A smarter tool for monitoring accuracy of genetic evaluation
AbacusBio consultants Grace Johnstone and Bram Visser have been working with Australian Dairy Herd Improvement Scheme (ADHIS) to develop a tool for monitoring the accuracy of genetic evaluation – among dairy bulls in Australia.
ADHIS recognises the fact that data is king, especially when it comes to providing world-class analytics for genetic improvement.
However, the key consideration of data analytics lies in its efficiency and accuracy.
The Integrity Diagnostic Tool for Australian Breeding Value Predictions has been integrated into ADHIS’s existing system to produce a suite of results comparing early genetic evaluations to more recent and reliable validation datasets across 34 traits and selection indices.
Written in R programming language, the tool is useful for identifying traits where sire performance deviates significantly from what was predicted in early breeding values and also determines whether any bull breed or country of origin deviates to a greater degree than others.
Some sires used in Australia have no or a few daughter observations contributing to their breeding value estimation.
Notably, this is the case for breeding values based on Interbull or genomic merit evaluation.
Grace remarks that “the performance of these breeding values are of particular interest. To enable farmers to make the best sire selection for their herd, it is vital these breeding values are comparable to those of traditional Australian progeny proven sires.”
Like New Zealand, Interbull breeding values are used in Australia for imported semen from international bulls, which have been proven overseas – but do not have adequate Australian records to generate an Australian breeding value.
As such, conversion factors are traditionally used to adjust the foreign Interbull breeding values to an Australian breeding value.
The genomic evaluations are generated for young unproven bulls, where the marketed breeding value is calculated based on pedigree information blended with genomic predictions of merit.
Genomic prediction involves the future prediction of genomic merit from the combination of genetic markers in the bulls’ DNA genomes.
“By allowing the industry to access high-merit bulls earlier than progeny testing schemes by a few years, genomic prediction has the potential to increase genetic gain rates significantly among international dairy herds. It also facilitates the better selection of bulls entering the traditional progeny testing schemes,” Bram says.
When these files are entered into the software tool, the user can have a comprehensive analysis of historic breeding value performance in just a few minutes.
“This allows ADHIS to monitor their performance, target areas for further development, and provide transparent information to the industry,” Grace adds.
A smart tool, for a group like ADHIS that is committed to continual improvement in genetic evaluation for the Australian dairy industry.