Compound identification in metabolomics

Jessica Cooperstone

How do we go from features to metabolites?

  • Individually (one feature at a time)
    • Manual, very high confidence in ID
  • In bulk (many features at a time)
    • Computational, generally lower confidence in ID

Starting with a m/z

  • A feature of interest differentially present in commercial vs. wild tomatoes. We want to know what this is.
  • UHPLC-QTOF-MS, ESI+, reversed phase C18, Methanol extract
  • A useful adduct calculator

Relative intensity of 1092.5590 in tomatoes m/z in tomatoes

Search in MS databases - HMDB

Search > LC-MS Search

Screenshot of an HMDB MS1 search
  • Enter your mass and mode
  • Indicate adduct type (can put unknown if you don’t know)
  • Select a mass error (5 ppm is good for a QTOF)
  • Search

Evaluate search results

Think about:

  • Which structures are plausible?
  • Make sense in your biological system
  • Make sense in your extraction
  • Make sense based on retention time
  • Which adducts are more likely?
  • The top result is not necessarily your ID!

Access or collection MS/MS fragmentation data

MS/MS spectra of 1092.5590 at 65eV in +ESI

Download structure

http://www.hmdb.ca/spectra/ms_ms/59530

Download an .sdf file of your structure

Import into ChemSketch

Import your .sdf file into ChemSketch

Set ChemSketch to provide you structural information

  • Select which parameters you want printed: Tools > Calculate > Select properties to calculate > Select “Molecular Formula”, “Monoisotopic Mass”, “[M+H]+”
  • Have ChemSketch calculate those parameters: Tools > Calculate > Selected properties > Copy to editor.

Screenshot of acetoxytomatine in ChemSketch with parameters calculated

Breaking bonds to rationalize fragments

  • Use the eraser (delete) to break bonds (learn more here re: using ChemSketch)
  • Select a bond and it will be deleted
  • The details will stay with the larger fragment. Highlight the smaller piece to recalculate parameters.
  • Account for any rearrangement or additions/subtractions

Click on the delete eraser

Breaking bonds to rationalize fragments

Cleaving at the acetoxy group

Breaking bonds to rationalize fragments

Cleaving at the acetoxy group

How can we improve the confidence of our ID?

  • Reference against publicly available MS/MS spectra
  • Purchase authentic standard and compare
  • Synthesize standard, confirm by NMR, and hope you’re right 🥹
  • Compare to MS/MS spectra of similar compounds
  • Compare to a sample that you know has your compound of interest

MS/MS online databases

  • Experimental spectra
  • Predicted spectra

MS/MS predicted spectra of acetoxytomatine

http://www.hmdb.ca/spectra/ms_ms/59530

Predicted spectra for (23R)-acetoxytomatine

Predicted vs. actual

MS/MS spectra of 1092.5590 at 65eV in +ESI

Acetoxytomatine predicted spectra from CFM-ID

MS/MS spectral databases

LC-MS metabolite ID workflow

From Fitzgerald et al., ACS Omega 2017

Do you need to have an ID to quantify?

© Jessica Cooperstone, 2024