Difference between revisions of "Intake and Feed Efficiency"

From BIF Guidelines Wiki
Line 86: Line 86:
  
 
== Dry Matter Intake vs Residual Feed Intake and Residual Gain ==
 
== Dry Matter Intake vs Residual Feed Intake and Residual Gain ==
Organizations producing genetic predictions for feed consumption and partial efficiency differ in the expression of the EPD.  While some EPD are expressed as measured daily dry matter intake (DMI), others are published as in an index form to quantify partial efficiency such as residual feed intake (RFI) and residual average daily gain (RADG).  
+
Organizations producing genetic predictions for feed consumption and partial efficiency differ in the expression of the EPD.  While some EPD are expressed as measured daily dry matter intake (DMI), others are published in index form to quantify partial efficiency such as residual feed intake (RFI) and residual average daily gain (RADG).  
 
<blockquote>
 
<blockquote>
 
===RFI and RADG===
 
===RFI and RADG===
Line 94: Line 94:
 
EPDs produced for DMI are produced simply by fitting an analytical model that does not adjust for genetically correlated covariables.  Instead, these other traits may serve as indicators of DMI in a multiple trait model.  This is analogous to how all other traits in the genetic evaluations are considered.   
 
EPDs produced for DMI are produced simply by fitting an analytical model that does not adjust for genetically correlated covariables.  Instead, these other traits may serve as indicators of DMI in a multiple trait model.  This is analogous to how all other traits in the genetic evaluations are considered.   
 
</blockquote>
 
</blockquote>
Kennedy et al. (1993)<ref>Kennedy, B. W., J. H. J. van der Werf, and T. H. E. Meuwissen. 1993. Genetic and statistical properties of residual feed intake. J.Anim. Sci. 71:3239–3250.</ref> showed the equivalence of [[Selection Index | selection indexes]] that incorporated intake or RFI when the economic weights were calculated correctly.  Of course, this assumes the production of the RFI phenotype is performed sensibly when this method is used.  Additionally, the inclusion of RFI in selection indices is problematic.  By definition, RFI is not an [[Economically Relevant Traits | economically relevant trait]] given it only accounts for a portion of feed consumed and thus cannot be a sensible trait in an economically rational [[Breeding Objectives | breeding objective]].  All RFI EPD are produced using phenotypes obtained on breeding bulls and heifers and are, therefore, [[Indicator Traits | indicator traits]].  While likely correlated to the residual components of feedlot feed consumption, feed consumption pre-weaning, and mature cow feed consumption, useful covariance parameters are not available for the residual components.  These relationships are potentially available for DMI EPD produced from data from young breeding bulls and heifers.  Thus, RFI from tested breeding animals is, at best, difficult to use in an economically optimal selection index for commercial situations.  It has been argued that RFI should be published because not all producers use selection index methods.  However, this logic promotes sub-optimal selection practices including single-trait selection or at best two-trait selection methods, and the challenge of inclusion of RFI in selection indices has been previously pointed out.  Therefore, ''BIF recommends that if an EPD for growing animal intake and/or partial efficiency be published that DMI EPD be made available and not RFI and RADG EPD''. Moreover, BIF recommends that economic selection indices be made available to select for feed efficiency in an economic context with other appropriate economically relevant traits related to more comprehensive breeding objectives. RFI EPD can be obtained using an index that includes DMI EPD and the EPD of the RFI covariables.
+
Kennedy et al. (1993)<ref>Kennedy, B. W., J. H. J. van der Werf, and T. H. E. Meuwissen. 1993. Genetic and statistical properties of residual feed intake. J.Anim. Sci. 71:3239–3250.</ref> showed the equivalence of [[Selection Index | selection indexes]] that incorporated intake or RFI when the economic weights were calculated correctly.  Of course, this assumes the production of the RFI phenotype is performed sensibly when this method is used.  Additionally, the inclusion of RFI in selection indexes is problematic.  By definition, RFI is not an [[Economically Relevant Traits | economically relevant trait]] given it only accounts for a portion of feed consumed and thus cannot be a sensible trait in an economically rational [[Breeding Objectives | breeding objective]].  All RFI EPD are produced using phenotypes obtained on breeding bulls and heifers and are, therefore, [[Indicator Traits | indicator traits]].  While likely correlated to the residual components of feedlot feed consumption, feed consumption pre-weaning, and mature cow feed consumption, useful covariance parameters are not available for the residual components.  These relationships are potentially available for DMI EPD produced from data from young breeding bulls and heifers.  Thus, RFI from tested breeding animals is, at best, difficult to use in an economically optimal selection index for commercial situations.  It has been argued that RFI should be published because not all producers use selection index methods.  However, this logic promotes sub-optimal selection practices including single-trait selection or at best two-trait selection methods, and the challenge of inclusion of RFI in selection indexes has been previously pointed out.  Therefore, ''BIF recommends that if an EPD for growing animal intake and/or partial efficiency be published that DMI EPD be made available and not RFI and RADG EPD''. Moreover, BIF recommends that economic selection indexes be made available to select for feed efficiency in an economic context with other appropriate economically relevant traits related to more comprehensive breeding objectives. RFI EPD can be obtained using an index that includes DMI EPD and the EPD of the RFI covariables.
  
 
In both situations, DMI or RFI, data need to be evaluated for heterogeneity of variation due to alternative test periods, diets and feeding conditions.  Methods have been proposed for normalizing the variation<ref name="macneil" />
 
In both situations, DMI or RFI, data need to be evaluated for heterogeneity of variation due to alternative test periods, diets and feeding conditions.  Methods have been proposed for normalizing the variation<ref name="macneil" />

Revision as of 13:18, 2 June 2020

THIS PAGE CONTAINS DRAFT INFORMATION UNDER CONSIDERATION
FOR INCLUSION IN THE OFFICIAL GUIDELINES

There is no guarantee that this proposed revision will be included in the official guidelines. To view the current official version of the Guidelines please navigate to Intake and Feed Efficiency.

This information is intended to cover feed intake and feed efficiency in growing cattle, rather than for cows.

Background

The following is an updated addendum to the feed intake measurement recommendations in the 9th edition of the Beef Improvement Federation Guidelines. The key updates below focus on warm-up and testing periods, as well as newly suggested approaches to appropriate contemporary grouping for application to genetic evaluation, and the (de)coupling of gain and feed intake measures. Testing equipment and test diet guidelines were not revised from the original recommendations. Historically, measures of feed utilization incorporated both feed consumption and measures of body weight gain [1] [2] [3]; however, when expressed in a linear form (e.g. selection index) application of values for costs and returns results in outcomes more closely associated with net return (value – cost) and therefore a selection index considering both cow/calf performance, postweaning growth and carcass merit measures is likely optimum for the beef industry [4]. Application of either approach could result in genetic change in feed utilization. However, for genetic improvement programs selection should be based on EPD (or EBV) resulting from multiple-trait genetic evaluation of feed intake [5]. As such, recommendations below address both approaches where the desired phenotypes are simultaneous measurement of gain and feed intake with defined endpoints being start/end of the feed intake test (coupled) or measurement of feed intake with gain measured outside the bounds of the feed intake test (de-coupled). This article discusses feed intake for young animals. Feed intake for mature cows is discussed elsewhere.

Warm-up period recommendations

Accurate feed utilization testing in beef cattle is dependent on collecting reliable and sufficiently precise measures of daily feed intake and body weight gain. Measurement of both phenotypes is subject to some degree of error. Therefore, much care should be given to the development and implementation of testing procedures that systematically minimize the errors associated with measuring these two components. A warm-up or acclimation period of at least 21 days should be included in the feed intake and gain test protocol. The goal of the acclimation period is to reduce within contemporary group variation in feed intake or gain due to non-genetic factors (e.g. pre-test environment) and to acclimate animals to both the test diet and the testing equipment. Animals should be transitioned from receiving to test diets gradually to minimize digestive system upset. Frequently these transitions will include moving animals from a primarily roughage-based receiving/backgrounding diet (either fed ad libitum or with restricted intake) to a higher energy, concentrate-based growing diet fed ad libitum. If calves entering a de-coupled test protocol have been previously transitioned to a diet that will be used in the formal feed intake test, then the acclimation period may be substantially reduced (by a week or more) to accommodate acclimation to the feed intake equipment only. However, users should be cautious that animals acclimated to a high concentrate diet are not restricted from that diet while training to prevent acidosis issues when they return to normal consumption levels. The overall test period (warm-up and testing) may be reduced by adjusting the start date of the trial through observation of daily intake records. When intakes have stabilized across the entire group following a week or more of feed intake observations, the trial may begin. While adjustment of start date may practically save days on feed in the feed intake facilities, it typically won’t reduce the days on feed for growing or developing animals and incurs additional labor and data analysis costs to reliably determine the ‘start’ date of the test. In practice, it may be simpler and more reliable for producers and central tests to set a minimum warm-up period (21 days) as the standard operating test protocol.

Test period recommendations

From an industry-wide genetic improvement perspective where there are limited feed intake measuring facilities and resulting limitations to the number of animals upon which phenotypic measures can be collected, a trade-off between precision of measure and number of animals measured exists. For instance, given limited feed intake facilities, one might choose longer test periods for more precise measures of feed intake and gain, realizing fewer animals could be phenotyped on an annual basis. Alternatively, time periods for testing could be shortened, reducing precision of intake measures but increasing the numbers of animals phenotyped. From a cost perspective, the latter approach reduces costs on a per animal basis compared to the former. This trade-off is illustrated in Table 1, where shortened test lengths are increasingly less related to previous industry standard measurement period of 70 days. While not estimated in that study, genetic correlations between these measures are invariably higher.

Table 1. Average correlation and regression coefficients between standard (70-day) and varying shortened test period lengths. (adapted from [6])

Test Length ADMI Pearson Correlation RFI Pearson Correlation ADMI Regression RFI Regression--Bruce L. Golden (talk) 22:04, 7 May 2020 (UTC)
28 days 0.94 0.835 0.83 0.765
42 days 0.975 0.905 0.93 0.87
56 days 0.99 0.95 0.985 0.955

In a de-coupled test where the primary outcome is a measure of average daily feed intake, recommendations are for a 42-day test length. As previously outlined in the guidelines, days where animals are removed from the pen for any reason are not considered a valid measurement day and should not be included in the calculation of ADMI. A 42-day test length will typically ensure 35 days of reliable and precise feed intake measures—a minimum number of reliable days of feed intake measures. In tests that use coupled feed intake and gain, the testing period should continue at 70 days due to lower correlations with the 70-day test (Table 2). The expanded test period is necessary for more reliable measurements of body weight gain. Animals should be weighed twice on adjoining dates (once on each of two consecutive days) at end of acclimation period and used as a on-test weight. At a minimum, animals should then be weighed twice at the end of the testing period. A preferable approach would be to weigh animals every two weeks during the testing period and a regression approach used for calculation of ADG. However, the additional costs and loss of feed intake data for those days must be considered.

Table 2. Average correlation and regression coefficients between standard (70-day) and varying shortened test period lengths for average daily gain (ADG) and metabolic mid-weight (MMWT). (adapted from [6] )

Test Length ADG Pearson Correaltion MMWT Pearson Correlation ADG Regression MMWT Regression
28 days 0.66 0.975 0.275 0.92
42 days 0.835 0.99 0.525 0.98
56 days 0.945 1 0.825 1.005


Contemporary grouping for national cattle evaluation

The most efficient use of feed intake and gain measures to achieve genetic progress are in national cattle evaluation with EPD(EBV) delivered to breeders for use in selection decisions. Given the sparse nature of feed intake data (in comparison to birth, weaning, and yearling weights), alternative approaches are suggested for contemporary grouping to maximize the value of feed intake measures in NCE. Feed intake test contemporary groups should consist of the weaning contemporary group (association defined) in addition to any post weaning treatment differences including ‘on test’ date, test or acclimation duration, test location and/or test equipment, and minimum age on test. The combination of these factors into a single feed intake contemporary group should result in a relatively equal pre-test environment resulting in equal ability to express genetic differences during the test. Given testing stations may combine animals from multiple sources and weaning contemporary groups in a single test pen, recommendations are to fit test group (pen, period/test, diet) as a separate effect in the model in addition to fitting weaning contemporary group as a separate effect. This approach should increase the value of these measures in genetic evaluation by maximizing contemporary group size while still attempting to account for environmental differences. Thus, in the evaluation there will be fewer class effects to account for overall. An alternative and potentially better approach would be to include weaning and test contemporary groups in the genetic evaluation as a random effect (as opposed to fixed effect) to regress effects relative to the information content of the pen. Given pens will often be confounded with genetics of the animal (due to animals from the same source being grouped), this is likely the best compromise in model fitting. Appropriate contemporary grouping insures that the within group performance differences are minimally affected by differences in non-genetic factors thereby reducing bias in genetic predictors.

Dry Matter Intake vs Residual Feed Intake and Residual Gain

Organizations producing genetic predictions for feed consumption and partial efficiency differ in the expression of the EPD. While some EPD are expressed as measured daily dry matter intake (DMI), others are published in index form to quantify partial efficiency such as residual feed intake (RFI) and residual average daily gain (RADG).

RFI and RADG

Phenotypic-based RFI attempts to adjust observed intake for phenotypically correlated sources of variation, so RFI is not correlated with indicator traits. Most commonly these include gain and metabolic mid-weight, although measures of body composition have also been used. This process creates a restricted selection index based on phenotypes, whereby selection for RFI will reduce intake without changing gain. Alternately, RADG is a restricted index that allows change in gain whilst holding feed intake constant. To generate EPD for RFI, two alternative methods have been proposed. A phenotypic-based RFI can be calculated using estimated relationships between maintenance requirements, and anticipated requirements for growth and fat deposition, and this phenotype becomes the dependent variable in the genetic evaluation. More commonly, the DMI phenotype is a dependent variable in a model that includes correlated factors as covariables - e.g., weight, gain, fat thickness, etc. The second approach ensures that the resulting genetic prediction of RFI is genetically independent of the covariates included in the model. The same issues and approaches exist for producing genetic predictions of RADG.

DMI

EPDs produced for DMI are produced simply by fitting an analytical model that does not adjust for genetically correlated covariables. Instead, these other traits may serve as indicators of DMI in a multiple trait model. This is analogous to how all other traits in the genetic evaluations are considered.

Kennedy et al. (1993)[7] showed the equivalence of selection indexes that incorporated intake or RFI when the economic weights were calculated correctly. Of course, this assumes the production of the RFI phenotype is performed sensibly when this method is used. Additionally, the inclusion of RFI in selection indexes is problematic. By definition, RFI is not an economically relevant trait given it only accounts for a portion of feed consumed and thus cannot be a sensible trait in an economically rational breeding objective. All RFI EPD are produced using phenotypes obtained on breeding bulls and heifers and are, therefore, indicator traits. While likely correlated to the residual components of feedlot feed consumption, feed consumption pre-weaning, and mature cow feed consumption, useful covariance parameters are not available for the residual components. These relationships are potentially available for DMI EPD produced from data from young breeding bulls and heifers. Thus, RFI from tested breeding animals is, at best, difficult to use in an economically optimal selection index for commercial situations. It has been argued that RFI should be published because not all producers use selection index methods. However, this logic promotes sub-optimal selection practices including single-trait selection or at best two-trait selection methods, and the challenge of inclusion of RFI in selection indexes has been previously pointed out. Therefore, BIF recommends that if an EPD for growing animal intake and/or partial efficiency be published that DMI EPD be made available and not RFI and RADG EPD. Moreover, BIF recommends that economic selection indexes be made available to select for feed efficiency in an economic context with other appropriate economically relevant traits related to more comprehensive breeding objectives. RFI EPD can be obtained using an index that includes DMI EPD and the EPD of the RFI covariables.

In both situations, DMI or RFI, data need to be evaluated for heterogeneity of variation due to alternative test periods, diets and feeding conditions. Methods have been proposed for normalizing the variation[5]

Impacts of Changing Technologies

New remote-sensing technologies continue to be developed and older technologies improved such as automated animal weighing systems and ear tags monitoring feeding behavior. Given the rapid advancements, the guidelines for measuring individual feed intake and gain, will likely need review on an ongoing basis. These technologies will likely result in changes to the current recommendations.

References

  1. Koch, R. M., L. A. Swiger, D. Chambers, and K. E. Gregory. 1963. Efficiency of feed use in beef cattle. J. Anim. Sci. 22:486-494. doi:10.2527/jas1963.222486x.
  2. Dickerson, G. E., N. Kunzi, L. V. Cundiff, R. M. Koch, V. H. Arthaud, and K. E. Gregory. 1974. Selection criteria for efficient beef production. J. Anim. Sci. 39:659-673. doi:10.2527/jas1974.394659x.
  3. Berry, D. P., and J. J. Crowley. 2012. Residual intake and body weight gain: A new measure of efficiency in growing cattle. J. Anim. Sci. 90:109–115. doi:10.2527/jas.2011-4245.
  4. Nielsen, M. K., M. D. MacNeil, J. C. M. Dekkers, D. H. Crews Jr., T. A. Rathje, R. M. Enns, and R. L. Weaber. 2013. Review: Life-cycle, total industry genetic improvement of feed efficiency in beef cattle: Blueprint for the Beef Improve--~~~~ment Federation. Prof. Anim. Sci. 29:559–565.
  5. 5.0 5.1 MacNeil, M. D., N. Lopez-Villalobos, and S. L. Northcutt. 2011. A prototype national cattle evaluation for feed intake and efficiency of Angus cattle. J. Anim. Sci. 89:3917-3923. doi:10.2527/jas.2011-4124.
  6. 6.0 6.1 Culbertson, M. M., S. E. Speidel, R. K. Peel, R. R. Cockrum, M. G. Thomas, and R. M. Enns. 2015. Optimum measurement period for evaluating feed traits in beef cattle. J. Anim. Sci. 93:2482-2487. doi:10.2527/jas.2014-8364
  7. Kennedy, B. W., J. H. J. van der Werf, and T. H. E. Meuwissen. 1993. Genetic and statistical properties of residual feed intake. J.Anim. Sci. 71:3239–3250.