Skip to main content

Topic: Resolving Closely Related Structural Isomers (Read 132 times) previous topic - next topic

  • hhabra
  • [*]
Resolving Closely Related Structural Isomers
Hello,

Forgive me if a topic like this has been brought up before (though I looked and didn't find a satisfactory answer).

So I'm showing here a picture of one feature with median values mz = 232.1542 and RT = 171.41 s (2.8568 minutes).
It appears that this multipeak feature corresponds to a pair of closely related structural isomers: Butyrylcarnitine and Iso-Butyrylcarnitine. On a different preprocessing program (Agilent Profinder), they seemed to have been resolved with Iso-Butyrylcarnitine having retention time 2.85 minutes and Butyrylcarnitine having retention time 2.93 minutes.

So my question is are there possible ways for XCMS to resolve structural isomers such as this? What are my options if I'm really interested in identifying and quantifying such isomers?

I have included my code for consideration:

Code: [Select]
xset <- xcmsSet("stellaPOS2", 
  method = "centWave",
  peakwidth       = c(10, 40),
  ppm             = 30,
  noise           = 0,
  snthresh        = 6,
  mzdiff          = 0.01,
  prefilter       = c(3, 500),
  mzCenterFun     = "wMean",
  integrate       = 2,
  fitgauss        = FALSE,
  verbose.columns = FALSE)

xset <- retcor(xset,
  method   = "obiwarp",
  profstep = "0.5")

xset2 <- group(
  xset,
  method  = "density",
  bw      = 5,
  mzwid   = 0.033,
  minfrac = 0.5,
  minsamp = 1,
  max     = 50)

xset3 <- fillPeaks(xset2)

Thanks for any and all insight!





  • University of Michigan

  • Jan Stanstrup
  • [*][*][*][*][*]
  • Administrator
Re: Resolving Closely Related Structural Isomers
Reply #1
A quick suggestion:
I think mzdiff should be left at the negative default. Otherwise it does not allow overlapping peaks as far as I know.
Maybe peakwidth need to be lowered. Your peaks are less than 10 sec.

If that doesn't work you need to understand if the grouping is merging them. Look in peaks() if there are multiple peaks per sample.
  • Last Edit: July 08, 2018, 10:00:29 AM by Jan Stanstrup
  • University of Copenhagen, Denmark
Blog: stanstrup.github.io