Re: [Psidev-ms-dev] Smaller files

[Matthew C. <mat...@va...>]

 From Josh Tasman, unable to post directly:

Hi Rune,

As you've seen in the source code, only the most basic implementation is 
currently in the SPC converters; and as Matt points out it is in the 
mzML spec so could be implemented pretty easily.

Thanks for your good suggestion, especially integer counts for Waters 
(although my experience is that if you acquire in centroid you get real 
values.)

Josh


Matthew Chambers wrote:
 > Compression and data type/precision are cv params for the
 > BinaryDataArray elements:
 > MS_zlib_compression = 1000574
 > MS_no_compression = 1000576
 >
 > MS_32_bit_integer = 1000519
 > MS_16_bit_float = 1000520
 > MS_32_bit_float = 1000521
 > MS_64_bit_integer = 1000522
 > MS_64_bit_float = 1000523
 >
 > Rune Schjellerup Philosof wrote:
 >> I looked for the capability in the latest (20080327) snapshot xsd from
 >> Lennart and couldn't find it.
 >> Isn't integer intensities also quite common.
 >> Data from the Waters Premier seem to be integer ion counts.
 >>
 >> -Rune
 >>
 >> Matt Chambers wrote:
 >>  
 >>> Hi Rune, mzML has both of these capabilities. With separate binary
 >>> arrays (like mzXML 3.0 theoretically supports but so far is rare in the
 >>> wild), the m/z array can be 64 and the intensity array can be 32, which
 >>> is most likely to fit current instrument capabilities. Also,
 >>> chromatogram time values probably don't need 64 bit either. SPC's
 >>> converter tools support writing and reading compressed mzML last time I
 >>> checked.
 >>>
 >>> -Matt
 >>>
 >>>
 >>> Rune Schjellerup Philosof wrote:
 >>>  
 >>>    
 >>>> In mzXML 3.0 there are two options for generating smaller files.
 >>>> 1. Saving in 32bit instead of 64bit.
 >>>> 2. Compressing the scans using zllib.
 >>>>
 >>>> Any special reasons why these features hasn't survived into mzML 
(yet)?
 >>>>
 >>>> --
 >>>> Regards
 >>>> Rune