Bacterial microbiota diversity in sheep milk during refrigerated storage under different processing conditions

Abstract

Sheep milk, while nutritionally dense due to high calcium and B vitamins, spoils rapidly due to its diverse microbiota. Contamination risks arise from environmental Listeria or biofilm-forming microbes during milking and handling. This study employed 16 S rRNA gene sequencing to characterize microbial dynamics in sheep milk treated with ultrasound, pasteurization, and ultrasound/pasteurization during refrigerated storage. Analysis of 71 samples yielded 3,301,066 high-quality reads, revealing significant shifts in microbial community composition and diversity. Results revealed that Lactococcus was the primary bacteria causing deterioration in untreated sheep milk stored at 4 °C. In treated samples, the significant growth of Lactococcus and Hafniella obesoides were identified as the main cause of spoilage during 4 °C storage. This research provides valuable insights into bacterial influences on sheep milk under different processing conditions, offering potential for reducing milk losses during transportation, processing, and lowering costs in the dairy industry.

Keywords: 16S rDNA; Pasteurization; Spoilage; Ultrasonic sterilization.

Fig. 1

Principal component analysis based on the distance matrix calculated using the Bray–Curtis distance matrix for sheep milk microbiota. The sheep milk microbiota compartments cause the largest source of variance amongst bacterial communities of ultrasonic sterilization (T1), pasteurization (T2), combined ultrasonic-pasteurization (T3), and Control group.

Fig. 2

Composition and structure of bacterial community composition across different treatments and storage times. (A) represent phylum level of bacteria. (B) represent genus level of bacteria. The top 10 relative abundances are shown, and the remaining abundances are indicated as ‘Others’.

Fig. 3

LEfSe analysis at multiple taxonomic levels comparing the sheep milk bacterial community across different treatments and storage times. Histogram of the LDA scores greater than 3.5, computed for groups with differential abundance among the bacterial communities at 0 d (A) and 8 or 10 d (B). The length of each bar represents the contribution of species with significant differences in abundance.

Fig. 4

PICRUSt2-predicted COG functions from 16 S rRNA data. Core cellular processes (J/A/K) dominate across treatments (83–87%), with < 6% combined contribution from secondary metabolism (Q) and uncharacterized functions (S). Error bars show SEM (n = 3).