removing neutral loss fragments from spectra

[Gautam Saxena]

I know that some pre-processing tools exist to deisotope spectra and even do charge deconvolution; however, I don't know of any/many examples in which a similar process is done to remove the ms2 signals that "appear" to correspond to neutral losses of H20, NH3, or CO. Is it because, in practice, removal of any possible ms2 fragment signals that may be a H20, NH3, or CO is more damaging than helpful, as some of those signals may in fact be real b or y ions? In particular, would it be reasonable to conclude that in order for any such neutral-loss-removal algorithm to be "beneficial", the ratio of "true positive" to "false-positive" neutral-loss ions must be very very high, eg 100 to 1?

[dtabb1973]

Hi, Gautam.

Many of the identification tools from the Center for Computational Mass Spectrometry at UCSD interpret "PRM graphs" from tandem mass spectra.  The process of building a PRM combines information from multiple isotopes of fragment ions, neutral losses from fragment ions, and complementary pairs of fragment ions.  I think Sam Payne may have visualized this the best in his 2008 paper on phospho-InsPecT: 10.1021/pr800129m.

Thanks,
Dave

[David Bouyssié]

Maybe my answer is very naive but I think the reason of absence of such pre-processing tools is that it is usually managed by the search engine itself, since it is more PSM related (neutral losses have to be inferred from a given peptide sequence).

However one could imagine to easily implement this filter for spectral libraries for instance.

Do you have a specific use case?

David #2

[Gautam Saxena]

Thanks David, I'm going through the paper now....