Predictive Biology: What does the future hold for mastitis control?

Green and his colleagues developed software for farmers. Using data, they’ve been able to make predictions on cost effectiveness. This information is available to farmers through the AHDB Dairy Mastitis Control Plan.

Ultimately, disease control programs have one goal: to predict disease and take measures in order to prevent disease from occurring. According to Dr. Martin Green, Professor of Cattle Health & Epidemiology at the University of Nottingham, more can be done to prevent mastitis. Green spoke on the subject at the World Buiatrics Congress in Dublin in July.

Early in his presentation, Green discussed current approaches to mastitis control, but pointed out that most used are reactive, rather than prospective. “Whilst automation has increased and improved, and in many developed dairy nations, classical contagious mastitis pathogen are now at low prevalence, there remains a reliance on generic control strategies that originated in the 1960s,” Green wrote in his presentation report.

In particular, he spoke about current on-farm investigations conducted by veterinarians and advisors who are asked to assess farm records and management practices. Following that, they may suggest a new approach, one that incorporates the principles of herd health management, including goal setting and on-going disease monitoring. Management practices are amended until incidents of disease decrease or no longer occur.

The key question now, suggests Green, is how to move from current approaches to mastitis control. A more ideal method would be to make use of current data to predict what is likely to happen in the future, he said. These opportunities, more and more, are becoming a reality.

Four examples of predictive biology for mastitis management

1. Predicting cow susceptibility

Predicting the susceptibility of an individual cow can be extremely valuable not only when it comes to making breeding decisions, but also in terms of individualized management practices. Green believes that susceptibility is, to some extent, dictated by genetic influences. He suggests that a polygenic approach is necessary as it is clear that more than one gene is responsible for resistance to mastitis. Susceptibility, too, is only partially influenced by the individual cow’s initial genetic makeup. This is especially true since environment tends to alter susceptibility as well.

Some work in this area has been done, said Green in an interview. Most work, though, has been done in terms of milk production and milk constituents. “But in relation to health and mastitis, not as much has been done,” he said. “We haven’t fully cracked it yet.”

One area that shows great promise in understanding susceptibility and immunity, though, is the study of the udder microbiome. Research has shown that microbial communities in milk are quite complex, and that they seem to differ between individuals. “One of the key goals of research in the different areas is to try to link the makeup of the bacterial community with future disease susceptibility,” said Green.

“The bacterial communities are identified usually using genetic tests, so you can identify species of bacteria from particular genes and the sequences of particular genes that they carry,” he explained. “Essentially what the research does is to try to see if the mixture of different species makes you more or less prone to disease.”

Green believes that an understanding of the mammary microbiome could help to more accurately predict resilience. Furthermore, it could be possible to manipulate the microbiome in order to optimize mammary gland health. Research in this area is ongoing, he said.

2. Understanding pathogen phenotypes

Dr. Martin Green
Dr. Martin Green

The pathogens that cause mastitis can be categorized at different levels, but for the most part their name refers to the species level. Within the species, though, there are different strains or subtypes, which may be broken down into groups according to their behaviour. Phenotypic traits, said Green, include being associated with ‘contagious’ or ‘environmental’ transmission; chronic or acute disease; and long or short duration rise in SCC. These characteristics, however, will not be the same at the sub-species level of the individual mastitis-causing pathogens. It is, therefore, probable that making predictions on the farm is possible through a better understanding of sub-species and phenotypic infection patterns. During the conference, Green gave a recent example of how subspecies typing of the pathogen S. uberis has been used to differentiate strains showing ‘contagious’ as opposed to ‘environmental’ transmission patterns.

3. Predicting the cost effectiveness of management changes

Especially given the current state of milk prices and the volatility of the market, the resources of dairy farmers are sometimes quite limited. For this reason, it would be good to know not only which interventions work best, but also which ones provide the best return on investment. Data in this area are, however, quite sparse, said Green.

“Whenever we make changes on-farm we can never be exactly sure how they are going to turn out,” said Green. “There are lots of reasons for that, but it’s often because when researchers conduct studies on management it’s conducted on specific farms, in certain circumstances.”

When you assess this information on a whole, you can determine how likely a change is to be effective and how likely the changes are to be cost effective, said Green. Having this information, he said, can help farmers to prioritize changes that they make on-farm. In the UK, Green and his colleagues have conducted research in this area and developed software for farmers. Using data, they’ve been able to make predictions on cost effectiveness. This information is available to farmers through the AHDB Dairy Mastitis Control Plan.

4. Predictions from simulation

Finally, Green believes that data simulation could also be used to combine information to create “what-if” scenarios. It is an area of study often overlooked, he said, but one that could be used to assess the impact of uncertainty in research findings. In order to make predictions, simulation studies use information gathered from ‘real life’ situations, where possible, and look at them in combination in order to identify patterns and identify circumstances when control measures are likely to be useful. Those control measures are then ranked in order of their usefulness.

Green discussed one example in a recent interview. “In one simulation technique, we were looking at trying to predict the affects that mastitis would have on herd fertility,” he explained. “There’s been quite a lot of research that shows how mastitis affects fertility individual circumstances. What you need to do in a simulation is to put the data from those studies together with other factors that in this case can affect fertility as well as mastitis.”

“For each of those different possibilities, you also simulate other things that affect fertility and put them together to work out which have the biggest impact,” he continued. “Again, it’s about prioritization, but it helps you to understand when you change one thing what other things might change, and to predict the overall outcomes per farm.”

In the future, the quality and quantity of data and technology will only continue to improve and thus provide further opportunities to make more accurate biological predictions. Perhaps one day, rather than responding in a reactive way, veterinarians will be able to able to advise farmers on how to predict mastitis more prospectively before it occurs.