MS-DIAL / Problem with annotation of lipidomics data MSDIAL v. 4.9.221218.Last post by capkavl -
I came across a problem when annotating lipidomics data from Thermo instrument (.raw) using the latest MSDIAL version 4.9.221218. On average, I see ~70% fewer reference matched annotation compared to version 4.90 when using the same parameters (same parameter file) on the same raw data. Both versions detect roughly the same total number of features but the number of ref. matched is ~70% less in v. 4.9.221218
Here is an example using commercially purchased brain extract:
Total number of features: 8606
ref. matched: 852
Total number of features: 9688
ref. matched: 202
Interestingly, I only observed this using Thermo data. Processing Sciex data worked OK.
Is there something that needs to be changed as far as processing parameters go for the latest version vs version 4.90?
MS-DIAL / Re: Problems with Lipidomics mode: No peaks/zero ref matchesLast post by capkavl -
Also, I do not see this problem when processing Sciex data.
I hope this helps.
MS-DIAL / Question about Bruker QTOF raw data in MS-DIALLast post by silasmellor -
I am using MS-DIAL to process data coming from an older qTOF compact we have in our institution. We have encountered some issues with these data when converting to mzXML/mzML using msconvert, wherein the conversion incorrectly uses the uncalibrated precursor m/z in MS2 scans, which meant a lot of MS2 spectra ended up being unconnected to corresponding MS1 peaks.
I wonder, since MS-DIAL seems to read bruker .d natively, can we assume that this issue is handled correctly when processing the datafiles (assuming they are calibrated before processing)?
MS-DIAL / Re: Problems with Lipidomics mode: No peaks/zero ref matchesLast post by Tomiwa Oyedokun -
To be more clear on my previous question. Actually, it's not that MS DIAL did not detect any hits. It did detect over 28,000 hits but after I compared the hits with the reference library, it did not display any peaks with the reference spectra. The hits were reduced to zero. Please see the two screenshots showing hits generated, and zero hits after I compared them to the reference library.
And as for the file conversion used, we used ABF converter: https://www.reifycs.com/AbfConverter/
Thank you so much once again.
MS-DIAL / Re: Fail to detect shimadzu qgd fileLast post by biswapriya -
Thanks for educating!!!! : )
MS-DIAL / Re: Help with Library Selection/ResourcesLast post by biswapriya -
HMDB is NOT a spectral library but reflects other spectral libraries. So if you want to focus on human metabolites, better to do as much annotations possible using multiple spectral libraries such as the folllowing and then sort the annotations using HMDB IDs.
1. MoNA .msp "ALL spectra":https://mona.fiehnlab.ucdavis.edu/downloads
2. GNPS .msp all spectra:https://gnps-external.ucsd.edu/gnpslibrary
3. Combine the rest from RIKEN MSDIAL server
You can combine all .msp's accessible as .text files, following instructions as here :https://www.protocols.io/view/steps-for-building-an-open-source-ei-ms-mass-spect-eq2ly33rqgx9/v1 when it was done for GC-EI-MS spectral libraries, but the idea is simple as once Hiroshi taught me this way!
Note: Please "ignore" using all 'w/o MS2" at all- and you are very right there.
Also when searching against the spectral libraries you got to ignore the RT values. Only MS/MS matches at 80-90% cosine similarity fora any further analysis or interpretation !
Hope this helps.
MS-DIAL / Re: Which libraries were the 'MSMS_Public_EXP_VS17.msp' made up ofLast post by biswapriya -
(1) If (2) is correct and is VERY likely- but wait till Hiroshi confirms; then (1) take precedent / priority annotation over all other libraries. Make sure your data is from negative ionization mode and if that's true given limited amount of -ve mode MS/MS spectra, this limited annotations in Public_Exp_Neg_VS17 is understandable!
MS-DIAL / Re: Error occurred in JointAlignerLast post by biswapriya -
[A] Very likely that one of the files is corrupt or does not have good data. Or, one of the file is run on different ionization modes.
Just leave out a few in the set and see which is the culprit file!
At least these 2 have affected me in the past, so worth trying.....
MS-DIAL / Re: Problems with Lipidomics mode: No peaks/zero ref matchesLast post by biswapriya -
Please check the following things and very likely:
1. Ensure that your LC-MS instrument acquired the data in MS/MS
2. While converting the data you are converting in a way that MS2 level data is retained and not lost.
3. More info on the instrument used and file converted used and sharing a downloadable link to the file help.
4. Another culprit could be too stringent filters set for RT matches (put it to 100 mins) or MS1 resolution could be impacting analysis.