Figure 1.
Phenotype of the inherited disproportionate dwarfism.
(A) Photograph of a female affected Labrador Retriever. Notice the relatively long body in relation to the length of the legs. (B) Affected mother (left) with her non-affected daughter (right). The affected dog has shorter forelegs. (C) Photograph of three littermates. The affected male has short legs and different proportions compared to his non-affected sister. (D) Radiograph of the left front limb of a male control Labrador Retriever and (E) his affected full brother at 12 months of age. In the affected dog ulna and radius are 2.7 cm and 1.9 cm shorter and slightly more bent than in the control, respectively. The diameter of the diaphysis of the long bones is not affected (Table S1).
Figure 2.
Pedigree of 33 Labrador Retrievers with mild disproportionate dwarfism.
Affected dogs are shown as filled symbols together with their laboratory identifiers. One dog with severely shortened legs and deformed paws that did not carry the disease-associated haplotype is highlighted in red (see also Figure S1). Three dogs that could not unambiguously be classified are indicated with question marks. Please note that due to space restrictions many non-affected littermates of the affected dogs are not shown in this pedigree. All affected dogs in this pedigree can be traced to a common ancestor in their paternal and maternal ancestry, respectively (arrow). This male dog was born in 1966 and may represent the founder of SD2.
Figure 3.
Mapping of SD2 in Labrador Retrievers.
(A) A genome-wide association study using 23 cases and 37 controls indicates a strong signal with multiple associated SNPs on CFA 12. (B) The detailed view of CFA 12 delineates an associated interval of ∼18 Mb at the beginning of the chromosome. (C) Homozygosity mapping. Each horizontal bar corresponds to one of the 23 analyzed cases. Homozygous regions with shared alleles are shown in blue. A shared homozygous interval of ∼4 Mb in 22 of the 23 cases delineates the exact boundaries of the critical interval from 467,795 bp to 4,906,914 bp (CanFam 3 assembly). The 23rd case did not carry the associated haplotype and was also phenotypically slightly distinct from the other 22 cases (Figure S1).
Table 1.
Variants detected by whole genome re-sequencing of an affected Labrador Retriever.
Table 2.
Five non-synonymous variants in the critical interval of an affected Labrador Retriever that were absent from 13 other dog genomes.
Table 3.
Association of non-synonymous variants with the dwarfism phenotype.
Figure 4.
The p.R48P variant is located in a region of moderate sequence conservation between the N-terminal signal peptide (amino acids 1–27) and the laminin G-like domain (amino acids 57–228). The sequences correspond to the following accessions: XP_538855.2 (canis familiaris), NP_542411.2 (homo sapiens), NP_001039664.1 (bos taurus), NP_034056.1 (mus musculus), NP_997693.1 (rattus norvegicus), NP_001073461.1 (danio rerio).
Figure 5.
Genotype-phenotype correlation.
The figure displays the shoulder heights of male and female Labrador Retrievers with the three different genotypes at the COL11A2:c.143G>C variant. The red lines indicate the averages of each group. The size distributions confirm that SD2 is inherited as recessive trait. The female outlier indicated in red was phenotyped as affected by disproportionate dwarfism, but did not carry the COL11A2 variant (see also Figure 2 and Figure S1). Please note that there is a considerable overlap in the shoulder height distributions of SD2 cases and controls, reflecting the fact that height is a complex polygenic trait in dogs. We used shoulder height as a proxy for the SD2 phenotype in this analysis as it was the only quantitative phenotype that was available to us on a sufficiently large number of animals. However, dogs with the C/C phenotype typically had different proportions from dogs with the other two genotypes, even if they had a normal shoulder height.