Currently the FamPlex grounding map contains mappings for entities that aren't actually protein families or complexes, and are therefore not directly relevant to the main role of FamPlex. These mappings would be better maintained as part of INDRA, and the FamPlex grounding map could just focus on lexicalizations for FamPlex entries.