Title:Successes and Pitfalls in Automated Dereplication Strategy Using Mass Spectrometry Data: a CASMI Experience
VOLUME: 5 ISSUE: 1
Author(s):Samuel Bertrand, Catherine Roullier and Yann Guitton
Affiliation:Groupe Mer, Molécules, Santé-EA 2160, UFR des Sciences Pharmaceutiques et Biologiques, Université de Nantes
Keywords:Annotation, automated method, CASMI contest, dereplication, mass spectrometry, natural products.
Abstract:Background: Automated annotation of MS spectra remains highly challenging.
Therefore, the Critical Assessment of Small Molecule Identification
(CASMI) Contest (http://casmi-contest.org/) represents a unique opportunity to
blindly evaluate an annotation workflow.
Objective: The 2014 CASMI Contest consisted of 42 MS and MS2 spectra (from
“molecules detected in mammalian biofluids and tissues”, “natural products of
plant, fungal or bacterial origin”, “synthetic or semi-synthetic molecules”), which
had to be identified (molecular formulae and structures).
Methods: An R script based on the Rdisop and RMassBank packages was devised
for the automated annotation of the provided spectra. Searching within various databases yielded the
assignment of structures. Further discrimination between annotations was achieved using phylogenetic
information and in silico fragmentation by competitive fragmentation modelling using CFM-ID.
Results: Successes and failures of the proposed script were investigated after release of the CASMI
Contest solutions.
Conclusion: The CASMI Contest allowed identifying key points for the successful annotation of MS
spectra. As an example, this study pointed out the importance of taking a large number of adducts into
consideration (such as [M+H]+, [M+H-H2O]+, [M+Na]+, [M+K]+ and [M]+ in positive ionization) to
avoid failed annotations.
