Hi,
I send you another private message. Please give me a short feedback if you were able to download all the files.

Thanks!

Regards
FalcoB