Recent, full-length gene retrocopies are common in canids

Abstract

Gene retrocopies arise from the reverse transcription and insertion into the genome of processed mRNA transcripts. Although many retrocopies have acquired mutations that render them functionally inactive, most mammals retain active LINE-1 sequences capable of producing new retrocopies. New retrocopies, referred to as retro copy number variants (retroCNVs), may not be identified by standard variant calling techniques in high-throughput sequencing data. Although multiple functional FGF4 retroCNVs have been associated with skeletal dysplasias in dogs, the full landscape of canid retroCNVs has not been characterized. Here, retroCNV discovery was performed on a whole-genome sequencing data set of 293 canids from 76 breeds. We identified retroCNV parent genes via the presence of mRNA-specific 30-mers, and then identified retroCNV insertion sites through discordant read analysis. In total, we resolved insertion sites for 1911 retroCNVs from 1179 parent genes, 1236 of which appeared identical to their parent genes. Dogs had on average 54.1 total retroCNVs and 1.4 private retroCNVs. We found evidence of expression in testes for 12% (14/113) of the retroCNVs identified in six Golden Retrievers, including four chimeric transcripts, and 97 retroCNVs also had significantly elevated FST across dog breeds, possibly indicating selection. We applied our approach to a subset of human genomes and detected an average of 4.2 retroCNVs per sample, highlighting a 13-fold relative increase of retroCNV frequency in dogs. Particularly in canids, retroCNVs are a largely unexplored source of genetic variation that can contribute to genome plasticity and that should be considered when investigating traits and diseases.

Interchromosomal Colocalization with Parental Genes Is Linked to the Function and Evolution of Mammalian Retrocopies

“…Similar to Zhang et al (2021), we did not distinguish homozygotes and heterozygotes, only accounted for the presence or absence of each allele in different individuals. Our list captures most retroCNVs identified in other studies (Abyzov et al 2013; Ewing et al 2013; Schrider et al 2013; Zhang et al 2017; Feng and Li 2021; Batcher et al 2022) ( supplementary table S5 ) and the neighbor-joining tree based on retroCNV presence frequency is topologically the same as the tree built from genome-wide SNPs ( supplementary fig. S37 ), indicating the overall accuracy of retroCNV identification.…”

Duplications and retrogenes are numerous and widespread in modern canine genomic assemblies

“…The most well-studied retrogenes in dogs are derived from the FGF4 gene which cause the short-leg phenotype found in some breeds such as corgis and beagles [12]. FGF4 retrogenes have occurred multiple times independently, with 7 independent copies identified to date [13, 66]. Additionally, dogs have a high rate of SINE polymorphism [34, 67].…”

“…Other observations are consistent with a high rate of retrogene formation in dogs. A recent survey suggests that polymorphic retrogene insertions are 13-fold more common in dogs than in humans [13]. The most well-studied retrogenes in dogs are derived from the FGF4 gene which cause the short-leg phenotype found in some breeds such as corgis and beagles [12].…”

“…Gene conversion and changes in pseudogene sequence can also give rise to actual and novel function, altering the name from pseudogenes into functional retrogenes. Retrocopies commonly occur in dogs [13,14] and confound the interpretation of genetic variation [15]. Novel developments in sequencing technology have allowed for the rapid and thorough analysis of thousands of samples.…”

Identification of Genetic Risk Factors for Monogenic and Complex Canine Diseases

“…(a) Different SV types are indicated schematically. The insertion of intronless retrogenes is more common in dogs than in other mammalian species(145). (b) Schematic representations of a conventional linear reference genome assembly and a variant-aware graph assembly.…”