diagnostics_and_testing.md

Experiments with information theoretic tree generation

We’ll generate 10 blocks of sequence data for 5 species where each block corresponds to a tree randomly generated in the TN93 model.

trees <- list()
sequences <- list()

for(i in 1:10){
  who_dat(sites = 1000, 
          model = 6, 
          parameters = "5 5", 
          gamma = "0.5 4", 
          equilibrium = "0.25 0.28 0.34 0.13")
  system("paml4.8/src/evolverRandomTree 5 paml4.8/MCbaseRTree.dat") 
  trees[[i]] <- write.tree(read.nexus("mctrees.nex"))
  sequences[[i]] <- ReadCharacters("mc.nex")
}

Let’s compute some diagnostics!

1. Robinson-Foulds distances:

#trials <- c(1:100)
divisive <- lapply(lapply(sequences, infotree), function(x) paste(x,";", sep = ""))
agglomerative <- lapply(lapply(sequences, agg_clustering), function(x) paste(x,";", sep = ""))
nj_trees <- sequences %>%
            lapply(as.DNAbin) %>%
            lapply(dist.dna) %>%
            lapply(nj) %>%
            lapply(write.tree)

## [1] "Average distance of divisive trees from original: 0"

## [1] "Average distance of agglomerative trees from original: 0"

## [1] "Average distance of NJ trees from original: 0.2"

2. Low tolerance ultrametricity:

summary(metricity)

##  as.character.trees. as.character.divisive. as.character.agglomerative.
##  Mode:logical        Mode :logical          Mode :logical              
##  TRUE:10             FALSE:10               FALSE:10                   
##  as.character.nj_trees.
##  Mode :logical         
##  FALSE:10

3. High tolerance ultrametricity:

summary(low_metricity)

##  as.character.trees. as.character.divisive. as.character.agglomerative.
##  Mode:logical        Mode :logical          Mode :logical              
##  TRUE:10             FALSE:10               FALSE:10                   
##  as.character.nj_trees.
##  Mode :logical         
##  FALSE:10