International genetic evaluations for feed intake in dairy cattle through the collation of data from multiple sources

Abstract

Feed represents a large proportion of the variable costs in dairy production systems. The omission of feed intake measures explicitly from national dairy cow breeding objectives is predominantly due to a lack of information from which to make selection decisions. However, individual cow feed intake data are available in different countries, mostly from research or nucleus herds. None of these data sets are sufficiently large enough on their own to generate accurate genetic evaluations. In the current study, we collate data from 10 populations in 9 countries and estimate genetic parameters for dry matter intake (DMI). A total of 224,174 test-day records from 10,068 parity 1 to 5 records of 6,957 cows were available, as well as records from 1,784 growing heifers. Random regression models were fit to the lactating cow test-day records and predicted feed intake at 70 d postcalving was extracted from these fitted profiles. The random regression model included a fixed polynomial regression for each lactation separately, as well as herd-year-season of calving and experimental treatment as fixed effects; random effects fit in the model included individual animal deviation from the fixed regression for each parity as well as mean herdspecific deviations from the fixed regression. Predicted DMI at 70 d postcalving was used as the phenotype for the subsequent genetic analyses undertaken using an animal repeatability model. Heritability estimates of predicted cow feed intake 70 d postcalving was 0.34 across the entire data set and varied, within population, from 0.08 to 0.52. Repeatability of feed intake across lactations was 0.66. Heritability of feed intake in the growing heifers was 0.20 to 0.34 in the 2 populations with heifer data. The genetic correlation between feed intake in lactating cows and growing heifers was 0.67. A combined pedigree and genomic relationship matrix was used to improve linkages between populations for the estimation of genetic correlations of DMI in lactating cows; genotype information was available on 5,429 of the animals. Populations were categorized as North America, grazing, other low input, and high input European Union. Albeit associated with large standard errors, genetic correlation estimates for DMI between populations varied from 0.14 to 0.84 but were stronger (0.76 to 0.84) between the populations representative of high-input production systems. Genetic correlations with the grazing populations were weak to moderate, varying from 0.14 to 0.57. Genetic evaluations for DMI can be undertaken using data collated from international populations; however, genotype-by-environment interactions with grazing production systems need to be considered.

Reducing the carbon footprint of Australian milk production by mitigation of enteric methane emissions

“…For example, de Haas et al (2012) combined DMI phenotypes from Dutch and UK cows with Australian heifer phenotypes and found that the accuracy of genomic prediction was 5.5% higher when a multi-country reference population was used, than with single-trait models. Since then, there has been further international collaboration through the global Dry Matter Initiative (gDMI) to build an even larger reference population (Berry et al 2014). Initial results on genomic prediction using this reference population look promising.…”

Genomic prediction of dry matter intake in dairy cattle from an international data set consisting of research herds in Europe, North America, and Australasia

“…For such difficult and costly-to-measure traits, there is growing interest in combining data from international research populations for genetic analysis de Haas et al, 2012;Veerkamp et al, 2012;Berry et al, 2014). The justification for combining genotypes and phenotypes from different research organizations includes greater statistical power for genome-wide association studies as well as improved accuracy of genetic or genomic predictions.…”

Genetic and phenotypic associations of type traits and body condition score with dry matter intake, milk yield, and number of breedings in first lactation Canadian Holstein cows

“…Vallimont et al (2010) also reported heritability estimates for 305-d feed intake traits ranging from 0.15 to 0.18, which are close to the estimates obtained in the present study. Berry et al (2014), using feed data collected experimentally, found heritability estimates of 0.11 and 0.19 for feed intake in the first 70 d of lactation in Canadian Holsteins under two different statistical models. They found an overall heritability estimate of feed intake of 0.34 when experimental data were collated from nine countries.…”

“…Another possibility is to consider feed intake of cows in a genetic selection program. Feed intake in dairy cattle has moderate heritability (van Elzakker and van Arendonk 1993;Ageeb and Hayes 2006;Banos and Coffey 2010;Berry et al 2014), and therefore, improvement in feed intake through genetic and (or) genomic selection is possible . Feed intake might be considered in a selection index using appropriate weights to improve biological efficiency and to help reduce health and fertility problems in dairy cows.…”