Mathematical approaches to detect low concentrations in progesterone profiles

Area of Science:

• Animal Science
• Reproductive Biology
• Dairy Cattle Management

Background:

• Accurate assessment of dairy cow fertility is essential for genetic selection and herd management.
• Current methods for detecting estrus events can lack objectivity and accuracy.
• Low progesterone concentrations are reliable indicators of estrus in dairy cows.

Purpose of the Study:

• To develop and evaluate mathematical approaches for objective detection of low progesterone concentrations in dairy cows.
• To establish reliable starting points for selecting genetically superior dairy cows based on fertility indicators.
• To improve the accuracy of identifying estrus events in dairy cattle.

Main Methods:

• Utilized three mathematical methods: an exponentially weighted moving average control chart and two in-house threshold methods.
• Analyzed two datasets comprising 97 and 160 insemination records from Holstein dairy cows (first-lactating, primiparous, and multiparous).
• Evaluated model performance using sensitivity, specificity, and error rates with a progesterone threshold of 1.2 ng/ml skimmed milk.

Main Results:

• High sensitivity (93.13%–100%) and specificity (97.92%–99.93%) were achieved across the models.
• True-positive low progesterone concentrations ranged from 0.18 to 0.28 ng/ml (first dataset) and 0.21 to 0.26 ng/ml (second dataset).
• A revised threshold of 0.8 ng/ml skimmed milk is suggested for indicating luteal activity in defatted foremilk.

Conclusions:

• Mathematical models effectively detect low progesterone concentrations, aiding in estrus event identification.
• The developed methods offer improved objectivity and accuracy for dairy cow fertility assessment.
• This approach provides a foundation for further research, including estrus interval calculations.