Skip to main content ↑ ↓
Topic: XC-MS data analysis: MS/MS matching –replicates- and methods (Read 5029 times) previous topic - next topic

XC-MS data analysis: MS/MS matching –replicates- and methods

In our laboratory we work on the non-targeted metabolic profiling of plant extracts.
Our objective is the identification of the metabolites present in plant extracts and compare their accumulation levels in ctrl vs treatment.
After loading the data in the XC-MS website:
1)   Which is the most suitable method to analyze LC-MS ESI triple quadrupole data?
2)   Saying that I used HPLC orbitrap method in pairwise job (ctrl vs treatment), is there a way to group the replicates. (In the heatmap etc..)
3)    The identification using MS/MS matching, the matrix model is limited to: human, mouse, fly, and yeast. However we are looking for plant secondary metabolites. Is there a way to change the matrix model?
Thanks in advance

Re: XC-MS data analysis: MS/MS matching –replicates- and met

Reply #1 –
Hello,

My 2cents:

Regarding the job conditions,it took me weeks to understand and optimise the parameters for XCMS pre-analysis (Orbitrap in my case), so I guess there is no instant answer.
Perhaps do some technical reading and look for a preset with lower resolution to fit the QqQ?

If I were you, I would utilise XCMS for the pre-analysis, and then prepare the diffreport for further work on other flexible online tools, such as metaboanalyst and Invex, which cater more readily to what you are asking for. no offense to XCMS guys - they are doing an amazing job!

cheers

Re: XC-MS data analysis: MS/MS matching –replicates- and met

Reply #2 –
Thank you naamaka for your answer, I`m going to use SIMCA for the statistical analysis, xc-ms was interesting to me because it allows automatic MS/MS matching for metabolites identification. I wanted to avoid studying the MS/MS data peak by peak. However, as you said apparently I should study more to optimize the initial parameters to get the best results. 
if anybody tried or have experience with the identification part please help with your advises.

Re: XC-MS data analysis: MS/MS matching –replicates- and met

Reply #3 –
Hi again,

Although XCMS provides easy and comfortable tool for initial screening,I reckon that there is no way around re-checking MS spectra and peak integration individually, especially when you are not on top of controlling the XCMS parameters . Also, the MS-MS fragmentation patterns differ between instruments and conditions, so I'm not sure how well it will match yours.

I'll give you an example for non MSn data: out of ~800 peaks integrated in serum in one job, when I manually checked them on XCMS and Qual browser (the instrument software), I got rid of ~200 which were noise peaks, contaminants, funny products of big peaks, and tails of big peaks (there is always a compromise because not all peaks have the same characteristics for integration). Now, this was still after strict filtering, hence at least 50% of samples in one group has the peak; above s/n of 6, threshold intensity 5000 etc. I reckon it would have been better if I did blank subtraction before I imported the files to XCMS, but didn't due to the unstable background in the blanks. Also, there is a lot of work after you are given a putative id: some peaks that XCMS identified for me (not MS/MS data though) were fragments of others, and for quite a few peaks there was no hit because not all the possible metabolites exist in metlin.

good luck!