Developing an automated and cost-effective method for early detection of subclinical mastitis to reduce the use of antibiotics on dairy farms. Invisible (subclinical) mastitis decreases milk quality and production. Invisible mastitis is linked to an increased use of antibiotics and the risk of the emergence of antibiotic-resistant bacteria, a major public health concern worldwide. Early detection of the infected cows is of great importance. Here, we developed a new approach for early mastitis prediction based on milking parameters. We employed machine learning to generalize the application of the developed approach

Principal investigator

Esmaeil Ebrahimie
Magnifying glass

Area of science


Systems used


Applications used

cellRanger, STAR, Rapidminer, CLC Genomics Workbech, RSubread
Partner Institution: The University of Adelaide and La Trobe University | Project Code: b222a778708d4f8ba8994d5078e75f07

The Challenge

Invisible (subclinical) mastitis, caused by bacterial intra-mammary infections, is the most widespread and economically challenging disease of dairy cattle. Subclinical mastitis is up to 40 times more common than clinical mastitis and far more difficult to detect, thus having a greater economic impact. Subclinical mastitis reduces milk production and quality and is associated with an increased use of antibiotics, a major public health concern worldwide. To control subclinical mastitis, accurate surveillance strategies are required.

Test-day somatic cell count is the most common detection method. However, test-day somatic cell count fluctuates widely between days, causing major concerns for its reliability. In addition, this method is a late indicator when the infection is spread. Consequently, there would be great benefit to identifying additional efficient indicators from large-scale and longitudinal studies.

The Solution

We demonstrated that the development of machine-learning expert systems using milking features may produce a predictive pattern for early detection of subclinical mastitis. In this research, for the first time, we integrated machine learning and meta-analysis to correct the farm variation effects and discover early prediction models of invisible mastitis. The results are available for the farmers across the Australia and New Zealand as an automated and cost-effective machine-learning expert system for early detection of hidden mastitis using milking parameters.


The Outcome

Allocated Pawsey resource (Nimbus) was used or developing an automated and cost-effective method for early detection of subclinical mastitis to reduce the use of antibiotics in dairy farms and increase the milk quantity and quality. To this end, a variety of machine learning models were applied on big datasets of a variety of milking parameters in a longitudinal multi years study. The results are published in internationally recognized journals.

List of Publications

1- Ebrahimie E, Mohammadi-Dehcheshmeh M, Laven R, Petrovski KR (2021) Rule Discovery in Milk Content towards Mastitis Diagnosis: Dealing with Farm Heterogeneity over Multiple Years through Classification Based on Associations. Animals 11, no. 6: 1638.

2- Mohammadi-Dehcheshmeh, M., Moghbeli, S. M., Rahimirad, S., Alanazi, I. O., Shehri, Z. S. A., & Ebrahimie, E. (2021). A Transcription Regulatory Sequence in the 5′ Untranslated Region of SARS-CoV-2 Is Vital for Virus Replication with an Altered Evolutionary Pattern against Human Inhibitory MicroRNAs. Cells, 10(2), 319.

3- Ebrahimie E, Zamansani F, Alanazid IO, Sabie EM, Khazandi M, Ebrahimi F, Mohammadi-Dehcheshmeh M, Ebrahimi M (2021) Advances in understanding the specificity function of transporters by machine learning. 138: 104893