Menu
▾
▴
Re: [Psidev-ms-dev] Retention time in PSI-MS ontology
|
From: Pierre-Alain B. <pie...@is...> - 2006-11-13 11:56:30
|
Thanks Randy for these precisions. This tells me that we really need a god sample of use cases. Eric has a number of high-level descriptions that we have prepared during one of the sessions in Washington (this was done by Jim, Sean, Ronan, myself and a couple of others I am sorry to have not in mind right now). For each of them, It would be nice to think at generating 1) a more formal list of these usecases in the documentation of the standard and 2) a number of typical instanciations. Maybe in Seattle these usecases can be listed again and "checked" if considered or not during these 2 days. Pierre-Alain Randy Julian wrote: >Hi everyone, > >Short version: > >This is confusing and we need to deal with it directly in the CV. Below is >a long discussion - the short version is that 1000038/1000039 are meant to >be relative acquisition times which in chromatography system would be >'retention time' but this is not true for non-chromatography systems, so we >need to clean up the nomenclature. > >Long version: > >There are two terms in the CV which are meant to represent relative >acquitisiton time from the start of overall acquisition process. The data >type is expected to be a floating point number - not a date/time stamp. > >As such, it is ideally suited (and intentionally designed) to hold a >retention time. However, since 'retention time' is not globally applicable >to all separation methods, and since not all flow-based methods or >sequential acquisition methods involve separation at all (flow-injection, >etc.), then calling this term 'retention time' was overly specific. > >For MALDI system where multiple spectra would be collected from a single >spot, the idea of recoring the time from the first spectrum acquired for a >spot seemed to match this defintion of acquisition time, and since mzData >1.05 files cannot handle multiple samples, then each spot on the plate would >get it's own mzData file whose admin description could include creation >dates and times. > >As for the precise meaning of a relative acquisition time (in minutes or >seconds), we are somewhat at the mercy of the base acquisition system. > >There are two time-frames which need be considered: analyzer time, and >flow-system time. > >A typical analyzer takes some finite time to perform a single acquisition. >Most systems combine multiple low-level acquisitions into a single reported >acquisition (like most ion traps and TOF systems). Some systems record the >time representing the start of an atomic operation which results in the >production of a single reported spectrum. Other systems record the >concluding time of the spectrum. > >It is not usually critical to determine if the time indicated is at the >start of the spectrum scan, or the end, since it is a very short time in >almost all systems and is consistently done within individual analyzers. > >For most systems used in protoemics, there flow-system time is a completely >different thing. In data-dependent acquisitions, there can be large >variations is the cycle-time of the instrument, so there will usually not be >a 'sampling frequency' in the usual sense - although you can configure >experiments so have s fixed sampling frequency. > >This means that the acquisition time marks the time from the start of the >acquisition at which the particular spectrum is recorded. If several >spectra are combined then you are working in flow-system time, and there are >several ways to describe what has been done. > >The very best way to handle this is to use the >'acqSpecification/acquisition' elements to record exactly which acquisitions >are combined - including their acquisition times. > >As for the acquitision time of the combined spectrum, it is then possible to >indicate either the start of the sequence, or, if you are processing >chromatographic peaks, you can record a peak parameter like the apex time, >or the centroid of the peak (which is what most chromatographers think of as >'retention time'). > >This suggests that we should have new CV terms to allow the specification of >'retention time' from the chromatographic point of view, and perhaps clean >up the nomenclature to specify that 1000038/1000039 are relative acquisition >times. We should also consider adding a true datetime stamp CV term which >could be used to store a timestamp for the sample - especially if we move to >allowing multiple samples in a single file (proposed for dataXML). > >Suggestions about additions/changes to the CV? > >Randy > >-----Original Message----- >From: psi...@li... >[mailto:psi...@li...] On Behalf Of David >Creasy >Sent: Thursday, November 09, 2006 11:23 AM >To: len...@eb... >Cc: psi...@li... >Subject: Re: [Psidev-ms-dev] Retention time in PSI-MS ontology > >Hi Lennart, >The intention was certainly that these are for the retention time, but I >can't see any documentation that confirms this. Certainly Mascot >Distiller, Mascot, Bruker CompassXport, the Sciex wiff converter, >Bioworks and the Insilicos viewer all assume that this is the case. > >Best regards, >David > > > >Lennart Martens wrote: > > >>Hi Kent, >> >> >>Thanks for pointing that out, but I had seen these myself and assumed >>that they were used to document the actual acquisition time during which >>the final spectrum was acquired (the final spectrum represented in the >>mzData binary arrays is thus a summation over all the individual >>detector recordings during that time slot). >> >>As such, I would expect this value to be either constant (I guess most >>mass specs do not allow dynamic setting of this value, but I could be >>wrong) and certainly small (say from 0.5 to 8 seconds). Having '900' in >>the seconds field, say, would therefore probably lead to confusion? >> >> >>Cheers, >> >>lnnrt. >> >>Kent Laursen wrote: >> >> >>>Hi Lennart, >>> >>>There is not currently a term labeled retention time in the MS CV. The >>>current terms for the purpose are: >>> >>>[Term] >>>id: PSI:1000038 >>>name: Time In Minutes >>>def: "Acquisition time in minutes." [PSI:GPS] >>>is_a: PSI:1000460 ! Unit >>>relationship: part_of PSI:1000459 ! Spectrum Instrument Description >>> >>>[Term] >>>id: PSI:1000039 >>>name: Time In Seconds >>>def: "Acquisition time in seconds." [PSI:GPS] >>>is_a: PSI:1000460 ! Unit >>>relationship: part_of PSI:1000459 ! Spectrum Instrument Description >>> >>>Regards, >>> >>>Kent >>> >>> >>> >>>>-----Original Message----- >>>>From: psi...@li... [mailto:psidev-ms-dev- >>>>bo...@li...] On Behalf Of Lennart Martens >>>>Sent: Thursday, November 09, 2006 9:27 AM >>>>To: psi...@li... >>>>Subject: [Psidev-ms-dev] Retention time in PSI-MS ontology >>>> >>>>Hi, >>>> >>>> >>>> >>>>I am processing some mzXML files into mzData for insertion into the >>>>PRIDE database, and I can't seem to find an ontology term for 'retention >>>>time' in the PSI-MS ontology. >>>> >>>>If there is one, and I just can't find it for some silly reason, please >>>>let me know. >>>> >>>>If there isn't any, could this be added? It seems like a sensible thing >>>>as it is a common enough piece of data. >>>> >>>> >>>>Thanks in advance, >>>> >>>> >>>>lnnrt. >>>> >>>> >>>> >>>> >------------------------------------------------------------------------- > > >>>>Using Tomcat but need to do more? Need to support web services, >>>> >>>> >security? > > >>>>Get stuff done quickly with pre-integrated technology to make your job >>>>easier >>>>Download IBM WebSphere Application Server v.1.0.1 based on Apache >>>> >>>> >Geronimo > > >>>>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 >>>>_______________________________________________ >>>>Psidev-ms-dev mailing list >>>>Psi...@li... >>>>https://lists.sourceforge.net/lists/listinfo/psidev-ms-dev >>>> >>>> > > > -- -- Dr. Pierre-Alain Binz Swiss Institute of Bioinformatics Proteome Informatics Group 1, Rue Michel Servet CH-1211 Geneve 4 Switzerland - - - - - - - - - - - - - - - - - Tel: +41-22-379 50 50 Fax: +41-22-379 58 58 Pie...@is... http://www.expasy.org/people/Pierre-Alain.Binz.html |
