Update schema.md

docs/schema.md

@@ -1,59 +1,13 @@
-**Status**: Draft
+## CL KG Schema
 
-# OWL/RDF to Neo4j Schema
+Full details of the schema are now here:
+[CL_KG user stories, schema and roadmap](https://docs.google.com/document/d/1CIvy_NV1poK1wK-lY9E_sksOIRDxMyyBc-ZZLzD8OrM/edit#heading=h.vq3lz7r6domf)
 
-Defined in [documentation of owl2neo library](https://github.com/OBASKTools/neo4j2owl?tab=readme-ov-file#entities).
+For ontology representation see: 
+[OWL-2-NEO mapping](https://github.com/OBASKTools/neo4j2owl/blob/master/README.md#owl-2-el---neo4j-mapping-direct-existentials)
 
-## Nested cell sets:
 
-Cell sets are individuals representing author category cell type annotations.
 
-```cypher
- (c1)-[:INSTANCEOF]-(:Cluster { label: 'cluster' } ) // 'cluster' (PCL:0010001) # This should be improved!
-// Where one cell set subsumes another it is represented as
- (c1)-[:subcluster_of]->(c2) subcluster_of [RO:0015003](https://www.ebi.ac.uk/ols4/ontologies/ro/properties/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FRO_0015003)
-```
-subcluster_of is transitive, so a transitive reduction step MUST be used in generating the graph.
-
-All cell sets representing author cell type annotations MUST be present, however, if cell sets have identical membership, they are unified into a single node. Configuration specifies an order of preference for which annotation will become rdfs:label if nodes are unified.  All other names are stored with their original keys.
-
-TBD: Should we also represent overlaps between author annotations.  These could use RO:overlaps and record percent_overlap on the edge (should think about how this fits with confusion matrix generation)
-
-## Cell sets to Cell ontology terms
-
-The cell_type fields in the CELLxGENE schema also define cell sets.
-
-**All cell ontology terms MUST be represented.** 
-
-Where there is a 1:1 relationship between a cell set defined by a cell_type annotation and one represented by an author annotation, this is represented by:
-
-```cypher
-(c:Cluster)-[:composed_primarily_of]->(cl:Cell:Class) 
-```
-
-'composed primarily of' ([RO:0002473](https://www.ebi.ac.uk/ols4/ontologies/ro/properties/http%253A%252F%252Fpurl.obolibrary.org%252Fobo%252FRO_0002473))
-
-Where a cell set defined by a cell_type annotation doesn't map to single cell set defined by author category annotation, but subsumes >1 of these, we generate a cluster (cell set) node for the cell_type & relate this as above. One advantage of this is that it allows for CxG metadata to be consistently attached to an author annotation node.
-
-
-## Cell sets to standard [CxG metadata](https://github.com/chanzuckerberg/single-cell-curation/blob/main/schema/5.0.0/schema.md) (apart from cell ontlogy terms)
-
-```cypher
-(c:Cluster)-[:CxG_metadata_key { percentage: <float> }]-(x)
-```
-
-Where percentage = percent of cells in cell_set defined by author annotation that are in cell_set defined by metadata annotation.
-
-e.g.
-```cypher
-(c:Cluster)-[:tissue { percentage: 50.5 }]->(:Class { label: 'cornea', short_form: 'UBERON_'})
-```
-
-Above properties are reprented as OBASK builtin
-
-## Markers/marker sets
-
-TBA