Combining ultrasound data and MBV for selection tools

Oct 15, 2010
by WLJ

The world is rapidly changing. Technologies that were barely imagined just a few years ago have become a normal way of life.

Genomics is the new buzz word for beef, and research into the genetic make-up of cattle holds great promise. But no new science is complete without practical context, and in beef selection and production, that means measuring physical attributes, or phenotypes, to assess end result.

Larry Kuehn, research geneticist in the Genetics and Breeding Research Unit of the U.S. Meat Animal Research Center (US- MARC), part of the USDA Agricultural Research Service, makes the case for continuing to collect phenotypes, beginning with the limitations of genotypic data. “Genomic-enhanced expected progeny differences (EPDs) show a decent amount of accuracy, but they’re not perfect,” he explains. “Ultrasound measurements provide sources of variation in actual carcass merit in addition to that provided by the genomic tests; and that contributes a substantial amount of accuracy.”

Kuehn’s colleague Mark Thallman, also a research geneticist at USMARC, adds: “Current genomic tests are reported in the form of molecular breeding values (MBVs), which are similar to EPDs, but based on DNA tests instead of phenotypes. In the national cattle evaluation system, MBVs are merged with phenotypes and pedigree to produce EPDs. The DNAbased approaches are an exciting development, but currently we’re accounting for relatively minor proportions of genetic variations with them. We still get better accuracy by accounting for more of that variation by including additional sources of information.”

Matt Spangler, assistant professor of animal science at the University of Nebraska - Lincoln, further explains the advantage of using two sources of data. “If we take a DNA sample on an animal (assuming that DNA actually does predict some of the differences and the MBV is included in the EPD calculation), we’ll increase the accuracy of the EPD,” he says. “Likewise, if we collect ultrasound records and include that data in EPD calculation, the accuracy of that EPD increases because we have a record on an individual.”

Published work by Spangler, J.K Bertrand and R. Rekaya in the Journal of Animal Science backs that up. “Jointly considering molecular and phenotypic information showed greater long term response as compared to tandem selection, showing that discrimination of candidates for selection based solely on molecular information is not optimal.” (Spangler, Bertrand, Rekaya 2006) M.D. MacNeil and his team also make the case for both sources of information in citings included in their work “Genetic Evaluation of Angus Cattle for Carcass Marbling Using Ultrasound and Genomic Indicators” (MacNeil et al., 2007) “If both phenotypic and molecular data are available, their joint consideration is the most powerful selection strategy.” (Dekkers and Hospital, 2002; Spangler et al., 2007) Kuehn explains why ultrasound data, in particular, is needed to make the equation work. “In terms of ultrasound,” he says, “the advantage is rapidly getting pedigree, or progeny, phenotypes. In other words, if both the ultrasound and MBV tests had the same correlation with an actual carcass measure, both would increase the accuracy of that carcass EPD by the same amount. But ultrasound can also be done on that bull’s progeny, and that will make accuracy shoot up quite a bit higher. Getting a MBV on progeny doesn’t do anything for that.”

Again, MacNeil and company concur. “In predicting breeding values, records of MBV from relatives do not increase accuracy, whereas IMF records from relatives improve the accuracy of prediction.” (Beef CRC, 2009).

Keeping it honest

Continuing to collect phenotypic data is also crucial to ongoing accuracy of the genetic information.

“These MBVs probably aren’t good forever,” says Kuehn. “They’re kind of time dependent.” He says they do a good job for current populations, but over time, the relationship of the MBV with the current set is going to deteriorate. “They will become reorganized, fixed in some cases. They aren’t going to work forever with the same level of accuracy.” That means it’s important to continue phenotypic collection in order to continue to make marker panels that are useful down the road. “That’s nothing negative against the MB- Vs,” adds Kuehn, “but phenotypes are still important.

They are the last thing we want to get rid of.”

Spangler agrees. “We need something to continue to validate those genomic tools,” he says. “It’s not like a DNA marker panel comes out onto the market and for perpetuity, it predicts marbling extremely well. It doesn’t work that way. We have to ‘retrain’ it every few generations, and by that I mean validate it again to see how good it’s doing. And the only way to do that is to have phenotypes to validate it against.”

Again, MacNeil’s team agrees: “Continued data collection will be needed to update EPDs and MBV to account for erosion in LD, for changes in allele frequency that result from selection, and because some sires used to construct the MBV had low accuracy EPDs. To date, high accuracy EPDs continue to require the collection of phenotypic data for economically relevant traits.” (Mac- Neil, Northcutt, Schnabel, Garrick, Woodward and Taylor, 2009)

Working data

Thallman breaks the issue of continued phenotype collection down into two practical applications.

“In the micro view, for the individual producer, we need to look at it from the perspective of how do we evaluate each individual animal. In order to achieve the kind of accuracy people would prefer to have in their breeding program, we need more information than a single source would provide. Probably more than just a DNA test would provide today,” says Thallman. “and just having that additional info from ultrasound is going to increase accuracy substantially, rel ative to what we can do with just a DNA test.”

From a more macro perspective, continued collection of phenotypic data and its incorporation into EPDs is essential for the industry. “It’s really a benefit to the industry to continue to collect all these phenotypes,” says Thallman, “in order to really know how much impact DNA tests are having, and how much variation they’re accounting for. To understand what the relationships are between DNA tests, ultrasound, and carcass data, we’re going to have to collect all those sources of information.”

For the sake of science

From a pure research standpoint, there is an even stronger case to be made for the use of ultrasound and phenotypic measurement.

“As a center [USMARC], it’s important for us to make sure these MBV tests are being done very well. And to do them right, we need a lot of phenotypes, and most of us don’t have anywhere near what we need right now,” says Kuehn. “In terms of trying to develop new novel phenotypes for things like disease resistance, cow longevity, or whatever, we need a lot of phenotypes. If we’re going to try to use this marker technology, we’re going to need more than we’ve been collecting, not less.”

Spangler adds: “Fundamentally, the collection of phenotypes isn’t going to go away. Ultrasound has provided a very reliable data set across a multitude of breeds to include in national cattle evaluations, and make very valuable EPDs, and we’ve seen the amount of change that’s occurred because of that,” he says. “So ultrasound is a very valuable and reliable technology from that sense.

“Genomic technology holds a tremendous amount of promise,” he continues, “but for it to be truly as useful as it could be, genomic information has to be integrated into EPDs. And so that becomes the benefit of ultrasound right now. Across the board, it is.” — WLJ