make choice of BA.2 "reference" at root more robust

[jbloom]

Right now the spike for the reference is hardcoded into the script Trevor generated (81e43a2#diff-54f0382e7ad5485edc0f8936ebc650ad9a30d03bd39bcd4e0b1080800bf757afR15-R17), which doesn't seem ideal. How do we make this better and read from data?

[trvrb]

Right now, we're using SouthAfrica/NHLS-UCT-LA-Z957/2022 as basal BA.2. This manifests in multiple locations:

Data files:

• data/ba2_outgroup_metadata.tsv
• data/ba2_outgroup_sequences/fasta

refine:
  root: "SouthAfrica/NHLS-UCT-LA-Z957/2022"

in escape_profiles/escape/builds.yaml

reference is set in escape_profiles/escape/sequence_to_escape.py. I grabbed this from the workflow artifact results/escape/translations/aligned.gene.S.fasta

[trvrb]

I'm actually not very worried about having to have multiple places where this shows up. However, it would be good to make sure that this is a properly basal BA.2 virus.

[jbloom]

Sounds good. Is there a recommended way to determine what is a good basal BA.2 virus?