Donkey Milk Study: Genetic Markers for Milk Yield (Fang 2025)

In 112 Xinjiang donkey mares, 236 genetic variants were found that were statistically associated with milk yield, fat, protein, or lactose content. In a test group of 82 animals, the daily milk yield differed by genotype for eight variants (e.g., 2.67 vs. 1.57 kg).

The study in brief

• Original title: Unveiling Genetic Markers for Milk Yield in Xinjiang Donkeys: A Genome-Wide Association Study and Kompetitive Allele-Specific PCR-Based Approach
• Authors: Fang C., Farnir F., Liu L. et al.
• Journal: International Journal of Molecular Sciences, 26(7), 2961 (2025)
• Country: China · Study type: Animal study

How it was investigated

112 Xinjiang donkey mares (birth years 2013 and 2014, 5th or 6th lactation) from two farms in Yopurga County (Kashgar Region, China) were phenotyped; average monthly milk yield (AY = daily yield × 30), fat, protein, and lactose content were derived from 4704 sample milk data. The genome was fully sequenced from blood samples with approximately 10-fold coverage. After quality filtering, around 12 million single-base variants (SNPs) entered a genome-wide association analysis using mixed linear models, calculated with two analysis programs; the significance threshold was p < 1 × 10⁻⁶.

Important figures

• 112 – Xinjiang donkey mares from two farms in the GWAS
• 4704 – evaluated sample milk data sets
• 38.51 kg – average monthly milk yield (SD 14.87; 13.60–72.75 kg)
• 1.02 % – average fat content (0.12–4.52 %)
• 1.44 % – average protein content (0.99–1.94 %)
• 6.46 % – average lactose content (4.47–6.96 %)

Results

The 112 donkey mares achieved an average of 38.51 kg of milk per month at 1.02 % fat, 1.44 % protein, and 6.46 % lactose; fat content (SD 0.93) and monthly yield (SD 14.87) varied widely among the animals.

After quality control, 12,039,909 SNPs were available, predominantly in intergenic regions (43.20 %) and in introns (39.36 %). Principal component analysis showed no clear substructure between the two farms (PC1 7.48 %, PC2 6.21 %).

The two analysis programs found 222 and 228 significant SNPs, respectively; a total of 236 SNPs were associated with one of the four traits, of which 11 were consistent for milk yield, 95 for fat, 5 for protein, and 103 for lactose. No gene was associated with three or four traits at the same time.

Limitations of the study

• Small sample size for a GWAS: 112 animals from two farms and one breed are few for a genome-wide analysis; the authors themselves call for larger validation studies. Statistical associations do not prove causal genetic variants, and the functional interpretations of the candidate genes are based on literature from other species.
• Contradictions in the markers: The abstract mentions five significant markers and assigns position g.9133371 to the TOP1MT gene; according to Table 2, it is in NOTCH1, TOP1MT is at position 38323301. The results section, however, mentions eight significant SNPs.
• Conspicuous values in Table 4: For the two PLK1 variants, the values for TC (21 animals) and GC (21 animals), as well as for TT (40) and CC (40), are identical. The genotype numbers result in only 70–80 animals per SNP instead of 82, without explaining the missing values.
• Milk yields difficult to compare: The average monthly yield of 38.51 kg corresponds calculated to approximately 1.28 kg per day, while the genotype groups of the test group are at 1.57–2.75 kg per day. How daily and total milk yield were collected in the test group is not described. The discussion also mentions 38.52 instead of 38.51 kg.
• Methodological ambiguities: According to the methods section, no individual additive random effects were included, but the stated model includes a polygenic random effect. The threshold 1 × 10⁻⁶ is described as Bonferroni-corrected and at the same time as "suggestive"; the number of genes per trait is stated inconsistently (e.g., 18 and 50 for milk yield).

Level of evidence: Animal study – describes measured values in the studied animals and their milk, no statement on effects in humans.

Our assessment

The study concerns the genetics of milk yield in donkey mares in China and provides clues for animal breeding. Statements about the quality of milk for humans or about commercial products cannot be derived from it.

Source: Fang C., Farnir F., Liu L., Xiao H. (2025). Unveiling Genetic Markers for Milk Yield in Xinjiang Donkeys: A Genome-Wide Association Study and Kompetitive Allele-Specific PCR-Based Approach. International Journal of Molecular Sciences. DOI: 10.3390/ijms26072961 · Summarized by the DON DONKEY'S editorial team, as of October 2026.

End of the study summary. The study was conducted independently by DON DONKEY'S® and does not refer to our products. Our donkey milk is a food.