以下只用到了R中的ape包

1. 从BirdTree下载大量树文件

https://data.vertlife.org/?basetree=birdtree&start_folder=Stage2/

有Hackett和Ericson两种backbone。文献中见到的更多的是Hackett。

一个文件是1000个树（不同boostrap的topology），根据目标类群的大小和亲缘关系决定具体下多少。以下假定下载了所有树

2. 确定目标物种

假定目标物种都存在了向量sp里

3. 对每一个树做独立的剪枝

#### trim tree
filenames = c('AllBirdsHackett1.tre','BirdzillaHackett2.tre','BirdzillaHackett4.tre','BirdzillaHackett6.tre',
                'BirdzillaHackett8.tre','BirdzillaHackett10.tre','BirdzillaHackett3.tre','BirdzillaHackett5.tre',
                'BirdzillaHackett7.tre','BirdzillaHackett9.tre')

new_trees = list()
tree_count = 0
n = 1000
file_count = 0

for (filename in filenames){
  file_count = file_count+1
  trees = read.tree(paste0(in_tree_parent_path,filename))
  
  for (tree in trees){
    tree = drop.tip(tree, tip=tree$tip.label[!tree$tip.label %in% sp])
    tree_count = tree_count + 1
    new_trees[[tree_count]] <- tree
    cat("\rFile", file_count, "Finished", tree_count, "of", n)
  }
}

write.tree(new_trees, out_name) ### 写入10000棵树剪枝后的结果

4. 获得多数一致树（majority consensus tree）

#### consnsus tree
consensus_tree = consensus(new_trees, p = 0.5, 
                           check.labels = TRUE, rooted = T)

consensus_rooted_tree = root(consensus_tree, 'Struthio_camelus') ### 重新root一下

write.tree(consensus_rooted_tree, out_name3) ### 输出

也可以用Feng et al., 2020, Nature的文章中提供的一万多种鸟的树文件
Dense sampling of bird diversity increases power of comparative genomics