Gene gain and loss across the metazoan tree of life

Abstract

Although recent research has revealed high genomic complexity in the earliest-splitting animals and their ancestors, the macroevolutionary trends orchestrating gene repertoire evolution throughout the animal phyla remain poorly understood. We used a phylogenomic approach to interrogate genome evolution across all animal phyla. Our analysis uncovered a bimodal distribution of recruitment of orthologous genes, with most genes gained very 'early' (that is, at deep nodes) or very 'late', representing lineage-specific acquisitions. The emergence of animals was characterized by high values of gene birth and duplications. Deuterostomes, ecdysozoans and Xenacoelomorpha were characterized by no gene gain but rampant differential gene loss. Genes considered as animal hallmarks, such as Notch/Delta, were convergently duplicated in all phyla and at different evolutionary depths. Genes duplicated in all nodes from Metazoa to phylum-specific levels were enriched in functions related to the neural system, suggesting that this system has been continuously and independently reshaped throughout evolution across animals. Our results indicate that animal genomes evolved by unparalleled gene duplication followed by differential gene loss, and provide an atlas of gene repertoire evolution throughout the animal tree of life to navigate how, when and how often each gene in each genome was gained, duplicated or lost.

Figure 1. Gene gain, loss and duplication ratios show a similar pattern across animal phyla.

Phylogenetic hypothesis of animal phyla interrelationships. Number of taxa per phyla corresponding to dataset 1 is indicated in each case. Nodes of interest are highlighted in different colors. Gene gain, loss (histograms in main branches) and duplication ratios (coloured circles) as inferred for each phyla in dataset 2 are shown. Colors correspond to nodes as defined in the legend. Mean (x) and standard deviation (S) values of gene gain and loss are indicated in each histogram. Size of circles representing duplication ratios are proportional to the size shown in the corresponding legend, excepting for Opisthokonta and Holozoa, where values are shown in brackets.

Figure 2. Gene gain and duplication ratios are high at deeper nodes, and gene loss at shallower ones.

Box plot graphs of gene gain (a), loss (b) and duplication ratios (c) per node as inferred from dataset 1. Mean and standard deviation values are shown for each node. Colors indicatenodes as shown in the legend.

Figure 3. Genes related to the neural system are amongst the most highly duplicated.

a. List of top 10 most duplicated gene families as defined by OrthoFinder2. Group of taxa in which these families were inferred (metazoa or outgroups), mean and maximum number of duplications across animals and putative functions are indicated. b. Treemap representation of GO enrichment analysis of duplicated genes related to neural system (light blue) in all nodes in Ctenophora, Porifera, Placozoa and Craniata (see also Suppl. Mat.). Other colors represent other functions as shown in the legend. Each main square represents hierarchical GO term relationships. Square size is proportional to the p value for each GO term as found in the enrichment analyses. Scales are irrelevant.

Figure 4. Pairwise gene loss is pervasive across phyla.

a. Chord diagram representation of the dynamics of gene loss observed between each pair of taxa measured as the percentage of genes in each genome that were lost, as inferred for dataset 3 (phylome approach). Each seed phyla is represented by one color. Thicker chords between two given phyla represent higher percentages of gene loss. Directionality of chords is defined by each color (i.e., if a chord and a phylum share a color, the percentage of gene loss is calculated based on the genome size of that phylum). b. Percentage values of gene loss as inferred for dataset 3. Values are polarized with the genome size of each seed genome shown in the left-most column.

Figure 5. The core gene repertoire of metazoans includes genes from a plethora of KEGG pathways that have undergone different degrees of duplication.

a. Heatmap of pathway conservation and duplication level in the core gene repertoire of metazoans. The reference pathways were selected from the KEGG pathways collection. The color in each cell depicts the relative number of duplicates within each KEGG category found in each phyla as inferred for dataset 2 (for total number of duplicates and specific KEGG pathway annotation per orthogroup see Suppl. Mat. S9). Tree topology and colors for each clade are as in Fig. 1. Asterisks indicate specific pathways expanded in 5b and 5c. b. Percentage of each KEGG category annotated in all orthogroups from the metazoan core gene repertoire. c,d. Heatmap of KEGG nervous system (5c) and signal transduction (5d) pathway conservation and duplication level in the core repertoire of metazoans. The color in each cell is calculated as in 5a.