Phenotypic means for pGFE and ECM
Gross feed efficiency (GFE) is an exceptionally important trait in dairy production, due to its impact on profitability and environmental sustainability. Energy-corrected milk is a major predictor trait for GFE, and is also of great economic importance. Besides genetic variation between populations, differences in measurement or prediction methods may account for discrepancies in mean GFE among studies (Köck et al. 2018; Tarekegn et al. 2021; Becker et al. 2022). On the other hand, mean ECM is based on yield of milk, which is invariably measured directly, using standard measuring devices (Kirchgeßner 1997).
Previous studies (Batch et al. 2006; Ishler and Heinrichs 2016) reported relatively higher mean values for actual GFE than those obtained for pGFE in the current study. However, the observation that cows were more efficient in early compared to later stages of lactation concurs with earlier research on intensively-fed Holstein cows elsewhere (Batch et al. 2006; Ishler and Heinrichs 2016). As expected from the normal lactation curve, higher daily yields of ECM were also produced in early than later lactation stages. It has been noted that cows in early lactation are more feed efficient because they mostly utilise their body reserves to derive energy for milk production, which causes an artificial increase in gross feed efficiency (Ishler and Heinrichs 2016; Ledinek et al. 2019). On the other hand, late-lactation cows will be gaining weight; thus lowering their calculated gross feed efficiency. The reduced gross feed efficiency in late lactation should, however, not be viewed negatively because cows need to regain body weight in late lactation, in order to have adequate body reserves for the next lactation. Exceptionally high gross feed efficiency in early lactation may, however, indicate that cows are losing too much weight, which might lead to metabolic disorders (Ishler and Heinrichs 2016; Ledinek et al. 2019).
Mean daily pGFE in first lactation was lower than values for actual GFE observed in other recent studies, despite mean daily ECM yields being comparable (Byskov et al. 2017; Li et al. 2018; Krattenmacher et al. 2019). Although mean daily pGFE was the same (1.26) for primiparous and multiparous cows, multiparous cows produced an average of 2.62 kg/day more ECM. Spurlock et al. (2012) also observed similar means for actual GFE of primiparous and multiparous American Holstein cows. The overall mean for daily pGFE across lactations was lower compared to values reported recently for actual GFE in Austrian, German and Swedish Holstein cattle (Köck et al. 2018; Tarekegn et al. 2021; Becker et al. 2022). Cows in the present study also produced relatively lower daily ECM on average than Austrian, German and Swedish Holstein cattle (Köck et al. 2018; Tarekegn et al. 2021; Becker et al. 2022).
Heritability estimates for pGFE and ECM
Heritability estimates for pGFE
A central objective of the current study was to assess the extent to which pGFE exhibits genetic variation and, hence, determine its suitability as a selection criterion for feed efficiency. The heritability estimates for pGFE within stages of lactation and across lactations were low to moderate, indicating scope for modest genetic improvement through selection. Late lactation had the second highest heritability, with the estimate for early lactation being marginally higher than that for mid lactation. Spurlock et al. (2012) also observed moderate albeit larger heritabilities for actual GFE in the first half of lactation, and a higher estimate in early compared to mid-lactation, in a study on American Holstein cattle. In further agreement with Spurlock et al. (2012), heritability was higher in primiparous than multiparous cows. Spurlock et al. (2012) analysed actual measured GFE in the first and second halves of lactation, which may partially explain the disparity in the magnitude of estimates from those of the current study. Additionally, Spurlock et al. (2012) used random regression models, which are better at modelling genetic and environmental variances along the lactation trajectory than repeatability models (Dzomba et al. 2010).
The heritability estimate for pGFE across lactations falls within the range of 0.10 ± 0.03 to 0.18 ± 0.03 obtained for actual GFE in Austrian dairy cattle (Köck et al. 2018). However, since this estimate is comparatively lower than the value observed for the late lactation stage, strategic selection based on measurements recorded only in late lactation may be more effective than considering the entire lactation. Due to the higher heritability of pGFE in primiparous compared to multiparous cows, selection considering first parity records only also appears to be justifiable. Thus, stage of lactation and parity should be taken into consideration when incorporating pGFE in the selection objective.
Heritability estimates for ECM
Energy-corrected milk is an important component of the feed efficiency trait complex; hence knowledge of its genetic attributes is essential to the inclusion of GFE in the selection objective. The heritability of ECM yield has been fairly studied in recent years, mainly based on first lactation records, and most of the estimates obtained were moderate (e.g. Manzanilla-Pech et al. 2014; Li et al. 2018; Krattenmacher et al. 2019). In general, these estimates are larger than the low values obtained in the current study. Manzanilla-Pech et al. (2014) found relatively larger values (>0.19), with higher estimates in mid compared to early and late lactation, in Dutch Holstein cattle. On the contrary, we observed higher estimates in late than mid and early lactation. Higher heritability estimates, which were larger in early than mid and late lactation, have also been reported in first-parity Holstein populations elsewhere (Li et al. 2018; Krattenmacher et al. 2019). Thus, there appears to be no consistency among studies on the relative magnitude of heritability of ECM by stage of lactation.
Disparities in the heritability of milk production traits between parities is well documented in the literature (e.g. Haile-Mariam and Pryce 2017; Meseret and Negussie 2017; Buaban et al. 2020; Tarekegn et al. 2021). In the current study, we obtained higher heritabilities for ECM in primiparous compared to multiparous cows, which is in agreement with Spurlock et al. (2012). Spurlock et al. (2012), however, found much larger estimates (>0.24) using random regression models.
A low heritability estimate for ECM across the first three lactations was also observed by Köck et al. (2018) in Austrian Holstein cattle, and suggests selection should be based on parity.
Genetic correlations and repeatability estimates for pGFE and ECM
Genetic correlations for pGFE
Genetic correlations between pGFE in different stages of lactation were estimated to determine if selection applied in one stage will result in improvement along the entire lactation. These correlations were positive and substantially large, supporting earlier findings by Spurlock et al. (2012) who observed a genetic correlation of 0.96 ± 0.18 between early and mid-lactation, for actual GFE in American Holstein cattle. These results suggest that pGFE in different stages of lactation is essentially an expression of the same trait. This further supports the idea to base selection only on data recorded in late lactation.
The genetic correlation between pGFE in primiparous and multiparous cows was also positive and extremely high (close to unity), suggesting that selection based on first lactation data will result in improvement in later lactations. Such an approach is further justified by our finding that the heritability of pGFE is higher in primiparous than multiparous cows.
Correlations between pGFE and ECM
Knowledge of the genetic correlations between pGFE and ECM may assist in improving accuracy of selection of pGFE, as well as incorporating it in the selection objective. These estimates were all positive and substantially high (>0.90) , within stages of lactation and across parities, confirming earlier findings by Köck et al. (2018) who reported strong positive genetic correlations between actual GFE and ECM across lactations in Australian Holstein cattle. Spurlock et al. (2012) also noted that improved GFE was closely associated genetically with increased ECM yield throughout the first half of lactation in American Holstein cattle. These results indicate that the two traits may be under the influence of the same or linked genes, and selection for higher ECM yield is likely to result in a correlated improvement in pGFE. More importantly, accuracy of selection for pGFE can be increased through multiple-trait analysis including ECM.
Repeatability estimates for pGFE and ECM
Repeatability was estimated to assess the extent to which repeated measures of pGFE and ECM across lactations are under the influence of permanent effects. Repeatability was moderate (0.37 ± 0.01 to 0.52 ± 0.02) for both pGFE and ECM, in agreement with a previous study by Köck et al. (2018) on Austrian Holstein cattle. Much higher repeatability estimates for ECM (>0.75) were, however, reported in recent studies on first lactations of Holstein cattle populations elsewhere (Byskov et al. 2017; Krattenmacher et al. 2019). It therefore appears that pGFE and ECM in first lactation is a fairly reliable indicator of performance in later lactations. Thus, culling decisions on pGFE or ECM may be made using only first lactation data.
Genetic trend for predicted gross feed efficiency
Genetic trend for pGFE, across all lactations, was ascertained to assess if there have been any changes in genetic merit for the trait in recent years, in the South African Holstein cattle population. Such a change may occur as a correlated response to selection for other traits with which it is genetically correlated. There has been sustained genetic selection for yield traits in the South African Holstein cattle population (Ramatsoma et al. 2014), which has unfortunately resulted in a correlated deterioration in functional traits (Banga et al. 2014). The current study found a marginal increase in genetic merit for daily pGFE over the period 2007-2017, which may also be a correlated response to selection for yield traits. This is plausible, given the high positive genetic correlation that we observed between pGFE and ECM. Other researchers (Spurlock et al. 2012; Köck et al. 2018) also noted a correlated genetic improvement of actual GFE due to an increase in milk production traits and a decrease in live weight. Thus, the exclusive focus on selection for yield traits in South African Holstein cattle has, fortunately, not been detrimental to feed efficiency. There is, however, a need to achieve more meaningful genetic improvement of feed efficiency by including it in the breeding objective.