my currently k-most-dissimilar approach is overengineered, which wouldn't be an issue except that it is going to be very useful to maintain coordinates w.r.t. WH1.
Note that this will also simplify (a lot) the alignment masking, since the algorithm can now look something like:
Clean up GISAID data
Align everything to WH1 (this gives an un-masked alignment)
Mask (don't trim) sites (this means coordinates are all maintained)
Estimate tree
This will save a lot of time compared to what I'm currently doing, and increase utility by maintaining alignment coordinates, and maintaining a completely un-masked copy of the alignment.
roblanf
commented
4 years ago
my currently k-most-dissimilar approach is overengineered, which wouldn't be an issue except that it is going to be very useful to maintain coordinates w.r.t. WH1.
Note that this will also simplify (a lot) the alignment masking, since the algorithm can now look something like:
This will save a lot of time compared to what I'm currently doing, and increase utility by maintaining alignment coordinates, and maintaining a completely un-masked copy of the alignment.