Validator - incorrect error about SearchType CV term

[vrkosk]

Mascot Server 2.6 and later support combined spectral library and FASTA searches. The software exports the type using a valid CV term:

    <SpectrumIdentificationProtocol id="SIP" analysisSoftware_ref="AS_mascot_server">
      <SearchType>
        <cvParam accession="MS:1002755" name="combined ms-ms + spectral library search" cvRef="PSI-MS" value="" />
      </SearchType>

Term definition in psi-ms.obo:

id: MS:1002755
name: combined ms-ms + spectral library search
def: "A combined MS2 (with fragment ions) and spectral library search." [PSI:PI]
is_a: MS:1001080 ! search type

The validator isn't happy and gives two errors:

Message 2:
    Rule ID: SearchTypeObjectRule
    Level: ERROR
    Context(/MzIdentML/AnalysisProtocolCollection/SpectrumIdentificationProtocol )
    --> At least one child term of 'search type' must occur in SearchType of the SpectrumIdentificationProtocol (id='SIP') element at /MzIdentML/AnalysisProtocolCollection/SpectrumIdentificationProtocol
    Tip: The SearchType element of SpectrumIdentificationProtocol must contain a CV term.

Message 3:
    Rule ID: SearchType_must_rule
    Level: ERROR
    Context(/searchType/cvParam/@accession ) in 2 locations
    --> The result found at: /searchType/cvParam/@accession for which the values is  ''MS:1002755'' didn't match any of the 6 specified CV terms:
  - The sole term MS:1001010 (de novo search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name.
  - The sole term MS:1001031 (spectral library search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name.
  - The sole term MS:1001081 (pmf search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name.
  - The sole term MS:1001082 (tag search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name.
  - The sole term MS:1001083 (ms-ms search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name.
  - The sole term MS:1001584 (combined pmf + ms-ms search) or any of its children. A single instance of this term can be specified. The matching value has to be the identifier of the term, not its name

The validator is wrong. The mzIdentML 1.1 specification section 6.63 says "MUST supply a child term of MS:1001080 (search type)" and MS:1002755 is a child of MS:1001080. mzIdentML 1.2 specification section 6.66 says the same.

[colin-combe]

MS:1002755 is missing from here:
https://github.com/HUPO-PSI/mzIdentML/blob/master/validator/resources/mzIdentML-mapping_1.2.0.xml#L128-L135
right?

[vrkosk]

I'm not familiar with the mapping file or how it's used by the validator, but looks probable.

[colin-combe]

I'm not familiar with the mapping file or how it's used

me neither

it looks like the validator maybe isn't maintained at the moment?

we could prob fix this, but maybe it makes more sense to look at it after mzIdentML 1.2.1 is finalised (there'll be a few changes to be made to the validator then)