Multiple values encoding about mzqc HOT 5 CLOSED

So, it would be nice to have a definite description in the cv term definition, but still a somewhat self-explanatory and insta-parsable content in the cv term representation. My biggest concern here would be units because having them together with the values makes good for a lot of validation. Not all our values have a unit, but we still might want to enforce unit on every value and accept these cases where we have to put "no-unit" there and then

from mzqc.

I think there are more than the two options if you consider term definition + units and current schema, then you could choose
Sub element:

[Term]
id: QC:xxxxxxxx
name: Interval Acquisition range m/z
def: "TODO: description of this NIST metric group" [PSI:QC]
is_a: QC:3000016 ! QC metric
is_a: QC:xxxxxxxx ! tuple with unit 

[Schema]
<qualityParameter ID="METRIC001" cvRef="PSI-QC-CV" accession="QC:xxxxxxxx" name="Interval Acquisition range m/z"> 
    <content cvRef="PSI-QC-CV" accession="n - tuple" value='2'>{'MS:1000043':[400,1500]}</content>
</qualityParameter>

Explicit units:

[Term]
id: QC:xxxxxxxx
name: Interval Acquisition range m/z
def: "TODO: description of this NIST metric group" [PSI:QC]
is_a: QC:3000016 ! QC metric
is_a: QC:xxxxxxxx ! tuple 

[Schema]
<qualityParameter ID="METRIC001" cvRef="PSI-QC-CV" accession="QC:xxxxxxxx" name="Interval Acquisition range m/z" value="{'MS:1000043':[400,1500]}" /> 

Implicit 1:

[Term]
id: QC:xxxxxxxx
name: Interval Acquisition range m/z
def: "TODO: description of this NIST metric group" [PSI:QC]
is_a: QC:3000016 ! QC metric
is_a: QC:xxxxxxxx ! tuple with unit MS:1000043

[Schema]
<qualityParameter ID="METRIC001" cvRef="PSI-QC-CV" accession="QC:xxxxxxxx" name="Interval Acquisition range m/z" value="{[400,1500]}" /> 

Implicit 2:

[Term]
id: QC:xxxxxxxx
name: Interval Acquisition range m/z
def: "TODO: description of this NIST metric group" [PSI:QC]
is_a: QC:3000016 ! QC metric
is_a: QC:xxxxxxxx ! unit MS:1000043

[Schema]
<qualityParameter ID="METRIC001" cvRef="PSI-QC-CV" accession="QC:xxxxxxxx" name="Interval Acquisition range m/z" value="{"1": 400, "2": 1500]}" /> 

from mzqc.

I guess what speaks against the implicit ones is that a good practice is keeping the attribute value short. Also, we would most likely run into (differing) length limitations, based on the XML parsing tool used -although the xml standard does not impose any length limitations afaik.

from mzqc.

I'd definitely prefer a list over key-value pairs, certainly if the keys are just indices as in the "Implicit 2" example.

Explicitly specifying the units seems like good practice to me. Most units in mass spec are standardized, so usually there shouldn't be any confusion, but still, sometimes people like to do things differently. Some people are just contrarian. 🙄 In that case explicitly specifying a unit could save us from a lot of trouble.

I'm not sure what the advantage of the sub element approach would be? This just adds extra complexity but wouldn't really make validation or other things easier as far as I can discern right now.

from mzqc.

XML-JSON no longer relevant after change to full JSON. See #45 for the discussion on how to store multiple values in JSON.

from mzqc.