Single-step evaluation for milking cow survival in Poland

In Poland, the current genetic evaluation of dairy cow survival is performed using Survival Kit with sire model. Genomic evaluation is implemented using a two-step approach. The Centre for Genetics of Polish Federation of Cattle Breeders and Dairy Farmers together with National Research Institute of Animal Production in Balice are currently undertaking the review of the Polish genetic evaluation system for all traits to implement single-step genomic evaluations using BLUPF90 family of programs. The goal of this research was to develop single-step evaluation of cow survival. The following approaches to defining cow survival phenotype were considered: 1) length of productive life in months form first calving to culling (with and without dry period included, continuous); 2) survival from one calving to the next (binary) implemented as either four-trait model (survival from 1st to 2nd, 2nd to 3rd, 3rd to 4th and 4th to 5th calving) or as a repeated records model (up to ten parities included); 3) survival to a given day in milk during lactation (binary) with lactations divided into three parts based on when most culling for specific reasons occurs (1-74, 75-250, and >250 to next calving days in milk). This approach was implemented using a nine-trait model (first 3 lactations split into 3 periods each) or a three-trait repeated records model (each part of lactation as separate repeated records trait in up to ten parities); 4) random regression model with survival defined per month from first calving up to 72 months; 5) number of completed lactations (categorical treated as continuous); 6) number of days survived within each lactation (continuous). Variance components were estimated for all phenotypes, and alternative modeling approaches were tested, with a primary focus on assessing the feasibility of correcting for levels of milk production in the model. Next, the list of phenotypes of interest was narrowed down to options 1-3 listed above. For those phenotypes, both conventional pedigree-based evaluations and single-step evaluations were performed using BLUPF90 with the APY approach. Formal validation was carried out for all runs, including the Interbull trend test and Mendelian sampling test. This paper will present the results of the validation work which leads to the choice of the cow survival phenotype for single-step implementation.