[Psidev-ms-dev] dataXML and spectral library support

[Karl C. <cl...@br...>]

Hi folks,

>- Do we support properly the spectrum "library" use case?

            dataXML is supposed to be for MS instrument raw data, not
interpreted data, i.e. with assignments. That is what analysisXML is
intended for.

I do not espouse the view that dataXML should contain peptide interpretation
for individual MS/MS scans (db, de-novo, or otherwise). dataXML is sensibly
intended for spectral information, raw or processed. Association of the
interpretation information and the individual spectrum in the dataXML file
could come from analysisXML or elsewhere.

 

The extent to which dataXML could usefully and easily support spectral
libraries is to facilitate that a dataXML file could contain a collection of
MS/MS spectra that may have been derived from one or more LC/MS/MS runs. I
personally envision that all spectra are at least derived from the same type
of instrument

To effectively meet that objective the schema simply needs to support:

1.       A spectrum name as a string (I expect this would not contain a full
path but instead be more like a .dta or .pkl name:
runame.firstScanNum.LastScanNum.charge.pkl)

2.       Spectral parameters on a per spectrum basis: i.e. resolution,
dissociation mode, collision energy

3.       Index attribute expansion where there can be additional attributes
other than scan number in the index element: i.e. spectrum Name, precursor
m/z, precursor charge. These would be used for sorting and rapid access to
selectively retrieve individual scans. Personally I favor the index as a
separate file rather than appended to the end as in mzXML.

 

--Karl

---------------------------------------------------

Karl R. Clauser

Research Scientist III

Broad Institute of MIT and Harvard

7 Cambridge Center

Cambridge, MA 02142

Ph. 617-324-9719

E-mail: cl...@br...