LAUSR

There is growing interest in improving feed efficiency (FE) traits in dairy cattle. The objectives of this study were to estimate the genetic parameters of RFI and its component traits (Dry matter intake (DMI), metabolic body weight (MBW) and average daily gain (ADG)) of Holstein heifers, and to develop a system for genomic evaluation for RFI in Holstein dairy calves. The RFI data were collected from 6,563 growing Holstein heifers (initial BW = 261 ± 52 kg; initial age = 266 ± 42 d) for 70 d, across 182 trials conducted between 2014 and 2022 at the STgenetics Ohio Heifer Center (South Charleston, Ohio) as part of EcoFeed® program which aims to improve FE by genetic selection. RFI was estimated as the difference between a heifer's actual feed intake and its expected feed intake, which was determined by regression of DMI against mid-point MBW, age, and ADG across each trial. A total of 61,283 single nucleotide polymorphisms were used in genomic analyses. Animals with phenotypes and genotypes were used as training population and 4 groups of prediction population, each with 2,000 animals, were selected from a pool of Holstein animals with genotypes, based on their relationship with the training population. All traits were analyzed using univariate animal model in DMU version 6 software. Pedigree information and genomic information were used to specify genetic relationships to estimate the variance components and genomic estimated breeding values (GEBVs), respectively. Breeding values of prediction population were estimated by using the 2-step approach, i.e., deriving the prediction equation of GEBVs from training population for the estimation of GEBVs of prediction population with only genotypes. Reliability of breeding values were obtained by approximation based on partitioning a function of accuracy of training population GEBVs and magnitudes of genomic relationships between individuals in the training and prediction population. Heifers had DMI (mean ± SD) of 8.11 ± 1.59 kg over the period of trial with growth rate of 1.08 ± 0.25 kg/d. The heritability estimates (mean ± SE) of RFI, MBW, DMI, and growth rate were 0.24 ± 0.02, 0.23 ± 0.02, 0.27 ± 0.02, and 0.19 ± 0.02, respectively. The range of genomic predicted transmitted abilities (gPTAs) of training population (-0.94 to 0.75) was higher compared with the range of gPTAs (-0.82 to 0.73) of different groups of prediction population. Average reliability of breeding values from training population was 58% and that of prediction population was 39%. The genomic prediction of RFI provided new tools to select for feed efficiency of heifers. Future research should be directed to find the relationship between RFI of heifers and cows, to select individuals based on their lifetime production efficiencies.