Display inclusive frequencies for named clades

[huddlej]

Context

We estimate frequencies for mutually exclusive clades/haplotypes in the MLR, but displaying the frequencies in this way causes ancestral named clades to appear at low frequency when their emerging haplotypes are at high frequency.

Description

We should display inclusive frequencies for named clades, summing frequencies like "K:80K", "K:88I", etc. into frequencies displayed for "K", for example.

Possible solution

We could calculate these inclusive frequencies in the visualization application or in the MLR JSONs in the workflow.

We might want to label the inclusive frequencies differently from other groups, to indicate that those frequencies are not the mutually exclusive values from the MLR. In the previous "flu_frequencies" workflow, inclusive frequencies had a * label like K*.

[huddlej]

Another way to address this issue would be to add a third variant_classification to the workflow of "clade" which only runs the MLRs on high-level clade annotations. This third workflow would run very quickly, but it would require a third set of panels to be displayed for each subtype/geo resolution combination which could get annoying.

The benefit of this approach is it shows when a large clade like H3N2's K is predicted to fix in the future while still communicating the patterns of clade competition modeled by the MLR. Then users can scrolls to the "emerging haplotypes" view to see which new subclades within K are outcompeting the ancestral clade.