HYPHY treats deletions in the data in one of two possible
ways:
- Keep Deletions: Deletions are treated as complete unknowns.
In case of codons, data like A-T will be treated as A?T (i.e.
either AAT, ACT, AGT or ATT).
- Skip Deletions: Any site which contains at least one deletion
will be excluded from the analysis. In case of codons, data like
A-T will result in all 3 data columns being skipped.
In general,
HYPHY treats ambiguities by averaging over all possible
translations, both in likelihood evaluation and in frequency
gathering.
|