Lab Notebook: August 20
- Wrap hyphy to make rate-estimation not suck. We can crib code from tapir for this, or roll something new for an all-R ecosystem.
- Track what sites go into which bins and visualize this somehow.
- Try using rate-partitioning on the McCormack mammal UCE dataset.
- Build a neighbor-joining tree with R, if possible
- Use hyphy to estimate rates
- Rate partition!
UCE analysis pipeline:
- Root gene trees from PhyML with a ranked list of taxa (not every gene tree will have every tip)
- Cluster UCEs based on the GTR+Gamma substitution matrix. Something like hcluster in R should be able to do this
- Divide UCEs into flank and core regions based on the original probe data
- Find shift points and print the labels of the nodes there and its descendants
- Investigate “discrete auteur”
- Ask Dan for BAMM code, which accounts for missing data
- Build a phlawd tree for gaddis, cod, or salmon, or use the darters from Near’s paper
- See if similar exploitation patterns hold up at these smaller scales