Developments in Single Step in U.S. Beef Cattle Evaluation

Single-step developer explains how the technology improves genetic predictions and what's next.

by Lindsay King, assistant editor, Angus Journal

LOVELAND, Colo., June 22, 2018 — Daniela Lourenco, assistant professor of animal science at the University of Georgia, is the mastermind behind single-step methodology. It is now the standard for genomic evaluation in beef cattle, and the adoption rate is astounding.

“The number of genotyped animals is increasing quickly. It just shows the commitment of breeders and producers to genetic selection,” Lourenco said. “The American Angus Association does not have the biggest data set in the U.S. Holsteins have that one, but Angus certainly has a massive set also.”

Lourenco’s analogy of the why behind single step comes in the form of a pickup using fuel, engine oil and additive to perform more efficiently. The truck represents the mixed model equation, fuel represents the phenotype, engine oil is the pedigree and the additive is the genomic information.

“We are just improving the relationships, making them better. Just like each component for the truck,” Lourenco explained. “We would expect full sisters to share 50% of their genotype, but if we run single step, it usually ends up being more or less. The difference is what single step uses to be more accurate.”

Validation measures are put in place to quantify how genomic information is doing within a system. Lourenco removed the phenotype from the animals and then ran the single step on the data. This revealed whether the new or old system was doing a better job of predicting future genotypes.

“Since calving ease is recorded in several different categories, we wanted to validate this trait,” Lourenco said. “The increase in accuracy for calving ease with this genomic info was only 1%. We wanted to know why that was, so we needed to validate the maternal traits.”

This forced Lourenco to find a different system of validation. She stumbled across a linear regression metric to check the consistency between consecutive evaluations. Lots of complicated statistics later, the linear regression method increased in accuracy for the validation.

“Sometimes we are not using the right method to compare things, and we blame genomics for it,” Lourenco said. “Not all organizations have a lot of genotypes. There are several methods able to work with multiple methods.”

When first working with Angus’ massive genomic data set, Lourenco was overwhelmed by the number. This led to the development of the truck-and-trailer analogy.

“When we have more genotyped animals than we have SNPs available for and there are some linear relationships, that means we cannot invert the matrix,” Lourenco said. “Some think APY does some magic, but it is just an algorithm that constructs G when inverting G is not computationally feasible.”

Some segments of the matrix are linearly independent, the “non-core animals.” Lourenco defined “core animals” as randomly selected individuals whose genotypes were used in the matrix. Core animals are linearly dependent and a certain percent of the total number of animals genotyped. To carry the load in the equation, core animals are in the hypothetical truck, while the non-core animals are in the trailer it is pulling.

“Single step is under constant improvement at the University of Georgia, other universities and in research groups. Scientists need to keep developing more methods for improved accuracy evaluations, but producers also have to keep collecting data,” Lourenco finished. “The methods and systems we develop are fully dependent upon the data, so keep collecting that.”

Editor’s Note: The articles used within this site represent a mixture of copyrights. If you would like to reprint or repost an article, you must first request permission of Angus Media by contacting the editor at 816-383-5200; 3201 Frederick Ave., Saint Joseph, MO 64506. Angus Media claims copyright to this web site as presented. We welcome educational venues and cattlemen to link to this site as a service to their audience.