The impact of truncating data on the predictive ability for single‐step genomic best linear unbiased prediction

Abstract

Simulated and swine industry data sets were utilized to assess the impact of removing older data on the predictive ability of selection candidate estimated breeding values (EBV) when using single-step genomic best linear unbiased prediction (ssGBLUP). Simulated data included thirty replicates designed to mimic the structure of swine data sets. For the simulated data, varying amounts of data were truncated based on the number of ancestral generations back from the selection candidates. The swine data sets consisted of phenotypic and genotypic records for three traits across two breeds on animals born from 2003 to 2017. Phenotypes and genotypes were iteratively removed 1 year at a time based on the year an animal was born. For the swine data sets, correlations between corrected phenotypes (Cp) and EBV were used to evaluate the predictive ability on young animals born in 2016-2017. In the simulated data set, keeping data two generations back or greater resulted in no statistical difference (p-value > 0.05) in the reduction in the true breeding value at generation 15 compared to utilizing all available data. Across swine data sets, removing phenotypes from animals born prior to 2011 resulted in a negligible or a slight numerical increase in the correlation between Cp and EBV. Truncating data is a method to alleviate computational issues without negatively impacting the predictive ability of selection candidate EBV.

Genomic predictions for yield traits in US Holsteins with unknown parent groups

“…Truncation to about 2 generations of data did not reduce accuracy and sometimes improved it. Similar results were found by Howard et al (2018). Data truncation in dairy cattle may be affected by decreasing generation intervals (Mäntysaari et al, 2020), change of trait definitions under improved management (Tsuruta et al, 2005), and decay of genomic information with selection (Muir, 2007).…”

Genomic and conventional evaluations for fertility traits in pigs

“…At present, the landscape of GS is changing: animals are being routinely genotyped and phenotypes for most traits are being collected (Howard et al, 2018). Genotypes within a species continue to increase in number and the relationship of recent selection candidates to the majority of the historic population is becoming more distant.…”

“…Genotypes within a species continue to increase in number and the relationship of recent selection candidates to the majority of the historic population is becoming more distant. Moreover, improvements in phenotype collections, changes in genetic architecture and/or changes in models across time create a situation, in which information from an older animal (in the reference set) has the potential to negatively impact the accuracy of selection candidates (Howard et al, 2018). The numbers of animals used for genomic predictions in this thesis were rather small (2,053 Large White animals in Chapter 2 and 495 (424) Landrace (Large White) animals in Chapter 3).…”

“…Thus, removing older data (truncating data) does not primarily have any particular practical relevance for the presented data (Chapter 2 and 3), but should be addressed with respect to the impact of using all available data on the accuracy for selection candidates. Howard et al (2018) assessed the impact of removing older data on simulation data (under idealized conditions without any pedigree or genotyping errors) and empirical pig data. For the simulation data, truncating was performed based on the ancestral generation number.…”

“…For the empirical data it was based on the year an animal was born, which is also the scenario with practical relevance. Howard et al (2018) had data on genotypes for Yorkshire (n=5,783)…”