Update nullable fields, add additional fields to AlleleDescription

[williamdlees]

In germline objects, remove all nullable: false entries and add x-airr: qualifiers instead.

• Move aligned sequence to SequenceDelineationV and change co-ordinates in SequenceDelineationV to be against the unaligned sequence rather than the aligned sequence. This makes AlleleDescription delineation-agnostic and removes reliance on an alignment for the co-ordinates
• Add fields to AlleleDescription for Allele Similarity Cluster designation

[williamdlees]

Now has the two items agreed at our last meeting for implementation in this release:

• use miairr 'importance' designation in germline objects
• de-nest objects in Genotype

[williamdlees]

fixed UndocumentedAllele name

[williamdlees]

Withdrawn while I work on updating the test data

[bussec]

@williamdlees please find my comments below:

• The following properties should IMO be essential:

  • RearrangedSequence.derivation
  • RearrangedSequence.observation_type
  • AlleleDescription.species
  • GermlineSet.species
  • MHCGenotype.mhc_class

• AlleleDescription.coding_sequence.description: Is it intended to refer to IMGT Ontology here?

• AlleleDescription.allele_similarity_cluster* needs more description and an example. The description should clarify whether "similarity" is different from "identical" and if yes to which degree.

• Could Genotype.receptor_genotype_set_id be renamed to genotype_set_id to avoid confusion with the receptor_* properties of the Receptor object.

• MHCGenotype.mhc_alleles is an array-of-objects (and not an array-of-refs)

• As a note: According to wiki the term "derivation" does not refer to the source of the process (or its at least ambigous).

[javh]

From the call:

• species, mhc_class, and observation_type as essential makes sense.
• Check with IARC on RearrangedSequence.derivation.
• coding_sequence.description can we be more explicit here? Can we use the following as a guide? https://docs.airr-community.org/en/stable/miairr/specification_miairr_ncbi.html.
• Can't change Genotype.receptor_genotype_set_id in a minor release. Maybe v1.5.
• Replacement of object (MHCGenotype.mhc_alleles) with reference should be fine for v1.4.2.

Good to merge in after the above fixes.

[williamdlees]

• species, mhc_class, and observation_type marked as essential.
• No strong views from IARC regarding RearrangedSequence.derivation (the name waa reviewed and agreed by IARC some years ago). Possible alternatives should they be required could be derived_from, sample_source, nucleic_acid_class..
• coding_sequence.descriptionchanged to nucleotide sequence of the core coding region, i.e. the coding region of a D-, J- or C- gene, or the coding region of a V-gene excluding the leader. I think this is pretty specific now.
• object (MHCGenotype.mhc_alleles) replaced with reference to a new top-level object MHCAllele.

[bcorrie]

Re-structuring of MHCGenotype and Genotype objects to include object references rather than embedded objects looks like it is correct to me.

[bcorrie]

MHCGenotype and Genotype object restructuring (no more embedded objects) looks like it is correct to me.

[javh]

From the call:

• Add additional description and clarification for AlleleDescription.allele_similarity_cluster* in the Germline documentation (see above comment from @bussec).