The evolution of the tree of life

Abstract

The tree of life is one of the most important organizing principles in biology. Updates and revisions are historically derived from improved data capture, increasingly refined models of evolution and expanded taxon sampling. Tracing the changes in the tree of life over the molecular era (1990-present) highlights the evolution of biologists' understanding of life on earth and serves as a foil placing the explosion of available data over this timeframe in context. Using current-day information, we explored the taxonomic growth captured in a tree of life through historic tree reconstruction. Data capture is now facilitating improvements in genome quality rather than expanding deep diversity, as the rate of novel phylum discovery is slowing for bacteria and archaea. Using dissimilarity metrics, the proportion of changes that each historic tree encompasses identified a diminishing influence of additional taxa on high-level topological revisions. No trees recapitulated current hypotheses for deep relationships on the tree of life, reflective of disadvantages associated with high taxon sampling and the divide-and-conquer methodologies required to analyse extremely large datasets. This work clarifies the effect of the interaction between data quality, data quantity and taxonomic diversity on our ability to construct a stable tree of life.This article is part of the discussion meeting issue 'Chance and purpose in the evolution of biospheres'.

Keywords: diversity; evolution; phylogenomics; taxon selection; tree of life.

Figure 1.

Timeline of universal trees of life, by total information included in the trees. Circles represent individual universal trees of life, with information calculated as # taxa × # alignment positions (×3 for amino acid alignments). (A) Published universal trees of life. (B): Published universal trees of life and reconstructed trees of life from this study (new scale). (C): Timeline in B, with other, large-scale two-domain trees included (orange), same scale as B.

Figure 2.

Taxon sampling versus molecular information used to construct universal trees of life. (A) All published universal trees and subsets of reconstructed trees from this study and from other significant two-domain trees. (B): Expanded view, including all reconstructed trees from this study (diamonds) and the full set of other significant two-domain trees included in figure 1 and electronic supplementary material, table S1. All universal trees are coloured by the 5-year window within which they were generated. Reconstructed (diamond) and two-domain (triangle) trees included on both plots are outlined to facilitate connection between the two x-axis scales.

Figure 3.

Reconstructed trees based on historical genomic sequencing data from 1999 to 2019. Trees are shown chronologically, each capturing an additional 5 years of data from the GenBank genome database. Clades are annotated at the phylum level with GTDB nomenclature, except for eukaryotes as they do not have associated GTDB representatives. To facilitate legibility, not all available phyla in trees b–d are shown. Colours indicate the reconstructed tree in which the phylum first appeared, in chronological order.

Figure 4.

Reconstructed universal tree of life using genome sequencing data from 1999 to 2024. The tree was constructed using available sequences in the GenBank sequence database as of August 2024. A total of 69 496 organisms were placed by uDance. Annotations are based on phylum-level taxonomy assigned by GTDB. Phyla with 35 or more representatives are shown, as well as selected phyla present in figure 3. Colours indicate the reconstructed tree that the phylum first appeared in, in chronological order.