The correlation between TBC and CC (r = 0.5) in this study corroborated previous studies that reported a moderate to high correlation [ 13, 15, 56, 69]. This indicates that TBC could be used as a single microbiological hygienic indicator for total aerobic bacteria present in raw milk [ 15]. Reta MA, Bereda TW, Alemu AN. Bacterial contaminations of raw cow’s milk consumed at Jigjiga City of Somali regional state, eastern Ethiopia. Int J Food Contam. 2016;3:4. https://doi.org/10.1186/s40550-016-0027-5. The samples were labelled with the farm code, sample type and date of collection with a permanent marker and transported to Asella Regional Veterinary Laboratory using an icebox for bacteriological analysis. The samples were stored at 4 °C and cultured the next morning (within 15–20 hours). Informed consent was obtained from each dairy farmer before sampling, and they were asked to rule out whether their cows had been given antibiotic treatment recently. Questionnaire survey

This study aimed to investigate the temporal dynamics of the microbiota in the bovine udder. The ten cows used for this purpose were divided in two groups based on the level of SCC before the first sampling. Five of the cows (L1–L5) had a stable low SCC (< 100,000 SCC/mL) on the three days leading up to the first sampling, while the remaining five cows (H1–H5) had a higher SCC (> 100,000 SCC/mL) in the same period. Quarter milk samples were collected from all ten cows at six samplings during a period of five months (January to May). Of the 240 quarter samples that were collected, six were missing during collection and were not included in the analysis. To study the microbial composition of the 234 remaining samples, amplicon sequencing of the 16S rRNA genes were performed for all the samples. The average depth of sequencing was 49,093 sequences per sample before filtering and 18,880 sequences per sample after filtering. In total 9132 high quality SVs were obtained from 234 samples. 14 samples were filtered out of the analysis because they did not pass the quality filtering of the Dada2 pipeline used to analyse the 16S data after sequencing. Of the 9,132 high quality SVs, 6962 SVs were used for taxonomy search, and 553 SVs were successfully assigned to family level. 61 quarter samples were classified as having an IMI based on definition “A” from Dohoo et al. [ 23]. This definition states that a quarter sample where > 10 colonies are cultured per 0.1 mL is defined as having an IMI. Eighty-nine percent of these were from group H and 11% were from group L. A limit of 100,000 SCC/mL was selected to classify the quarter samples as high or low SCC during samplings. 30 quarter samples had high SCC, 204 low SCC, and 6 samples had no recorded SCC. Of the 30 samples with a recorded high SCC, 93% were from group H and 7% from group L. Other additional data about the health, parturition, and recorded mastitis were retrieved from the Norwegian Cattle Health Recording System (Additional file 1). No cows were recorded to have mastitis during the sampling period, while two cows (L1 and H2) were treated for mastitis caused by Streptococcus dysgalactiae and Streptococcus uberis, respectively, after this period. The samplings occurred between 19 and 193 days in milk for the 10 cows (Additional file 2). This period encompasses the early and mid-lactation stages. Five of the cows (H4, H5, L3, L4, L5) were in their first parturition, while the remaining five (H1, H2, H3, L1, L2) were in their second parturition. Diversity analysis of the quarter samplesOlatoye O, Amosun A, Ogbu U, Okunlade Y. Bulk tank somatic cell count and associated microbial quality of milk from selected dairy cattle herds in Oyo state, Nigeria. Ital J Food Saf. 2018;7:95–100. Once it is identified, it is important to identify the severity of mastitis as this is crucial in determining what treatment to give. Porcellato D, Meisal R, Bombelli A, Narvhus JA. A core microbiota dominates a rich microbial diversity in the bovine udder and may indicate presence of dysbiosis. Sci Rep. 2020;10:1–14.

Bava L, Zucali M, Sandrucci A, Brasca M, Vanoni L, Zanini L, Tamburini A. Effect of cleaning procedure and hygienic condition of milking equipment on bacterial count of bulk tank milk. J Dairy Res. 2011;78:211–9. https://doi.org/10.1017/S002202991100001X. Observation and palpation of the udder is essential but tends to detect only later or more severe mastitis cases. Milk samples collected from cows with SCC < 100,000/mL and without IMI showed a stable eubiotic bacterial community throughout samplings as evidenced by higher species abundance and species richness estimates, which is commonly observed for milk samples obtained from healthy udders [ 4, 15, 17]. Interestingly, it is possible to observe that samples classified as “H”, IMI positive, and SCC > 100,000/mL are remarkably enriched by Gram-positive genera (i.e., Staphylococcus, Clostridium, Corynebacterium, Jeotgalicoccus, and Streptococcus), whereas milk samples from healthy quarters display a balance between Gram-positive and Gram-negative taxa. It is worth mentioning that Gram-negative genera such as Pseudomonas and Acinetobacter occur mainly in milk samples collected during the pasture season [ 1]. Although it is difficult to establish whether the low diversity found in milk samples from H-cows is a cause or a consequence of ongoing subclinical mastitis, our results reinforce the association of mastitis with an imbalanced udder microbiota [ 2]. Staphylococcus aureus is a major cause of bovine mastitis [ 20], resulting in large economic losses due to reduced milk production and quality [ 21]. S. aureus is frequently isolated from cows with mastitis [ 22, 23, 24, 25] and bulk milk [ 26, 27, 28] in Ethiopia. The presence of S. aureus in bulk milk samples indicates that infected cows are present in a herd [ 9]. The public health concern due to staphylococcal food poisoning is potentially serious in Ethiopia due to widespread raw milk consumption habits [ 29, 30, 31]. S. aureus produces enterotoxins that cause food poisoning when ingested in contaminated food. Milk and milk products are frequently implicated in staphylococcal food poisoning [ 32]. van Schaik G, Green LE, Guzma’n D, Esparza H, Tadich N. Risk factors for bulk milk somatic cell counts and total bacterial counts in smallholder dairy farms in the 10th region of Chile. Prev Vet Med. 2005;67:1–17.Fig 3: The main milking plant filter should not be used as a method of mastitis detection Conductivity Millogo V, Sjaunja KS, Ouédraogo GA, Agenäs S. Raw milk hygiene at farms, processing units and local markets in Burkina Faso. Food Control. 2010;21:1070–4. The problem with relying on udder examination as the first line of mastitis detection is that udder changes are detectable fairly late in the process, so by the time disease is detected considerable losses have occurred. Later identification also means delayed treatment, which tends to be less effective than early treatment and increases the risk of disease spread

O, Solbu H, Refsdal A, Roalkvam T, Filseth O, Minsaas A. Results and evaluation of thirty years of health recordings in the Norwegian dairy cattle population. J Dairy Sci. 2007;90:4483–97. Heikkilä A-M, Liski E, Pyörälä S, Taponen S. Pathogen-specific production losses in bovine mastitis. J Dairy Sci. 2018;101:9493–504.

Mild mastitis: Abnormality of the milk is the main sign with little evidence of change in the udder and no systemic signs such as dullness and loss of appetite. Amentie T, Eshetu M, Mekasha Y, Kebede A. Milk postharvest handling practices across the supply chain in eastern Ethiopia. J Adv Vet Anim Res. 2016;3:112–26. Lin H, Peddada SD. Analysis of compositions of microbiomes with bias correction. Nat Commun. 2020;11:1–11.