Pekingese
26 Pekingeses in the atlas. Every number on this page has a source.
26 Pekingeses in the Sniff Atlas. Population-genetic snapshot, Mendelian carrier frequencies from Donner 2023, and the data substrate's release version, sample sizes, and evidence tier on every claim.
In the atlas, the Pekingese clusters consistently as Pekingese (100% of the 26 dogs here). At the trait loci, FGF4_retrogene_CFA18 runs lower than average (0% here vs 77%); SMAD2 runs lower than average (4% here vs 74%). Dogs here sit in a relatively sparse region of the atlas, fewer close neighbors than typical.
Only 26 dogs of this breed in the atlas, modestly sampled.
Closest genetic neighbors in the atlas: Shih Tzu, Tibetan Spaniel, Tibetan Terrier, Pug, and Chihuahua.
What the genome says about Pekingese
Computed from the 18,477 research dogs in the Atlas.
- Shih Tzu2.14
- Tibetan Spaniel10.76
- Tibetan Terrier11.34
- Pug16.25
- Chihuahua17.48
Frequency of the alternate allele in this breed at each locus's representative SNP.
| IGF1 | 100% |
| HMGA2 | 8% |
| SMAD2 | 4% |
| LCORL | 85% |
| STC2 | 94% |
| ADAMTS17 | 85% |
| FGF4·CFA18 | 0% |
| FGF4·CFA12 | 27% |
| RSPO2 | 81% |
| FGF5 | 54% |
| KRT71 | 100% |
| MC1R | 96% |
| MSRB3 | 70% |
| BMP3 | 100% |
| SMOC2 | 94% |
What does the genome say about how a Pekingese looks?
Pekingeses look the way they do because of a small set of fixed and near-fixed morphology genes that, taken together, define the visible breed. Each translation below pairs the gene with the trait an owner actually sees, the breed's allele frequency at that locus, and a one-clause causal phrase.
Size and build
IGF1 is near-fixed at 100% for the small-body allele, which keeps the breed compact relative to its working-line ancestors.
HMGA2 is at 8%, leaving most of the size signal to other loci in the panel.
SMAD2 is at 4%, leaving the height signal mostly to other size genes.
LCORL sits at 85% at the NCAPG/LCORL height locus on chromosome 3.
STC2 is near-fixed at 94%, modulating growth-axis signaling toward the breed's body-size set point.
ADAMTS17 is at 85%, near-fixed for the size variant.
Leg length
The FGF4 retrogene on chromosome 18 is at 0%, the chromosome-18 leg-length variant, which keeps the breed short-legged like Corgis and Dachshunds.
The FGF4 retrogene on chromosome 12 is at 27%, leaving most of this breed clear of the chondrodystrophic intervertebral disc disease risk.
Coat type, length, and color
RSPO2 sits at 81% for the furnishings variant. Furnishings (the eyebrow-and-mustache pattern seen in Schnauzers and Wheaten Terriers) vary across the population at this intermediate frequency, and visible expression depends on the specific allele combination each dog carries.
FGF5 sits at 54% for the long-coat variant. Coat length is influenced by other loci as well, so intermediate FGF5 frequencies do not always correspond to intermediate visible coat lengths.
KRT71 is near-fixed at 100% for the wavy/curly variant. Coat curl phenotype varies across breeds at this fixation depending on modifier loci, and visible expression is not always curled even when the locus is fixed.
MC1R is at 96% at the representative SNP. MC1R controls the switch between red-to-gold and black-to-brown pigment, with the e/e homozygous genotype producing the gold-to-red spectrum by blocking eumelanin (black and brown pigment).
Ears
MSRB3 sits at 70% for the drop-ear allele, which is why ear set varies across the breed.
Skull shape
BMP3 is at 100%, contributing to the breed's brachycephalic skull shape.
SMOC2 is at 94%, the major locus contributing to the breed's brachycephalic face shape.
What genetic diseases do Pekingeses carry?
From a panel of 250 Mendelian-disease variants screened in 1,054,293 dogs (Donner et al. 2023), Pekingeses carry 6 of them at observable frequency. Carrier frequency is not clinical risk. Most recessive variants require two copies for disease expression; many dominant variants show incomplete penetrance. Read this as a population fingerprint of what's in the gene pool, not a per-dog prediction.
Where every number on this page came from.
This page draws on three primary data sources. Carrier frequencies for the Mendelian section come from Donner et al. 2023 (CC-BY-4.0). We grade these data at evidence Limited because the cohort is a direct-to-consumer ascertainment, which biases toward owners who chose to test their dogs. The panel also uses tag-SNP proxies for some variants rather than direct causal-variant assays. Limited is a study-design grade, not a quality grade: the Donner cohort is the largest open canine-genotype dataset in existence and we are grateful for it. We rate the confounding MEDIUM.
Population-genetic dimensions (heterozygosity, intra-breed PCA distance, nearest neighbors, trait-locus frequencies) come from CanVAS (Brundage 2026), harmonized through the Sniff Atlas. The exact release date and verification commit are pinned at the bottom of the page so a researcher can trace a number back to a specific snapshot. The disease-gene-variant graph comes from OMIA (Online Mendelian Inheritance in Animals; Nicholas, Tammen, and the Sydney Informatics Hub at the Sydney School of Veterinary Science, The University of Sydney; retrieved April 2026, DOI 10.25910/2AMR-PV70).
What this page does not yet have. Inheritance modes and per-disease penetrance evidence from Donner 2023 are now in the structured data for every variant the panel covers. Mondo, OMIM, Ensembl, and HGNC cross-references on gene pages remain pending — they arrive in December 2026 alongside the imputed 9.67M-variant CanVAS dataset via the OMIA SQL dump absorption. Until then, gene IDs carry NCBI Gene and OMIA phene URLs only; the wider human-homolog and disease-ontology cross-reference set fills in with that release.
How to cite this page. The computed dimensions on this page are derived from the open Sniff Atlas v1.0.1 (Gehring 2026, doi:10.5281/zenodo.20566358, CC-BY 4.0). Full citation formats including BibTeX, RIS, and CITATION.cff at sniff.world/cite.
We have 26 pekingeses. We do not have yours.
Every pekingese added sharpens the breed's genetic neighborhood. Enrollment is free. The data stays open. The star is permanent.
- Donner J, Anderson H, Davison S, et al. (2023). Frequency and distribution of 152 genetic disease variants in over 1,000,000 mixed-breed and purebred dogs. PLOS Genetics 19(2):e1010651. doi:10.1371/journal.pgen.1010651
- Brundage J, et al. (2026). CanVAS: a harmonized canine variant atlas. bioRxiv. doi:10.64898/2026.04.13.718238
- Nicholas, F.W., Tammen, I., & Sydney Informatics Hub. (2026). Online Mendelian Inheritance in Animals (OMIA) [dataset]. The University of Sydney. https://omia.org. doi:10.25910/2AMR-PV70 (retrieved April 2026).