Relax genome requirements

[jameshadfield]

Quick experiments into relaxing the criteria for strains to be in the genome build.

Commit 1 allows strains with 7 sequenced segments to be included. There are currently 5 eligible strains but 2 are pruned by the clock filter (A/red-tailed_hawk/California/24-003714-001/2024, A/skunk/Utah/24-008032-004/2024), leaving 3 new ones. I also add 2 color-bys to better understand this - the number of sequenced segments (shown below) and the ATGC% (across the entire genome) to help identify these strains.

Commit 2 force-includes host==Human, however 12 of these are subsequently dropped by our clock-filter so we end up with 8 additional strains. Namely, A/Washington/252/2024 A/Washington/253/2024 A/Washington/251/2024 A/Washington/254/2024 A/California/171/2024 A/California/174/2024 A/California/172/2024 A/California/173/2024 A/California/182/2024 A/California/193/2024 A/California/191/2024 A/California/192/2024 are dropped.

Here's the resultant tree filtered to the additional sequences (fewer than 8 segments sequenced):

[jameshadfield]

I'll defer to @lmoncla & @trvrb here - I think it's fine to relax to 7/8 segments but I don't think force-including (human) samples with <<8 genomes is a good idea. (One of the reason I added segment builds was to be able to analyse samples such as those.)