Long-read assembly of a Great Dane genome highlights the contribution of GC-rich sequence and mobile elements to canine genomes

Abstract

Technological advances have allowed improvements in genome reference sequence assemblies. Here, we combined long- and short-read sequence resources to assemble the genome of a female Great Dane dog. This assembly has improved continuity compared to the existing Boxer-derived (CanFam3.1) reference genome. Annotation of the Great Dane assembly identified 22,182 protein-coding gene models and 7,049 long noncoding RNAs, including 49 protein-coding genes not present in the CanFam3.1 reference. The Great Dane assembly spans the majority of sequence gaps in the CanFam3.1 reference and illustrates that 2,151 gaps overlap the transcription start site of a predicted protein-coding gene. Moreover, a subset of the resolved gaps, which have an 80.95% median GC content, localize to transcription start sites and recombination hotspots more often than expected by chance, suggesting the stable canine recombinational landscape has shaped genome architecture. Alignment of the Great Dane and CanFam3.1 assemblies identified 16,834 deletions and 15,621 insertions, as well as 2,665 deletions and 3,493 insertions located on secondary contigs. These structural variants are dominated by retrotransposon insertion/deletion polymorphisms and include 16,221 dimorphic canine short interspersed elements (SINECs) and 1,121 dimorphic long interspersed element-1 sequences (LINE-1_Cfs). Analysis of sequences flanking the 3′ end of LINE-1_Cfs (i.e., LINE-1_Cf 3′-transductions) suggests multiple retrotransposition-competent LINE-1_Cfs segregate among dog populations. Consistent with this conclusion, we demonstrate that a canine LINE-1_Cf element with intact open reading frames can retrotranspose its own RNA and that of a SINEC_Cf consensus sequence in cultured human cells, implicating ongoing retrotransposon activity as a driver of canine genetic variation.

Dog10K_Boxer_Tasha_1.0: A Long-Read Assembly of the Dog Reference Genome

“…Second, we identified 321 intervals encompassing 38.3 Mb of sequence based on excess depth of coverage from Illumina sequencing reads. These measures of duplication content are both less than that found in the Great Dane Zoey or CanFam3.1 assemblies [8], indicating that these duplicated sequences are not correctly resolved in the Dog10K_Boxer_Tasha_1.0 genome assembly.…”

“…Segmental duplications were defined as segments of four or more consecutive windows with an estimated copy number of at least 2.5. Comparable annotations for the CanFam3.1 assembly were obtained from [8].…”

“…The variant size distribution has clear peaks corresponding to the expected sizes of dimorphic LINEs and SINEs (Figure 2). Since LINE and SINE insertions are highly polymorphic among canines [8,40], we reasoned that the representation in the Dog10K_Boxer_Tasha_1.0 and CanFam3.1 assemblies may reflect the differential inclusion of heterozygous insertions. To assess this possibility, we identified structural variants relative to each assembly using the Tasha PacBio In the Dog10K_Boxer_Tasha_1.0 assembly, AMY2B is represented as a single copy on chromosome 6.…”

“…Furthermore, high-quality assemblies are an essential prerequisite for accurate annotation, which is required to assay the potential functional effects of detected variants. Recently, several high-quality genomes from different dogs became available [8][9][10].…”

Epigenetic clock and methylation studies in dogs

“…In total, 31,911 probes of the mammalian methylation array are aligned to loci that are proximal to 5,021 genes in the recently released long read reference assembly of the Great Dane (CanFam_GreatDane.UMICH_Zoey_3.1.100) 13 . The array has high inter-species conservation; thus, it can be extrapolated to other breeds and other mammalian species 9 .…”