Menu
▾
▴
Re: [Psidev-pi-dev] AnalysisXML PMF Example
|
From: Jones, A. <And...@li...> - 2008-10-23 14:13:48
|
If we represent different search engines within the same document, we have separate SpectrumIdentificationLists. So the composite key would work if it is (SpectrumIdentificationList ID, spectrumID, SpectraData_ref) I think? Cheers Andy From: Martin Eisenacher [mailto:mar...@ru...] Sent: 23 October 2008 15:05 To: Jones, Andy Cc: psi...@li... Subject: AW: [Psidev-pi-dev] AnalysisXML PMF Example Thank you for the changes! One remark: we run into problems with the composite uniqueness of (spectrumID,SpectraData_ref), because we assemble proteins with peptides found by different search engines. Each search engine run is a SpectrumIdentification and has its own SpectrumIdentificationList; there the same spectrum is (usually) referenced (with the same or another peptide sequences as SpectrumIdentificationItems, but other scores). We should drop that composite unique key… Bye Martin Von: Jones, Andy [mailto:And...@li...] Gesendet: Thursday, October 16, 2008 6:26 PM An: psi...@li... Betreff: Re: [Psidev-pi-dev] AnalysisXML PMF Example New schema uploaded by SVN with changes agreed on the call: - Made cvParam minOccurs = 0 on SpectrumIdentificationItem (to handle PMF data) - Altered documentation for SpectrumIdentificationItem to clarify that this is a single peptide-spectrum identification - Altered documentation for SpectrumIdentificationResult – one SpectrumIdentificationResult for one spectrum - Moved spectrumID and SpectraData_ref as attributes of SpectrumIdentificationResult, and made these composite unique, removed SpectrumElement Andy From: Pierre-Alain Binz [mailto:pie...@is...] Sent: 16 October 2008 14:26 To: Jones, Andy Cc: psi...@li... Subject: Re: [Psidev-pi-dev] AnalysisXML PMF Example Actually, there might be a score associated with each peptide - peak couple, similar to "weights" that coulld be attributed to a specific fragment ion match in a MS/MS spectrum. But at the end, even if a software goes through this, and maybe scores the peptides in an intermediary process, it will at the end always link a spectrum with one or more molecule(s) (in the case of PMF the molecule(s) is(are) proteins). It can annotate the spectrum, either with potential peptides of fragments, respectively. Anyway, we have not encoded any scores to the interpretation of fragment ions in MS/MS spectra. Similarly, we wouldn't do it for PMF. Annotation is the sole point of discussion, which I think we want to postone for v2, right? Pierre-Alain Jones, Andy wrote: Ah, I was getting confused, I was thinking that there would be one SpectrumIdentificationItem for one Spectrum, in fact there is one for each peptide. The problem with the way it's encoded (probably the only way possible without a schema change) is that it's not obvious how to tie together all the results from a single spectrum-protein identification. Even querying for the same spectrumID will not resolve it, since there may be multiple ranked spectrum-protein identifications from the same spectrum. I suggest we look at this in the call, I think we might need an alteration to the schema to model PMF more intuitively, unless I've misunderstood something. Cheers Andy -----Original Message----- From: David Creasy [mailto:dc...@ma...] Sent: 16 October 2008 13:54 To: psi...@li... Subject: Re: [Psidev-pi-dev] AnalysisXML PMF Example Hi Andy, For PMF, with Mascot, (and all other search engines as far as I know), there is no score assigned to matching a single peak to an individual peptide. A score for one peak in isolation would have little meaning - you can only really score the match of the complete spectrum. David Jones, Andy wrote: Hi David, You may be right that the parameter should be optionally anyway, but in this particular example is there not some score assigned to the peptide-spectrum identification? In the equivalent MS/MS example, we have: <pf:cvParam accession="PI:00171" name="mascot:score" cvRef="PSI-PI" value="62.72" /> <pf:cvParam accession="PI:00172" name="mascot:expectation value" cvRef="PSI-PI" value="0.000863428353297988" /> <pf:cvParam accession="PI:99999" name="mascot_rank" cvRef="PSI-PI" value="1" /> Are the same sort of scores not valid for PMF as well...? Cheers Andy -----Original Message----- From: David Creasy [mailto:dc...@ma...] Sent: 16 October 2008 12:29 To: psi...@li... Subject: [Psidev-pi-dev] AnalysisXML PMF Example Hi All, There's an example PMF file here: http://code.google.com/p/psi- pi/source/browse/trunk/examples/PMF_example.axml The only small problem that I can see is that for <SpectrumIdentificationItem> the schema has: <xsd:element ref="psi-pi:PeptideEvidence" minOccurs="0" maxOccurs="unbounded"/> <xsd:element ref="pf:Param" maxOccurs="unbounded"/> and I can't think of any required CV here To find it, search for <pf:cvParam accession="PI:xxxxx" name="TODO: Required content, but nowt to say here here" I suggest that we have minOccurs=0 for pfParam here unless anyone disagrees? David -- David Creasy Matrix Science 64 Baker Street London W1U 7GB, UK Tel: +44 (0)20 7486 1050 Fax: +44 (0)20 7224 1344 dc...@ma... http://www.matrixscience.com Matrix Science Ltd. is registered in England and Wales Company number 3533898 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Psidev-pi-dev mailing list Psi...@li... https://lists.sourceforge.net/lists/listinfo/psidev-pi-dev ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Psidev-pi-dev mailing list Psi...@li... https://lists.sourceforge.net/lists/listinfo/psidev-pi-dev -- David Creasy Matrix Science 64 Baker Street London W1U 7GB, UK Tel: +44 (0)20 7486 1050 Fax: +44 (0)20 7224 1344 dc...@ma... http://www.matrixscience.com Matrix Science Ltd. is registered in England and Wales Company number 3533898 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Psidev-pi-dev mailing list Psi...@li... https://lists.sourceforge.net/lists/listinfo/psidev-pi-dev ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Psidev-pi-dev mailing list Psi...@li... https://lists.sourceforge.net/lists/listinfo/psidev-pi-dev |
