bweedop/edgesims

EDGE Score Simulations

EDGE Score Simulations:

The R scripts used here have been created in order to 1) simulate phylogenies under a pure-birth Yule model, 2) manipulate the simulated phylogenies, 3) calculate ED (Evolutionary Distinctiveness) values, and 4) calculate the correlations between the original and manipulated simulated phylogenies' ED values. The manipulations carried out on the simulated phylogenies include dropping tips randomly, dropping tips if some threshold of a simulated characteristic value is reached, and applying an imputation of a particular clade of specified size.

edgesims's Issues

Analysis things to do

• Do a quantile regression on figure 1 (the infamous focal ED correlations). Check the lower, upper, and middle quantiles (25, 50, 75? ish?) to show that the value is 'converging' on 0 (ish)
• Drop number of species in focal clade down to 4 or 5; justify in methods because we're calculating correlations. I don't think we need to do something special for clades of size two (...I could be wrong...)
• Increase number of species in dropping simulation to 75% (why not?)

Parameters for simulation

• Number of species
• Number of loci
• DNA evolution model
• Sequence length
• DNA rate parameters: substitution rate and gap rate

Finishing off pipeline

• Setup phyloGenerator (...harder than it sounds!...)
• Sequence alignment (using MAFFT) using phyloGenerator
• Phylogeny search using BEAST using phyloGenerator
• Wrapper functions in R for the above