Skip to main content
Topic: LC/MS blank removal from samples (Read 4654 times) previous topic - next topic

LC/MS blank removal from samples

Hi all,

I'm doing an untargeted metabolomics project to find new bioactive Streptomyces compounds.  I'd like to subtract the media blanks from my samples, leaving the in theory only the produced metabolites showing in my chromatograms and spectra.  The vendor software (Waters) does have a function for this, but it's manual so obviously takes quite a bit of time.

Can anyone recommend any free software that can do this automatically?  I've tried XCMS Online and MZmine but haven't been able to find an appropriate setting (it's entirely possible I've missed it though, so please say if I have!).  I know it's possible to generate peak lists and just subtract the blank, but that doesn't give chromatograms/spectra.

Thanks,

Josh

 

Re: LC/MS blank removal from samples

Reply #1
There is no easy way to do this since your blank is probably also your solvent. You can overlap pure blanks on top of your chromatogram to see which peaks are supposedly coming from your blank. Agilent software or any opensource software can do that for you. I know you didn't want this for the processing steps but I would do it the following way:
If you mean to remove the intensity signals for the blanks in the diff report, then the best way to approach this is that, if you have run pure blank injections in between your runs, place them in one separate folder. Place the sample runs in another folder. Supposedly you would have two cohorts, the blanks being one of it. When you generate a diff report, look at the fold change values. Sort the spreadsheet on fold change. If the fold change for your samples is greater for a certain feature than keep those values. If the fold change for the blanks is greater than the fold change for your samples then with caution remove those values and impute then with zero. Before you do that make sure to look at the entire row of your samples to see if the intensity values are indeed lower than those shown for the blanks. Check the frequency of appearance for each cohort as a double check.