Add an attribute to add specificity in ontology hierarchy, including distinguishing 'entities' from 'groupings'

[cmungall]

Proposal:

1. a general purpose attribute that can be used on nodes that correspond to ontology classes to indicate "level" in the hierarchy
2. a vocabulary or vocabularies to use as values for this attribute. This could be course (e.g "entity" vs "grouping") or fine-grained (specific to the overall category)

Examples:

TM provider (cc @mikebada) uses PRO. PRO has proteins at different levels (see #230)

PR:000021998 neurotrophin à level=family
     PR:000011460 neurotrophin-4 à level=gene
          PR:Q80VU4 neurotrophin-4 (mouse) à level=organism-gene
               PR:Q80VU4-1 neurotrophin-4 isoform 1 (mouse) à level=organism-sequence

Note that "databases" such as uniprot are largely concerned with a level or small set of levels (e.g. uniprot is largely at level organism-gene, but they also have isoform IDs, e.g P12345-2)

NCBITaxon provides taxa at different levels from species up to kingdom (including unnamed levels)

the chemical working group (@vdancik) are considering splitting chemical "entities" (e.g. serine or L-serine) from "groupings" (e.g. amino acid). The alternative proposal is to follow chebi and treat these all as chemical entities and use a level attribute to distinguish groupings from entities (i.e. things that have a formula/inchi),

[sierra-moxon]

The DM committee is scheduled to adopt the chemical WG proposal for adding grouping entities. Removing the working-group/chemical tag in an effort to keep the remaining ideas in play for our 'conflation' work, but also to help clear out the chemical WG queue with the release of their 2.0.0 changes. :)