| Category | Agtech |
| Keywords | Cow Fertility, |
| Current development stage | General list: TRL3 Experimental proof of concept |
| Collaboration Opportunity | Sponsored Research with an option to License Research Results |
Abstract
The researchers developed a technique to detect cow fertility during in-line milking, which will reduce costs and increase dairy farm revenue.
Background
The modern dairy industry is based on intensive reproduction and fertility management grounded in tracking cow conception time, pregnancy diagnostic, parturition (birthing), postpartum, and lactation. Pregnancy is the ‘engine’ that drives lactation, therefore, knowing the pregnancy status of the cow is highly important for herd management.
Additionally, early detection of pregnancy loss provides the farmer with important information about the optimal time for the next insemination. This information might decrease the open days i.e., days between parturition and the subsequent pregnancy, thereby minimizing the economic loss.
The most precise and reliable method for pregnancy detection is invasive, time-consuming, and requires professional labor since either rectal palpation and/or ultrasound examination is required. Rectal palpation is generally performed between 40 to 44 days after artificial insemination (AI) while the ultrasound is effective earlier on, 26 to 35 days after AI.
Farmers generally use less invasive methods for monitoring cow fertility, including cow behavior (mounting) and hormone (progesterone) level sampling of the cow blood or milk. Progesterone levels are tested twice, on days 18 to 24 post insemination and later at day 60 during gestation to confirm pregnancy, since these assays are highly accurate for identifying non-pregnant cows, its diagnosis of pregnant cows is relatively poor, i.e., with a high number of false-positive identifications.
Unfortunately, there is no technique to easily monitor progesterone levels during in-line milking and thus, there is a real need to develop a non-invasive method to detect pregnancy and early embryonic loss in cows.
Our Innovation
The researchers developed a sensing device for measuring early pregnancy in cows that is based on Microwave Dielectric Spectroscopy (MDS) that can monitor pregnancy-specific attributes in the milk, during the milking process via an in-line milking device. The dielectric spectroscopy is coupled with machine learning, an artificial neural network algorithm. We believe that the system may reach an accuracy of 99%. This in-line milking technique gives farmers a fast measure of a cow’s fertility, heat, and pregnancy status without invasive techniques.
Technology
By using an in-line system to detect pregnancy, nonpregnancy, or pregnancy loss during each milking, cow fertility could be continually monitored on a daily basis without interfering with the cow’s daily routine. Integration of this information into a computerized software system would allow the dairy manager to review the pregnancy status of individual cows, might prevent mistaken diagnoses, and improve the reproductive management and the farmer’s profit margins.
At the Center of Electromagnetic Research and Characterization (CERC) at the Hebrew University of Jerusalem, an experiment was conducted to measure the milk from ten cows weekly (around 30 weeks) using MDS. The cows were monitored in different reproductive states, starting from a non-pregnancy condition, followed by insemination, and confirmed pregnancy by a veterinarian.
The most valuable insights were obtained from the dielectric parameter Δε during the transition from the non-pregnancy state into the inseminated state. The CC fitting parameters Δ𝜀, 𝜏, 𝛼, and 𝜎𝑑𝑐 were used to track the changes in the dielectric response of bulk water in milk before and after insemination. All of the fitting parameters (Δ𝜀, 𝜎𝑑𝑐𝛼, 𝜏) had a significant decrease in variability after the first two weeks of successful insemination (i.e., an insemination that followed by positive pregnancy diagnostic).
Findings
1) Transition to pregnancy status in terms of the fitting parameters in five cows confirmed as pregnant. A significant reduction in the variability of CC data can be observed during the transition, divided into three stages: (1) beginning with a nonpregnancy condition; (2) progressing to 15 days of pregnancy; and then (3) continuing to the 16th day. The first state had 66 nonpregnancy points, the second state had 8, and the final state had 14 points. All points in each state were arranged in a single column.
2) CC parameters of milk from confirmed pregnant cows. The variability of the fitting parameters significantly reduces when artificial insemination is successful. A more substantial reduction of variability is observed after the 16th day of pregnancy.
The results highlight the potential of Microwave Dielectric Spectroscopy (MDS) as a non-invasive method for pregnancy detection in cows. The sensitivity of this approach is based on the reduction of the variability in the dielectric parameter: Δε. However, it is important to note that this analysis method requires further improvements that can be achieved by large scale studies. Nonetheless, the ability to predict pregnancy in such a small number of animals, with successful insemination rates of below 40% indicate that our approach can be implemented in the management of dairy herds to follow reproduction events.
The findings of this study suggest that MDS holds promise as a potential method for the development of a non-invasive device for pregnancy detection. To advance this research, we propose the utilization of machine learning approaches, which typically require a larger volume of data. Given that milk can be readily obtained on the farm two to three times daily, it offers an opportunity to collect a substantial amount of data effortlessly. By incorporating machine learning techniques and leveraging the abundant data availability, we can further enhance the accuracy and effectiveness of pregnancy detection using MDS.
Patents and Publications
N. Argov-Argaman, Yuri Feldman, Z. Roth (2021) Method for determining progesterone levels in milk of a lactating cow ref.: International application (PCT) number WO 2021/234689, filed on May 13 2021, filed in National Phase in US and Europe . Yissum Ref 4183
Galindo CJ, Levy G, Argov-Argaman N, Roth Z, Feldman Y (2021) Bovine milk microwave dielectric response to pregnancy 13th International Conference on Electromagnetic.
