Tutorials and instructions
MS Prediction Tutorial
IntroductionThe Mass spectrometry peak prediction tools enable a user to search computationally generated databases of structures based on commonly occurring core structures and chains with a list of m/z values (ions). Examples of such databases are those derived from acyl/alkyl chain permutations for glycerolipids and glycerophospholipids. Alternatively, one may search sets of known lipids in a structure database with a list of m/z values (e.g. fatty acyls in the LIPID MAPS database). The user is presented with a number of option for refining the search, including mass tolerance window, headgroup (in the case of glycerophospholipids or sphingolipids), ion type and intensity threshold. It is also possible to limit the search to certain chain length and type. Finally, there are separate tools for the GL and GP categories in order to report predicted lipids as discrete structures (e.g PE(16:0/18:1(9Z))) or as "bulk" species (. PE(36:0), LPC(16:0)).
Example: Predict possible Glycerophospholipid structures
Using the 'Glycerophospholipid MS analysis' tool, the user pastes a tab or space-delimited list of precursor ion m/z values and corresponding peak intensities in the area on the right. An appropriate Intensity Threshold (default is 5000) is then chosen. All ions in the peak list with intensities lower than this threshold will be ignored. Depending on the mass accuracy characteristics of the MS instrument, a mass tolerance window is selected (default is +/- 0.2 m/z). Depending on whether the spectra were acquired in positive or negative ion mode, an appropriate ion-type(s) is selected from the pull-down "Ion" menu (default is [M+H]+).
The results summary (below) is grouped according to input m/z value (496,524,625, etc.) where the user-defined criteria are met. Columns C and DB refer to number of carbons and double-bonds in the side-chains only. Diagnostic product-ion calculations (e.g. loss of the sn1 acyl chain from the precursor ion (M-sn1)) are listed where appropriate. Results within each input ion sub-table are sorted by "Delta" value by default. Delta is defined as the absolute difference between the calculated and actual (input) m/z values for each predicted ion. There are several other sort options (abbreviation, number of chain carbons, formula, etc.)
Clicking on a hyperlink in the 'Abbreviation' column draws the structure of that predicted lipid as a MarvinView applet, using the LIPID MAPS structure drawing software. Alternatively, clicking on a hyperlink in the 'Formula' column displays tabular and graphical data for the isotopic distribution pattern for the selected ion (see below).
Example: Predict possible Glycerophospholipid structures - multiple ion types
Instead of using the default ([M+H]+) ion type, the user can expand the search by selecting multiple ions (e.g. [M+H]+,[M+Na]+,[M+K]+,[M+NH4]+). In this case, additional structures satisfy the search criteria, for example the [M+NH4]+ adduct below.
Example: Predict possible Glycerophospholipid structures - limiting search to certain chains
The user may specify a list of chains to search (for example all even-carbon chains) if information is known about the chain characteristics of the sample of interest. Once a list of chains is selected, the user must then make a selection from the "Limit search to certain chains" pulldown menu: either "Limit search to molecules containing ONLY the checked chains" or "Limit search to molecules containing at least 1 of the checked chains" (see example below).
Example: Predict possible Glycerophospholipid structures - specify a headgroup
A glycerophospholipid headgroup (e.g. PE) may be specified to limit the search to predicted structures for that headgroup (below).
Example: Predict possible Glycerophospholipid "bulk" structures
One may also use the "bulk" glycerophospholipid search interface to perform glycerophospholipid MS analysis, the key difference being that predicted matches are reported as isobaric "bulk" species where numbers of chain carbons and double-bonds are summed (e.g PE(36:0), LPC(16:0)).
This generates a shorter list of predicted species as opposed to the "discrete" search interface described above, because the individual isobaric molecules are not specified. There is no structure-drawing hyperlink in the 'Abbreviation' column in this case because exact structures are not being specified.