[Class hierarchy] T-types with reference clusters that do not subsume each other should not be subclasses of each other

[dosumis]

Lamp5 Lhx6 neuron SubClassOf Lamp 5 neuron, but transcriptomically they are distinct:

https://cellxgene.cziscience.com/e/e4ddac12-f48f-4455-8e8d-c2a48a683437.cxg/

Another way to think about this is that if these two classes were defined with suitably specific marker sets derived from the transcriptomics data, they would not subsume each other.

• see also SST and SST CHODL.