Genome sequencing of 2000 canids by the Dog10K consortium advances the understanding of demography, genome function and architecture

Abstract

Background: The international Dog10K project aims to sequence and analyze several thousand canine genomes. Incorporating 20 × data from 1987 individuals, including 1611 dogs (321 breeds), 309 village dogs, 63 wolves, and four coyotes, we identify genomic variation across the canid family, setting the stage for detailed studies of domestication, behavior, morphology, disease susceptibility, and genome architecture and function.

Results: We report the analysis of > 48 M single-nucleotide, indel, and structural variants spanning the autosomes, X chromosome, and mitochondria. We discover more than 75% of variation for 239 sampled breeds. Allele sharing analysis indicates that 94.9% of breeds form monophyletic clusters and 25 major clades. German Shepherd Dogs and related breeds show the highest allele sharing with independent breeds from multiple clades. On average, each breed dog differs from the UU_Cfam_GSD_1.0 reference at 26,960 deletions and 14,034 insertions greater than 50 bp, with wolves having 14% more variants. Discovered variants include retrogene insertions from 926 parent genes. To aid functional prioritization, single-nucleotide variants were annotated with SnpEff and Zoonomia phyloP constraint scores. Constrained positions were negatively correlated with allele frequency. Finally, the utility of the Dog10K data as an imputation reference panel is assessed, generating high-confidence calls across varied genotyping platform densities including for breeds not included in the Dog10K collection.

Conclusions: We have developed a dense dataset of 1987 sequenced canids that reveals patterns of allele sharing, identifies likely functional variants, informs breed structure, and enables accurate imputation. Dog10K data are publicly available.

Keywords: Canine; Demographic history; Dog; Genetic diversity; Genomics; Mitochondrial DNA; Variation.

Fig. 1

Overview of the Dog10K collection. a Sample collection and sample filtering for (b) the varied demographic, genome function, and architecture examined in the program. QC, quality control. ROH, runs of homozygosity. OMIA, Online Mendelian Inheritance in Animals

Fig. 2

Variant distribution across the genome. a SNV density in 50-kb windows, drawn from 1987 samples. Increased SNV density is observed at the X-PAR region and the ends of most autosomes. b Median copy number (CN) for 1824 dogs reveals a large, common duplication on chr9 relative to the reference genome. c Principal component (PC) analysis separates dog and wolf samples along the first axis while axis two separates dogs from Eastern and Western Eurasia. d SNV sharing between the three categories of samples

Fig. 3

Genetic relationships between 1563 samples spanning 292 breeds. The plot is annotated for major and minor breed groups, breed purpose, and key morphological features. The fraction of haplotype sharing is indicated by the heatmap

Fig. 4

Proportion of the genome covered by ROH (F_ROH). Mean and standard deviation are plotted for breed groups and are colored as per Fig. 3. Red dashed line shows mean F_ROH for breed dogs; gray dashed line shows mean F_ROH for wolves. Breeds containing individuals with the highest and lowest F_ROH are labeled

Fig. 5

Genotype imputation accuracy of the Dog10K reference panel. a NRC rates of imputed genotypes across autosomes and the PAR segment of chromosome X. Variant sites are filtered according to GLIMPSE and IMPUTE5 imputation quality scores (INFO > 0.9). b NRC rates of imputed genotypes across the non-PAR segment of chromosome X. Variants are not filtered by imputation quality score, as imputation software does not provide scores for haploid genotypes. c NRC rates of imputed genotypes across autosomes and the PAR segment of chromosome X prior to filtering on imputation quality. d NRC rates and total number of imputed sites for each platform. Sites were filtered according to imputation quality score > 0.9 and reference MAF > 1%. e NRC rates for downsampled and full chromosome 38 reference panels for sites with reference MAF > 1%. Results show both quality and non-quality filtered sites. Data points show NRC rates for a single downsampled reference panel. Horizontal bars indicate mean NRC rates for each reference panel population size. f Number of imputed variants for downsampled and full chromosome 38 reference panels for sites with reference MAF > 1%. Results show both quality and non-quality filtered sites. Data points show the number of imputed variants for a single downsampled reference panel. Horizontal bars indicate the mean number of variants for each reference panel population size

Fig. 6

Assignment and distribution of haplogroups in breed and village dogs. a Each breed and village dog was assigned to an existing mitochondrial haplotype. b Broad geographic categories based on sample origin

Fig. 7

Structural variation detected across 1879 samples. a Boxplots of the number of deletion, duplication, insertion, and inversion variants are shown broken down by sample category. b Histograms of the size distribution of each class of detected structural variants are shown. An increase in variant count at the ~ 200 bp size bin is apparent for deletions and insertions. This corresponds to the size of SINEC elements

Fig. 8

Signatures of selection inferred with ancestry components. a Population structure inferred from Ohana for the nine selected dog groups using K = 5. b Manhattan plots for four of the five ancestral components. Top to bottom: Spitz, Mastiffs, Scenthounds, Pointers, and Spaniels. The red dotted line represents the Bonferroni cutoff and genes either overlapping or within 100 kb from a significant site are indicated at each peak. The asterisk within the Manhattan plot for the Mastiff component contains 88 candidate genes listed in Table S7. c Population tree connecting the ancestral components with colors corresponding to the ancestries are shown in the admixture plot. Each ancestral component is labeled based on the dog group(s) for which it is maximized

Fig. 9

Comparison among variant data sets. a Number of variable positions shared between major databases. b Fraction of genome spaces under constraint (5% FDR, phyloP > 2.56). c Enrichment of constrained bases in OMIA trait (blue), and disease (red) sets compared to the genome as a whole (gray). d Relationship between allele count (AC), allele frequency (AF), and phyloP score for the coding (CDS, red) and non-coding (non-CDS, green) bases in the whole genome (gray)

Fig. 10

Structural and point variants impacting CYP1A2. a Distribution of copy number variation within all breed dogs, and the major clades. b Location of SNVs across CYP1A2 inclusive of filtering or impact status. c Mammalian phyloP scores (bounded by 10, − 10). d Illustration of the region from the reference genome perspective

Fig. 11

SLC28A3 locus copy number expansion. a Location and span of the copy number element in polymorphic Grand Basset Griffon Vendéen (GBGV) and basset hound (BASS) individuals contrasted with copy number two samples from the German Shepherd Dog (GRSD) and BASS breeds. b Distribution of the median copy number count across the 1879 samples assayed