Make the Diseases and Function data more useful

[MrHedmad]

Currently, the diseases and function tables are pretty useless, with just string descriptions of the diseases and functions. They are mostly unique entries. It would be better to either condense them into categories (so they are useful for gene clustering) or run some other sort of clustering on them.

I currently see three ways we could do this:

• Like the thesaurus, manually go over them and add the data (but its long and tedious);
• Using the word vectors provided by the iUPHAR, run clustering (but the name / description does not really matter for the actual meaning of the phrase, especially for diseases).
• Scrap the whole idea and use cherry-picked GO terms instead (which? how to be unbiased? GO download is also very slow)