Untargeted IMS Tentative Identification Lipidomics

Following pre-processing, tentative identification is performed using an in-house developed annotation software - annotine.

Generate an average mass spectrum of the dataset (in profile mode).

Scale the mass spectrum between 0 and 1, and peak-pick its profile to retrieve a list of m/z features (commonly 100s to 1000s).

Filter the peak list to retain only peaks that have a relative intensity above 0.001 (sensitive mode) or 0.01 (standard mode). Peaks whose intensity value falls below this threshold are removed.

( optional ) De-isotope the peak list to remove M+1, M+2, … and other potential isotopes from further consideration.

Generate an internal database on the basis of a user-supplied list of molecular species databases and a set of user-supplied expected adduct types:

Databases:

a. coreMetabolome, LMSD, SHexCer, HMDB5

b. (optional) a local LC-MS database

Adducts:

a. Positive mode:  [M+H]+, [M+Na]+, [M+K]+

b. Negative mode: [M-H]-, [M-CH3]-

Perform tentative identification by comparing the peak list with the built database of species and adduct combinations. If a peak is within a ±5 ppm window of an annotation in the database, that annotation is assigned to that peak. This process is repeated until each peak has been compared to the database.

Evaluate each tentative annotation using metrics.

( optional ) To reduce the number of unlikely annotations, calculate a false discovery rate (FDR) for every tentative identification. Annotations can be filtered based on the FDR (or any other) score.

( optional ) If LC-MS results are available (see LC-MS/MS lipidomics protocol below), associate these directly with the annotine results. This allows immediate highlighting of which tentative identifications have also been observed by LC-MS/MS, increasing the confidence of the identification.

Bulk Untargeted LC-MS/MS Lipidomics