BDSO - simplifying pipeline

[dosumis]

BDSO - simplifying pipeline

Aims:

• A curation system that allows editor to link cell types and properties to taxonomy and non-taxonomy node IDs without being overwhelmed by columns

• Auto-populated content input should be segregated from manually
  curated content input.

Questions: General question - do we need all the comment/annotation
stuff? It doesn't get used much.

Current patterns - to merge down

brainCellRegionMinimalMarkers.yaml

Description: Drives annotation axioms and subClassOf axioms, excluding
markers

vars: gross_cell_type: "'cell'" # From YAML config for whole
taxonomy taxon: "'Vertebrata <vertebrates>'" # From YAML config
for whole taxonomy brain_region: "'regional part of brain'" #
From YAML config for whole taxonomy projection_type:
"'projection type'" # From curation to nodes

data_vars: cell_set_preferred_alias: "xsd:string" #
From Dendrogram/taxonomy brain_regionabbv: "xsd:string" #
Currently from Taxonomy config species_abbv: "xsd:string" #
Currently from Taxonomy config

list_vars: minimal_markers:
"'sequence_feature'" # From markers allen_markers:
"'sequence_feature'" # From curation to node IDs layers:
"'layer of neocortex'" # From curation to node IDs

data_listvars: alias_citations: "xsd:string" #
Curation to node IDs (based on hard to parse info in taxonomy?)

taxonomy_class.yaml

Example ROBOT template:
CCN202002013_class.tsv

Description: Linking to node IDs - some autogenerated. most curated.

logic:

• Intermediate level classifications (e.g. LAMP5 GABAergic), but not MET classes
• Expression of markers used in classification (potentially redundant?)

annotation:

• synonyms from dendrogram
• curated synonyms
• publications and comments annotating classification expression and
  synonym axioms.
• prefLabel - Do we still want this?

list_vars: Expresses: "'thing'" # Expression linked to leaf nodes.

vars: Classification: "'cell'" # Classifications linked to taxonomy nodes.

data_list_vars: Synonyms from taxonomy
Curated_synonyms: "xsd:string"
Classification_pub: "xsd:string"
Expresses_comment: "xsd:string"
Expresses_pub: "xsd:string"

data_vars: prefLabel: "xsd:string" Comment: "xsd:string"
Classification_comment: "xsd:string"

taxonomy_equivalent_class.yaml

DESCRIPTION: Rolls EC definitions - really only needs one var.

ROBOT_TEMPLATE_EXAMPLE

vars: Exemplar_SC: "'thing'" # contents same as exemplar
Exemplar: "'thing'"

taxonomy_minimal_markers.yaml

ROBOT TEMPALTE:
CCN202002013_minimalmarkers.tsv

Description: NS_Forerst minimal marker axioms. + part_of/has_soma_location depending on type

list_vars: Markers: "'thing'" # NS_Forest markers

vars:

• part_of: "'thing'" # Driven from gross typing in taxonomy config
• has_soma_location: "'thing'" # Driven from gross typing in taxonomy config

taxonomy_non_taxonomy_classification.yaml

Description: Pulls in some additional/orthogonal classifications - CURATED

vars: Classification: "'cell'" # Driven from non-taxonomy nodes in YAML

Proposal

• We have 1 Mega DOSDP pattern - with some naming standard for vars
• The TSV for this comes from:
  • A minimal manually curated table including both tax and non-tax nodes.
  • Auto-generation that consumes:
    • taxonomy (dendrogram/tsv)
    • taxonomy config - including some spec for gross types and assoc loc relations
    • NS-forest