Current Progressive Dairy digital edition

3 Open Minutes with Dr. Kent Weigel

Progressive Dairyman Editor Peggy Coffeen Published on 18 July 2014

kent weigel

University of Wisconsin – Madison Professor and Department of Dairy Science Chair Dr. Kent Weigel believes a trait for improving feed efficiency will be “the next frontier of genetic selection in dairy cattle.”



Weigel recently addressed this topic at the Four-State Dairy Nutrition and Management Conference in Dubuque, Iowa, and he is on the docket for the upcoming World Congress on Genetics Applied to Livestock Production set for Vancouver in August.

Weigel shares his research and observations regarding improving dairy cattle feed efficiency through genetic selection with Progressive Dairyman Editor Peggy Coffeen and why he thinks genomics will be the key to unlocking this trait’s potential.

Describe the balance between increasing milk production and feed efficiency. At what point is it no longer cost-effective to push for more milk?

WEIGEL: I think we will always be pushing for more milk, but the marginal gains in profit from each extra pound of milk might not be as much as they were in the past. For example, going from 20,000 to 21,000 pounds of milk might not require as much extra investment in specialized feed additives, facilities and other management practices as going from 35,000 to 36,000 pounds.

This is sort of the “multiples of maintenance” argument, where we see diminishing returns in captured net energy (as milk) as a proportion of gross energy intake as the cow goes from consuming three times maintenance requirements to consuming five times maintenance requirements.


What have been the barrier in the past to effectively improving dairy cattle feed efficiency?

WEIGEL: I think we have made great strides in improving feed efficiency from a management point of view, with better ration formulation, pushing feed up more frequently, reducing storage losses and so on. In terms of genetic improvement, it has been a real challenge because we cannot measure individual animal feed intakes routinely on a large number of animals.

Before genomics, the basis of genetic improvement was progeny testing, and this meant that we had to measure 50 to 100 daughters of every young bull in the A.I. stud each year. Roughly speaking, this would cost about $20,000 per young bull per year in order to get a PTA for dry matter intake (DMI) or residual feed intake (RFI).

How has genomics changed the way you study this trait?

WEIGEL: For feed efficiency and other traits that are difficult or expensive to measure on a large number of animals, genomics offers great p otential. The reason is this: We don’t have to measure individual intakes on every cow in the country – we just need an adequate reference population from which we can create the genomic prediction equations (i.e., estimate the SNP effects).

The definition of “adequate” is a little tricky because measuring intakes on a group of 5,000 or 10,000 cows is still a huge task and that may only give us reliability (REL) values of 30 to 50 percent, which is not nearly as good as other traits but hopefully good enough to make some genetic progress.


Once we’ve measured the reference population animals, which in our case are mostly on experimental farms, we can generate genomic PTAs for other bulls and cows in the population, and then we just have to measure a few thousand additional animals over time to keep the prediction equations up-to-date.

If we know that cows with larger body size must eat more to meet greater energy requirements for production and body maintenance compared to their more moderately sized herdmates, why not just breed for smaller cows?

WEIGEL: That would certainly be an option – or maybe for starters we should just stop selecting for bigger cows. Stature of Holstein cows is changing (increasing) faster than any other linear type trait because people who like to show and merchandise registered cattle tend to like big cows.

Holstein Association USA is taking some steps in the classification program to stop rewarding enormous cows with high final scores, and hopefully this will slow down the trend. That said, body size isn’t the only answer, and it probably isn’t the best answer.

Depending on how one does the economic calculations, smaller cows fall somewhere between slightly more efficient and equally efficient as big cows. It’s pretty hard to make legitimate economic assumptions that will tell us big cows are more efficient, but the degree to which they’re less efficient might be less than people think due to fixed costs that accrue on a per-cow basis as well as high salvage value prices.

My view is this: We should try to measure and select for biological efficiency of feed utilization by measuring traits such as DMI and RFI while also putting a modest penalty on body size that will stop the ever-increasing stature of the Holstein cow.

DMI is a metric already familiar to dairy producers. Why not use that as the basis for feed efficiency?

WEIGEL: The genetic and phenotypic correlations between milk yield and DMI are very high, which is sensible since cows need to eat a lot in order to produce a lot of milk. Many of our European collaborators prefer the approach of putting DMI into the selection index directly rather than trying to convert it to RFI first.

If the selection index calculations are done properly, we should get equivalent results from selecting on an index based on DMI and an index based on RFI. The difference comes in the selection index weights because, for example, we would need a large positive weight on milk yield to get more milk and a large negative weight on DMI in order to make sure they don’t eat too much while doing it.

This is tricky, and in order to do it correctly, we need very accurate estimates of the genetic correlations between DMI and all other production, type, health and fertility traits.

On the other hand, if we calculate RFI based on the difference in DMI between a cow and her herdmates (after adjusting for body size, milk yield, milk composition, etc.) then RFI will be more or less independent of the other traits in the index, and we can just add it to Net Merit or TPI when it becomes available.

What is residual feed intake (RFI), and how can it be used to measure feed efficiency as a genetic trait?

WEIGEL: As noted above, it is the difference in DMI between a cow and her herdmates, after adjusting for the energy used for milk, bodyweight and bodyweight change. So it is independent of yield and body size, and we can select for higher milk yield and smaller or larger body size at the same time as selecting for lower RFI.

The idea is: RFI is extra feed consumed that provides no economic returns to the producer, and it is common for certain cows to eat several pounds more or less DMI than their herdmates on a relatively consistent basis. We know this extra feed is not used to make extra milk, higher fat or protein percentage, extra growth or body size maintenance.

The challenge is determining what RFI actually represents from a biological point of view. Is the more “efficient” cow eating less feed because she is digesting the feed more efficiently and then making less manure or methane? Or is she eating less feed because she’s losing less energy as heat?

Or is she eating less feed because she expends less energy on walking and other physical activity? Or is it that she’s eating less feed right now and appearing as efficient, but she is more likely to become sick or infertile in the future due to inadequate nutrition? We haven’t answered all of these questions yet.

How far away are we from having a genomic PTA that measures feed efficiency?

WEIGEL: We have a little less than two years left on our research project, but we’ve already got data on several thousand cows and have done a lot of the preliminary work toward a genomic evaluation for DMI and RFI.

Plus, we’re working closely with USDA-ARS scientists in Beltsville, Maryland, to make sure they’re part of the process, as this will speed the time to delivery of genomic evaluations once the project is finished.

I’d say within a year we’ll have some preliminary sire evaluations that can be looked at unofficially for the purpose of characterizing what types of bulls have daughters with high DMI, low RFI and so on.

Within two years, we should have predictions for all genomically tested bulls in the U.S., and at that point the Council on Dairy Cattle Breeding (A.I. studs, breed associations, milk-recording organizations and dairy records processing centers) and their stakeholders/members will have to determine whether this is something we want to evaluate routinely and incorporate into our selection indexes. PD

Click here to read more about Weigel’s presentation at the Four-State Dairy Nutrition and Management Conference.

peggy coffeen

Peggy Coffeen
Progressive Dairyman