HappyMoo, a new project which aims to link milk sample data with cow well-being parameters, is being developed.

The system aims to provide dairy farmers with early notification of issues relating to health and well being being of individual cows before they become a problem and affect performance.

National Milk Records is participating, along with major dairy organisations in north west Europe, through ‘machine learning’ – between milk spectral data and conditions exhibited in the cow.

“Results from milk samples tested through NMR’s mid-infrared machines in the laboratory from cows with a particular condition, such as lameness or mastitis, or body condition changes will be compared with those that don’t show these conditions,” says NMR service development manager Martin Busfield.

“Patterns that emerge will help to identify, through the milk sample, cows with a particular condition and can then be used to provide valuable management information for farmers.”

This 42-month Interreg North West Europe project, funded by the European Regional Development Fund, will use milk spectral data from participating project partners, which includes national dairy recording organisations in Germany, France, Belgium and Ireland.

“The data can be compared from laboratories across the project participants, thanks to the well-established monthly standardisation of machines across the European milk recording group,” he adds. “We can, therefore, be confident that we are comparing like-for-like data and, with large volumes of data, we can use machine-learning to build accurate and reliable predictions.”

Relating data from milk samples to physical traits is a well-established technique used for measuring milk constituents in bulk and individual milk samples by recording organisations and testing laboratories.

Looking further ahead, the Happy Moo project group is also looking at well-being factors, like hunger and stress, and using research farms and university units to carry out additional testing on body condition scores. These tests relate to energy balance and hunger, and on hair sample tests to detect cortisol – a hormone produced when animals are under stress.

“Again, milk samples will be tested from animals exhibiting these conditions and compared with those not affected.

“The more milk samples we have for each parameter, the more reliable the milk spectral patterns will be. The aim is to be able to integrate this information as part of the milk recording service, providing farmers with an early and non-intrusive cow health and welfare screen.”