Skip to main content
Topic: When to normalise urine using metaXCMS? (Read 3758 times) previous topic - next topic

When to normalise urine using metaXCMS?


I am using metaXCMS for human urine samples where there are 3 subjects and 2 conditions (fed and unfed) with 3 replicate injections and I need to normalise appropriately to account for sample dilution between subjects. It would be good to normalise across all of the samples together to the median fold change and this would usually be following peakpicking. However if I am treating each subject individually I cannot easily do this, perhaps there is a way as I don't want any false negatives/positives. I don't know how I could do this when I am generating a diffreport seperately for each individual? I read through the nature protocols paper and could not find any suggestions on fitting in a normalisation step. I have already used metaXCMS and it is working very well and some interesting features have arisen based on monoisotopic mass but I want to ensure my interpretation is accurate.

I am considering in future to normalise to osmolality and correcting each sample prior to data acquistion in future however for now does anyone have any suggestions of how I can normalise these urine samples so that I don't misintepret the second order analyses with metaXCMS and fit it into the work scheme?

Also does anyone have any suggestions on how I can search my own database or the soon to be released FooDB when generating the XCMS diff report instead of/ in addition to Metlin?

Thanks very much.  :)

Re: When to normalise urine using metaXCMS?

Reply #1
Dear all,

Another thing I wish to do is to reimport the detailed common features diff report of metaXCMS back into R that I can annotate it using CAMERA. Is this possible? One way I have tried to annotate the mass spectral features is by using the annotateDiffreport wrapper function prior to generation of the diff report for metaXCMS as part of the workflow. This also usefully calculates a potential monoisotopic mass for each mass spectral feature rather than a m/z range. It would of course be more useful to be able to annotate the metaXCMS common features and to obtain estimated monoisotopic masses as I want to assess the mass accuracy aswell.

It is possible to annotate adducts and isotopes utilising the Agilent Qualitative software and to perform a user-generated monoisotopic mass database search however I would like to be able to perform the same analyses using XCMS/metaXCMS if possible for the purposes of comparison. I prefer to conduct my multivariate analyses in SIMCA-P+ rather than Mass Profiler Pro (partly as I am more fimiliar with the former and also because I feel MPP is not very transparent and limited in it's versatility). I don't know if anyone else has had the same thoughts or encountered similar difficulties.


Re: When to normalise urine using metaXCMS?

Reply #2
When I process my XCMS data for metaXCMS on the R platform I receive this message during feature detection "Peak Data Insertion Problems-Please Try Lowering ppm parameter".