Donkey milk study: aroma marker for milk powder (Gou 2023)

61 volatile compounds were detected in cow, donkey, camel, and horse milk powder from Xinjiang. Eight of these reliably distinguished donkey and cow milk powder, including 3-hydroxy-2-butanone (0.55 vs. 5.60%).

The study in brief

• Original title: Discriminant Analysis of Aroma Differences between Cow Milk Powder and Special Milk Powder (Donkey, Camel, and Horse Milk Powder) in Xinjiang Based on GC-IMS and Multivariate Statistical Methods
• Authors: Gou Y., Han Y., Li J. et al.
• Journal: Foods, 12(21), 4036 (2023)
• Country: China · Study type: Laboratory study

How it was investigated

Cow, donkey, camel, and horse milk from various regions in Xinjiang (donkey milk from Hetian) was collected in 2022 in three batches (February, June, October), with three samples each, and stored frozen. To replicate common commercial processes, cow and camel milk were spray-dried, while donkey and horse milk were freeze-dried (residual moisture 5–6% each). The volatile compounds were determined using gas chromatography-ion mobility spectrometry (GC-IMS) (2 g of powder, 80 °C, measured in triplicate) and reported as a relative proportion of the total signal.

Key figures

• 4 – Milk powders: cow, donkey, camel, and horse from Xinjiang
• 3 × 3 – Batches (February, June, October 2022) with three samples each per milk type
• 108 – Detected peak signals
• 61 – Identified signals (incl. dimer signals)
• 15 – Candidates (VIP > 1) for cow vs. donkey milk powder
• 8 – Confirmed markers (AUC 0.95–1) for cow vs. donkey milk powder

Results

A total of 108 signals were detected, 61 of which were identified. The four powders were similar in the types of compounds present but differed in their quantities; ketones, aldehydes, and alcohols formed the main groups. The freeze-dried donkey and horse milk powders showed more signals than the spray-dried cow and camel milk powders.

In donkey milk powder, the ketone content was 10.83% compared to 23.41% in cow milk powder, and the aldehyde content was 22.21% compared to 16.65%. Hexanal reached the highest individual value of all four powders in donkey milk powder at 9.02%.

Typical milk aroma compounds such as 3-hydroxy-2-butanone (0.55 vs. 5.60%), 2-butanone (1.30 vs. 3.10%), and 2-heptanone (1.20 vs. 3.51%) were significantly lower in donkey milk powder than in cow milk powder; (E)-2-pentenal (1.30 vs. 0.75%) and 2-methylbutanal (0.55 vs. 0.10%) were higher.

Limitations of the study

• Animal species and drying method mixed: Donkey and horse milk were freeze-dried, while cow and camel milk were spray-dried. Differences between donkey and cow milk powder therefore cannot be reliably attributed to the animal species; the authors themselves attribute the higher ketone content, for example, to spray drying.
• Origin and number of animals unclear: Three batches of three samples each were collected for each milk type; the number and breed of the animals are not stated, and each milk type comes from a different region in Xinjiang. Whether the markers also apply to milk from other origins was not tested.
• Contradictory counts: The text mentions 22 dimer signals, while Table 2 marks 19; the stated compound classes (13 alcohols, 7 esters, 12 aldehydes, 7 ketones, 1 ether, 1 acid, 2 furans) do not add up to 61. Two aldehydes are listed under the alcohols in the table. For cow vs. horse, 23 candidates are mentioned, but 22 are counted.
• Marker numbers and designations: The abstract and Section 3.4 mention 8, 6, and 19 markers; the conclusion states "(E)-2-pentenal and eight more", "3-methyl-1-butanol and six more", and "2-nonanone and 19 more", respectively, thus one more each. Section 3.5 describes the comparison in Figure 4c as cow vs. camel instead of cow vs. horse.
• Relative proportions without reference: The contents are given as percentages, but the compound groups add up to only about 55–65% per powder; the reference value is not explained. No sensory test was conducted, and statements on odor impressions are based on literature.

Level of evidence: Basic research (laboratory) – describes the properties of the milk, no statement on effects in humans.

Our assessment

The study describes aroma differences between laboratory powders from Xinjiang and proposes markers for authenticity testing. Since donkey and cow milk were dried differently, the values describe the combination of milk type and process; no conclusions about individual commercial products can be drawn from this.

Source: Gou Y., Han Y., Li J., Niu X., Ma G., Xu Q. (2023). Discriminant Analysis of Aroma Differences between Cow Milk Powder and Special Milk Powder (Donkey, Camel, and Horse Milk Powder) in Xinjiang Based on GC-IMS and Multivariate Statistical Methods. Foods. DOI: 10.3390/foods12214036 · Summarized by the DON DONKEY'S editorial team, as of October 2026.

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