More than 1,800 dairy scientists from 48 countries came together to discuss innovative dairy research at the June 2018 American Dairy Science Association (ADSA) annual meeting in Knoxville, Tennessee.

Researchers shared their work and progress using vast amounts of data being collected on farms as well as new traits in development to aid in dairy cattle genetic selection.

Read on for a recap of a few key take-away presentations on genetics and technology at the conference.

The genetics of sire fertility

Presented by Francisco Peñagaricano and postdoctoral research assistants Fernanda Rezende and Juan Pablo Nani; University of Florida

Peñagaricano and his team previously identified genomic regions harboring mutations that have significant non-additive effects on sire fertility. Their current research, presented at this conference, built on that. It involves incorporating those mutations into a prediction equation to estimate the sire’s fertility before he ever produces semen.


The dataset used within this research included 11,500 Holstein and 1,500 Jersey bulls with sire conception rate data and genomic DNA information. When incorporating the non-additive genetic effects with the additive effects normally associated with breeding value calculation, they obtained predictive correlations up to 0.4 for Holsteins and 0.29 for Jerseys.

These accuracies are similar to some traits, such as sire calving ease, currently evaluated for U.S. dairy cattle.

The research Peñagaricano presented is another step toward developing a genetic prediction for sire fertility, which should help A.I. organizations provide more fertile bulls to dairy producers.

Furthermore, the breeding values we see every day on a proof sheet are meant to show how the progeny of a bull (or cow) are expected to perform for a given trait. The methodology Peñagaricano uses in this research includes the non-additive genetic effects. This is more suited to predict how that animal will perform for a given trait – in this case bull fertility.

This method may also be used for health traits, where predicting a cow’s susceptibility to a disease may be just as valuable as predicting how susceptible a sire’s progeny will be.

More accurate predictions of disease susceptibility could help in culling decisions or looking at preventative treatments for disease for specific cows.

The genetic reliability of feed efficiency

Presented by John Cole; USDA Animal Genomics and Improvement Lab

Dairy producers have a growing interest in feed efficiency. Over the past several years, universities and research groups from around the world have collected individual feed intake data for lactating cows. The goal is to create a feed efficiency trait for genetic selection.

Information from about 4,000 U.S. research cows was used in a cross-validation method to estimate the reliabilities of genomic predictions for residual feed intake. The sires of the cows in the study, who would essentially have daughter proofs for residual feed intake, had reliabilities up to 85 percent. But when trying to predict residual feed intake on genomic young sires commonly used in commercial dairy herds today, reliabilities reached only 12 percent.

Cole showed increasing the reference population of cows could increase those reliabilities. If individual feed intake and production data were available from 20,000 genotyped cows, a reliability of about 31 percent for genomic young sires could be achieved. If 50,000 cows were in this reference population, the reliability could reach up to about 50 percent. A major goal for the feed efficiency research group moving forward is to continue to add more cows to this reference population.

Virtual Dairy Farm Brain

Presented by Steven Wangen and Adam Christensen; Wisconsin Institute for Discovery

Independently, data from feeding systems, milking parlor software, genetic testing, management choices and economic analyses are useful in making on-farm decisions. Integrating these systems can provide a much richer picture of farm operations and help producers make more accurate and impactful decisions in real time.

Wangen and Christensen helped develop the Virtual Dairy Farm Brain project in conjunction with faculty from the University of Wisconsin – Madison.

The project looks to utilize data from on-farm management software, feed monitoring systems and parlor software to combine with genetic information, DHI test data and other third-party service data. This data is then cleaned, normalized and stored.

By combining information from many farms, predictive models and analytic tools can be developed. One example would be developing a prediction model to help detect health issues earlier and provide a data-derived treatment plan, unique to that specific situation.

By including economic information, Christensen explained they can create an individualized lifespan model for each cow. This will account for the cost and revenue of the cow over her lifetime and will change in real time as the cow ages and more data from that cow and farm is received.

For example, a cow will have a predicted lactation curve based on her farm, her genetics, and her reproductive and health history. As her initial lactation data and the data of her herdmates is received, her predicted lactation and economic value will be adjusted. This value could also change, given fluctuations or speculations in feed costs or milk prices.

In the future, they hope to develop analytical tools that can help producers and consultants navigate, visualize and analyze the data they are getting from an increasing number of sources. Their end goal is to provide accurate and manageable decisions that can be made in real time on an individual cow level, pen level or even a whole-farm level. end mark

PHOTO: From predicting sire fertility in bulls before they are old enough to produce semen, to progress toward a selection trait for feed efficiency, scientists are hard at work on projects that could revolutionize the dairy industry. Photo courtesy of Alta.

Doug Bjelland