Skip to content

Spectrums and subclasses

Peter Robinson edited this page Sep 24, 2017 · 1 revision

Representing spectra as subclasses

In a number of places in HPO there are cases where the subclass hierarchy is essentially trying to describe a spectrum. This is needed to adhere to the true path rule, the most severe is the child term, and the more atomic phenotype the parent term. This is logically consistent but biologically unintuitive.

This ticket is to discuss this later if there should be some other relationship between the terms. This is not dissimilar to something being somewhat hypoplastic, very hypoplastic, and aplastic, too. Need to think about how the current implementation or any changes would affect the semantic similarity or annotations.