taxonomy of human readable annotation row types for use in UI

Annotation rows, like sequence features fall into a set of canonical types - referring to the type of information they present: e.g. disorder predictions, etc.

Ideally, it should be possible to show/hide all rows of a particular type for some set of source data, or select one from a range of possible sources of analysis result of a particular type. This requires some canonical model to bind calcId strings to human readable descriptions of the data that the annotation rows present.