Figure 3. Widespread gene birth, gene loss and gene conversion in vertebrate IFIT genes.

Summary table of the IFIT gene repertoire of 51 vertebrate species (see complete phylogeny in Figure 3—figure supplement 1 as well as expanded regions of trees built from the 5' and 3' end of the IFIT gene alignment in Figure 3—figure supplements 2 and 3). Colored boxes indicate the presence of a given IFIT gene sequence, with multiple copies indicated as a number to the right of the colored box. Lightened boxes indicate that only a partial or pseudogenized copy of the gene can be found in the genome. On the phylogenetic tree to the left, X's indicate deletion or pseudogenization of either IFIT1 or IFIT1B genes from that lineage (see Figure 3—figure supplement 4 for primate examples). Also shown are predicted occurrences of gene conversion (red-around-blue circles) between IFIT1 and IFIT1B sequences based on discordance between phylogenetic trees generated from 3' or 5' regions of the genes (for carnivore examples see Figure 3—figure supplement 2). Accession numbers for all sequences are found in Supplementary file 1.

DOI: http://dx.doi.org/10.7554/eLife.14228.009

Figure 3—figure supplement 1. Phylogeny of vertebrate IFIT genes.

Maximum likelihood phylogenetic tree of over 200 IFIT genes from 51 vertebrate species. Bootstrap values greater than 80% are shown along the supported branch. An ancestral IFIT was duplicated multiple times early in mammalian evolution, resulting in many species having IFIT1, IFIT1B, IFIT2, IFIT3 and IFIT5. As in Figure 1B, IFIT1 and IFIT1B genes cannot be resolved into distinct phylogenetic clades based on the full-length gene alignment.

DOI: http://dx.doi.org/10.7554/eLife.14228.010

Figure 3—figure supplement 2. Recurrent gene conversion is restricted to the 5' ends of IFIT genes.

GARD and SBP analyses were run on the complete IFIT gene phylogeny shown in Figure 3—figure supplement 1. These analyses indicate recurrent gene conversion between IFIT1 and IFIT1B paralogs 5' to the breakpoint indicated in red, while there is no evidence for gene conversion in the 3' end of the genes. Below are portions of the maximum likelihood phylogenetic tree of the recombining 5' region (left) and the non-recombining 3' region (right) of vertebrate IFITs highlighting only the region of the tree showing carnivore IFIT1 and IFIT1B gene sequences. Bootstrap values greater than 80% are shown along the supported branch.

DOI: http://dx.doi.org/10.7554/eLife.14228.011

Figure 3—figure supplement 3. Primate IFIT1 and IFIT1B genes.

A portion of the maximum likelihood phylogenetic tree of the non-recombining 3' region of vertebrate IFITs is shown as in Figure 3 supplement 2. Bootstrap values greater than 80% are shown along the supported branch. Primate IFIT1 and IFIT1B gene phylogenies are expanded, with pseudogenization of chimpanzee and New World monkey (marmoset and squirrel monkey) IFIT1B genes by frameshift and nonsense mutations (see Figure 3—figure supplement 4) indicated by red-strikethroughs. Subsequent analyses suggest that human IFIT1B also lacks activity, although it encodes an intact open reading frame.

DOI: http://dx.doi.org/10.7554/eLife.14228.012

Figure 3—figure supplement 4. Pseudogenization of IFIT1B genes from several primate lineages.

Location of frameshift and nonsense mutations in pseudogenized primate IFIT1B genes are shown. Gene sequences were aligned to intact human IFIT1B and frameshift-inducing insertions or deletions (red triangles) or nonsense mutations (red asterisks) are shown in the corresponding position in the sequence.

DOI: http://dx.doi.org/10.7554/eLife.14228.013