Trees

Phylogenetic trees of the 10,575 genomes.

Quick start

• tree.nwk is the recommended reference phylogeny. It is identical to astral - branch_length - cons - collapsed - astral.cons.nid.e5p50.nwk. See below for details.
• tree.qza is a pre-compiled QIIME 2 reference phylogeny, which can be directly feed into phylogeny-aware analyses such as UniFrac and Faith’s PD.

Tree-building strategies

The species trees were reconstructed using two alternate strategies:

• astral: Reconstructed using the gene tree summary method ASTRAL.
• concat: Reconstructed using the traditional gene alignment concatenation method.

In addition, there is:

• genes: Individual gene trees used for building the ASTRAL species tree.
• ranks: Species trees collapsed to individual taxonomic ranks.

General rules

• Rooting: Species trees were rooted in between the two clades representing domains Bacteria and Archaea, respectively.

• Rotating: Internal nodes were flipped to follow the descending order (child nodes with less descendants are shown in higher position).

• Node IDs (nid): Internal nodes were assigned incremental numbers in a pre-order traversal: root = N1, crown Archaea = N2, crown Bacteria = N3, so on so forth. These node IDs can be used as unique identifiers in downstream analyses and applications. Each topology (regardless post-manipulations such as collapsing and branch length re-estimation) receives the same set of node IDs. Node IDs of different topologies cannot be cross compared.

• Branch supports: Internal nodes were labeled by branch support statistics (see below).

• Branch collapsing: In the trees under the collapsed subdirectories, branches with low supports were collapsed, i.e., they were deleted and their child clades were merged into their parental nodes, making them polytomies.

• Identical sequences: In the CONCAT trees and the gene trees, prior to tree-building, duplicate sequences were removed from the alignments. Those taxa were then appended to the corresponding tips of the tree as polytomies.

Node metrics

Node metrics are provided in separate files, in the format of node ID to metric(s) mappings.

• supports (example): Branch support values. Different types of branch support statistics are used by different methods.

• splits (example): Metrics of branch splits (unrelated to branch lengths), including:
  • taxa: Number of descendants (tips) under current node.
  • prelevels: Number of splits from root to current node.
  • lmin, lmax, lmean, lmedian, and lstdev: Statistics of number of splits from all descendants (tips) to current node (i.e., postlevels).
  • splits: Total number of splits from all tips to current node.

• lengths (example): Metrics of branch lengths, including:

  • length: Length of branch connecting current node and its parent.
  • height: Sum of branch lengths from root to current node.
  • dmin, dmax, dmean, dmedian, and dstdev: Statistics of sums of branch lengths from all descendants (tips) to current node (i.e., depth).

  • red: Relative evolutionary divergence (RED), introduced by Parks, et al. (2018):

    RED = p + (d / u) * (1 - p)
    

    where p = RED of parent node, d = branch length, u = dmean of parent node.

Please also find taxonomic annotation of nodes (internal nodes and tips) under the taxonomy directory.