Threshold feature metadata coloring

I very much like the ability to color sub-trees according to their uniqueness regarding metadata categories. I think it would be even more useful if I could somewhere define a minimal number of reads (or better: percentage of relative abundance) that the combined features in that subtree should have to be colored.
Thus, I could ignore features with only a handful of reads and focus on those that might have a biological impact.

Start by locating the metadata category coloring implementation and its related tests; the issue does not name specific files or entry points. Define how a minimum read count or relative-abundance threshold should be provided and applied to combined subtree features, then add coverage showing that low-abundance features are excluded from coloring.

Written by the indexing model from the issue text.