Buyer protection
Contents
27 donkeys were fed corn straw, wheat hulls, or wheat straw for six weeks. 43 volatile compounds were detected in the milk; 23 of them statistically distinguished the feed groups.
Study on animals and their milk – no statement about an effect in humans. Donkey milk is a food.
Study at a glance
| Feature | Information |
|---|---|
| Original title | Characterization and discrimination of milk volatiles from donkey fed different roughages using GC-IMS |
| Authors | Du X., Zhao Y., Ding L. et al. |
| Journal | Frontiers in Nutrition, Volume 12, Article 1652665 |
| Year | 2025 |
| Country | China |
| Study type | Animal study |
| DOI | 10.3389/fnut.2025.1652665 |
How the study was conducted
27 healthy female donkeys (5 years old, 60 ± 15 days in lactation, 286 ± 25 kg) were randomly assigned to three groups of 9 animals each. The ration consisted of 80% roughage – corn straw (G1), wheat hulls (G2), or wheat straw (G3) –, the rest (corn, soy, premix, salt) was identical. The experiment lasted 6 weeks (2 weeks pre-lactation, 4 weeks main lactation); at the end, a milk sample was taken from each donkey and stored at −80 °C. The volatile compounds (VOCs) were determined using gas chromatography-ion mobility spectrometry (GC-IMS) and evaluated with relative odor activity values (ROAV) and OPLS-DA multivariate analysis.
The most important figures
| Value | Meaning |
|---|---|
| 27 donkeys | Female donkeys, divided into three feed groups of 9 animals each |
| 6 weeks | Duration of the experiment (2 weeks pre-lactation, 4 weeks main lactation) |
| 80.00 % | Roughage content of the ration (corn straw, wheat hulls, or wheat straw) |
| 43 | volatile compounds (VOCs) detected in donkey milk |
| 27.91 % | Proportion of esters among the detected VOCs (12 compounds) |
| 25.58 % | Proportion of ketones among the detected VOCs (11 compounds) |
| 18.60 % | Proportion of aldehydes among the detected VOCs (8 compounds) |
| 13 | characteristic aroma compounds with a relative odor activity value (ROAV) above 1 |
The results
In the milk of all groups, the authors detected 43 volatile compounds: 12 esters (27.91 %), 11 ketones (25.58 %), 8 aldehydes (18.60 %), 7 alcohols (16.28 %), one pyrrole, alkene, and one acid (each 2.33 %) as well as 2 unidentified compounds (4.65 %). Ketones and esters were the most common.
Milk from the corn straw group (G1) contained significantly more esters (P < 0.05) and aldehydes (P < 0.001) than the two wheat groups; alcohols were higher in G1 than in G2. Alkenes (1-phellandrene), on the other hand, were highest in the wheat straw group (G3) and lowest in G1 (P < 0.001); the pyrrole was lower in G3.
13 compounds had a relative odor activity value (ROAV) above 1 and are considered characteristic aroma compounds in the study: five aldehydes, four esters, and four ketones. The largest contributions were made by methyl thiocyanate, ethyl acetate, and hexanal.
The OPLS-DA analysis separated the three groups; 23 compounds with VIP > 1 served as possible markers (9 esters, 8 aldehydes, 2 alcohols and ketones each, 1 alkene and pyrrole each). Thus, (E)-hept-2-enal was 278.29 in G1, 55.17 in G2, and 58.62 in G3 (standardized intensity); ethyl 2-methylpropanoate (monomer) was 1324.89, 382.52, and 311.02.
Some compounds were higher in the wheat groups, such as allylisothiocyanate (G1 188.22; G2 387.98; G3 370.23) and amyl acetate dimer (126.06; 160.75; 202.46).
The authors attribute the higher ester and aldehyde values in G1 to the higher crude fat content of corn straw (1.90 % of DM compared to 1.27 % and 1.02 %). This explanation was not directly tested in the study.
The authors conclude that the volatile compounds of donkey milk – especially esters and aldehydes – are influenced primarily by the type of roughage, and see this as a theoretical basis for controlling the aroma profile of donkey milk through roughage.
Critical assessment
- Limitation – No tasting: It was only measured instrumentally; there was no sensory panel. The odor descriptions (e.g., grassy, fruity) come from the literature, not from their own tests.
- Limitation – Only relative values: According to the methods section, peak areas do not allow for the calculation of quantities or concentrations. All group comparisons are based on normalized, relative signal strengths.
- Note – Errors in Table 3: Table 3 is titled "donkey meat" instead of milk. Several values are obviously incorrectly entered (e.g., "723.60 3.20.54b", "1024.40 2448.26c"), and some significance letters contradict the main text (e.g., hexanal monomer: highest value in G1 marked with "c").
- Note – Questionable assignments: Some database assignments are chemically questionable: At least one assigned compound does not appear plausible in milk; 2-propanethiol is classified as an alcohol, allylisothiocyanate as an ester. The discussion also refers to the rumen, although donkeys are not ruminants.
- Note – Limited sample size: 9 animals per group, a single milk sample per animal at the end of the experiment, one site in China. The conclusion mentions seven substance classes, while the results section mentions eight.
- Strength – Controlled feeding design: Randomized groups, identical housing and feeding times, identical ration except for the roughage, six-week duration, and ethics approval; the roughage analyses are fully provided.
Funding, conflicts of interest, and level of evidence
Funded by the National Key R&D Program of China (2022YFD1600103, 2023YFD1302004), the Shandong Provincial Natural Science Foundation (ZR2022QC130), the Donkey Industry Innovation Team of the Modern Agricultural Technology System of Shandong Province (SDAIT-27), the Support Program for Innovation and Technology in Universities of Shandong Province (2023KJ354), and the Livestock and Poultry Breeds Project of the Chinese Ministry of Agriculture (19211162).
The authors state that there were no commercial or financial relationships that could be interpreted as a potential conflict of interest.
Level of evidence: Animal study – describes measured values in the studied animals and their milk, no statement about effects in humans.
Our assessment
The study measures aroma compounds instrumentally and relatively; whether people can taste or smell the differences was not investigated.
Open research question: Do consumers actually perceive the feed-related differences in the aroma profile of donkey milk in a tasting?
You can find all measured values, interactive graphics, and the full critical assessment on our research page: Study 93: Roughage and aroma compounds in donkey milk. Further studies in the research archive.
Du X., Zhao Y., Ding L., Huang F., Ren W., Ma Z., Zhu M., Liu G., Khan M. Z., Wang C., Zhou M., Li M. (2025). Characterization and discrimination of milk volatiles from donkey fed different roughages using GC-IMS. Frontiers in Nutrition, Volume 12, Article 1652665. DOI: 10.3389/fnut.2025.1652665
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 study summary. The study was conducted independently of DON DONKEY'S® and does not refer to our products. Our donkey milk is a food.
