Buyer protection
Contents
Laboratory comparison of milk on day 3 after birth in four animal species: In donkey colostrum (2.29% protein, 1.10% fat), 303 proteins in the soluble fraction changed after heating at 72°C for 15 s, but none of the 38 measured fatty acids.
Laboratory analysis – no statement about an effect in humans. Donkey milk is a food.
Study at a glance
| Feature | Information |
|---|---|
| Original title | Source-Dependent Proteomic and Fatty-Acid Responses of Yak, Holstein Cow, Donkey and Bactrian Camel Colostrum to Short-Time Heat Treatment |
| Authors | Zhao J., Wang J., Liu Q. et al. |
| Journal | Foods, 15(20), 3609 |
| Year | 2026 |
| Country | China |
| Study type | Laboratory study |
| DOI | 10.3390/foods15203609 |
How the study was conducted
Milk from the third day after birth (transition stage between colostrum and mature milk) from three yaks, Holstein cows, donkeys, and Bactrian camels from farms in Qinghai Province, China, was examined. Each sample was divided into two subsamples: one remained untreated, while the other was heated in a water bath at 72°C and held for 15 s in the laboratory. The soluble protein fraction (after skimming and centrifugation) was analyzed using DIA proteomics, while the fatty acid profile was analyzed as fatty acid methyl esters (FAME) using targeted GC-MS. Differences between the animal species, changes due to heating within each species, and the agreement between the two datasets were evaluated.
Key figures
| Value | Meaning |
|---|---|
| 3 | Animals per species (12 samples in total, 24 subsamples) |
| 72°C / 15 s | Only heating condition tested in the laboratory |
| 2.29% | Protein content of donkey milk (day 3) before heating |
| 1.10% | Fat content of donkey milk (day 3) before heating |
| 6.59 | pH value of donkey milk (day 3) |
| 2583 | Identified protein groups across all 24 samples |
| 303 | Proteins in donkey milk changed by heating |
| 65.0% | Proportion of these donkey proteins with lower abundance after heating |
The results
In the untreated day 3 milk, donkey milk contained 2.29% protein and 1.10% fat, which was significantly less than yak (5.61% / 4.24%), Holstein cow (5.05% / 3.97%), and Bactrian camel (6.28% / 3.96%). The pH value for all four species was between 6.36 and 6.59, with no significant differences.
Proteomics identified 2583 protein groups. The samples separated clearly by animal species; heated and untreated subsamples of the same species clustered closely together. Before heating, there were 1220 different proteins between Holstein cow and donkey, 837 between donkey and Bactrian camel, and 1136 between yak and donkey.
After heating at 72°C for 15 s, 303 proteins in the soluble fraction of donkey milk were changed, 65.0% of which had lower abundance. In yak milk, there were 387, in Bactrian camel milk 276, and in Holstein cow milk 146. The authors emphasize that lower abundance in the soluble fraction does not directly prove degradation, but may also reflect aggregation or redistribution.
Most heat-responsive proteins were species-specific (donkey: 91). Only six proteins (ASL, C8B, FGB, FGG, RPLP2, TPP1) responded in all four species. Gelsolin (GSN) decreased in all species, most strongly in donkey milk; CTSA showed a particularly strong decrease in donkey milk.
In the fatty acid profile, donkey milk differed significantly from the other species, including relatively higher standardized proportions of methyl linoleate. Heating did not significantly change any of the 38 measured FAMEs in donkey milk; in yak milk, there were 14, and in Holstein cow milk 6.
Proteome and fatty acid data showed strong agreement at the species level (Mantel r = 0.943). The response to heating was less parallel in both datasets; the mean proteome response (RMS) in donkey milk was 0.966, while the fatty acid response was 0.105.
According to Zhao et al. (2026), the animal origin primarily determines the molecular composition of early milk; heating at 72°C for 15 s caused smaller, but origin-dependent changes, primarily in the soluble protein fraction and to a lesser extent in the fatty acid profile. The authors see this as a basis for further studies on bioactive components and the processing of colostrum; effects in humans were not investigated.
Critical assessment
- Limitation – Very small sample size: Only three animals per species were examined. The authors therefore explicitly describe the results as exploratory and origin-related, not as general characteristics of the species.
- Limitation – Origin and housing mixed: The animals came from different farms, altitudes (2100–3473 m), and housing systems with different feeding. Differences between the species may therefore also be due to the environment and feeding.
- Note – Laboratory question, no consumption recommendation: Only one condition (72°C, 15 s) was tested on 10-mL samples in a water bath; other temperatures and times were not examined.
- Note – Only soluble protein fraction: The protein fraction was analyzed after skimming and centrifugation; aggregates, cream, and membrane proteins are missing. The content or activity of individual proteins such as lysozyme was not determined separately.
- Note – Minor inconsistencies: In the methods section, the protein-FAME heatmap is described as a Pearson correlation, while the figure legend states Spearman. The number of interacting proteins is given as 48 and 73, respectively (different level of counting).
- Strength – Paired design, two methods: Heated and untreated subsamples each came from the same animal, all species were treated identically, and proteome and fatty acid analyses were combined. The temperature was monitored with a calibrated probe inside the sample.
Funding, conflicts of interest, and level of evidence
Science and Technology Plan Project of Haixi Prefecture, Qinghai Province, China (funding code 2025-JC-Q04).
The authors declare that there are no conflicts of interest.
Level of evidence: Basic research (laboratory) – describes the properties of milk, no statement about effects in humans.
Our assessment
This is a laboratory comparison of colostrum processing with a very small sample size. Statements about donkey milk products on the market or recommendations for heating milk cannot be derived from it.
Open research question: How do other temperatures and holding times change the protein and fatty acid profile of early donkey milk – and do the differences persist with larger animal numbers?
You can find all measurements, interactive graphics, and the full critical assessment on our research page: Study 98: Donkey colostrum in a species comparison: 72°C for 15 s. Further studies in the research archive.
Zhao J., Wang J., Liu Q., Zhuang J., Chen C., Zhang Y. (2026). Source-Dependent Proteomic and Fatty-Acid Responses of Yak, Holstein Cow, Donkey and Bactrian Camel Colostrum to Short-Time Heat Treatment. Foods, 15(20), 3609. DOI: 10.3390/foods15203609
Summarized and reviewed by the DON DONKEY'S editorial team based on the original publication · As of: October 2026. This summary does not replace the original work.
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.
