Non-trivial study setups

[althonos]

Originally opened by @sneumann on althonos/mzml2isa:

Steffen:
Hi, there could be non-trivial experimental designs which are a challenge
to creation of the sample table, e.g. for technical replicates where multiple
assays map to one sample row. A check could be to compare MetaboLights s_Sample.txt versus generated sample tables.

[althonos]

Hi @sneumann, I based the converter on the studies available on Metabolights, and none of these had that kind of layout, so I could take that in consideration.

Unfortunately there is no real way to know the name of each sample, so mzml2isa uses the mzML filename. Are you saying it should be made possible to use an external table (such as a spreadsheet) to map each mzML file to each sample instead ?

[sneumann]

Hi, main point of this issue is to be aware of it.

mzML and mzData do support sample names, e.g. ftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS338/Bur-189_2-D,5_01_37808.mzData has <sampleName>Bur-189</sampleName>. Bruker instruments automatically add the autosampler position 2-D,5 and serial number of the injection 37808 to the assay name.

Similarly, ftp://ftp.ebi.ac.uk/pub/databases/metabolights/studies/public/MTBLS341/pos-Exp1-Blatt_K-4_1-B,4_01_8086.mzML was converted with <software id="exportSoftware" version="3.0.9"> and
has <sample id="sample" name="pos-E495-Blatt_K-4">

Yours, Steffen

[rsalek]

Is this only for Bruker? As does seems to work with other vendors.

On 27 Oct 2016, at 18:50, sneumann notifications@github.com wrote:

Hi, main point of this issue is to be aware of it.

mzML and mzData do support sample names, e.g. MTBLS338/Bur-189_2-D,5_01_37808.mzData has Bur-189. Bruker instruments automatically add the autosampler position 2-D,5 and serial number of the injection 37808 to the assay name.

Similarly, MTBLS341/pos-Exp1-Blatt_K-4_1-B,4_01_8086.mzML was converted with and
has

Yours, Steffen

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[proccaserra]

@sneumann is there an official Bruker documentation about this we could reference or point to ensure traceability of the feature development?
Is there always an autosampler in the hardware chain? (i suppose so for all National Phenome Centres)