Feed Conversion

Page Index: Identifying differences in feed efficiency among group-fed cattle Use of a mathematical computer model to predict feed intake Genetic Prediction of Efficiency in the Future Using simulation models to predict feed intake Use of a mathematical computer model to predict feed intake in Angus cattle

Identifying differences in feed efficiency among group-fed cattle:

IMPLICATIONS
A growth model that accounts for the principal sources of variation in energy requirements can be used to accurately identify differences in feed-to-gain and gain-to-feed ratios between individually growing cattle fed in groups. In this study, the Cornell Value Discovery System-predicted dry matter required to average daily gain ratio accounted for 84% of the variation in actual feed conversion (feed:gain) with a mean bias of 1.94%. Feed-to-gain or gain-to-feed ratio comparisons should be made over the same stage of growth to avoid bias due to variation in proportion of fat in the gain during the test period. Further improvements can be made in our model to account for more of the variation in observed feed-to-gain and gain-to-feed ratios, requiring data from experiments with individually fed animals in which observed dry matter intake, accurate feed characterization, composition of growth, and other animal information are available.

Read full report: Click Here

Source: L. O. Tedeschi, 1 D. G. Fox, M. J. Baker, and D. P. Kirschten

623 Use of a mathematical computer model to predict feed intake:

Genetic parameters between observed and predicted values, and relationships with other traits.

The objectives of this study were to investigate the suitability of using DM required (DMR) as predicted by the Cornell Value Discovery System (CVDS) in genetic evaluations and to determine relationships between model predicted and individual DMI and other traits. Observed feed intake (FIo) records during the ?nishing phase were available from 115 individually fed Santa Gertrudis steers sired by 20 bulls. The data also contained records of ADG and mean BW (MW), carcass measurements (ribeye area, fat thickness, and marbling) and real-time ultrasound estimates (ribeye area, fat thickness, and marbling).

These inputs were used in the CVDS model to predict DMR. For the purposes of parameter estimation, CVDS predictions of DMR using ultrasound (DMRus) and carcass traits (DMRca) were considered surrogates for FIo. Genetic parameters were estimated with REML, using a sire model with ?xed effects of WW contemporary group and feedlot pen. Phenotypic correlations between FIo and DMRus, FIo and DMRca and DMRus and DMRca were 0.78, 0.79, and 0.99, respectively. Heritabilities for FIo, DMRus and DMRca were 0.09, 0.32 and 0.35, respectively.

Genetic correlations between FIo and DMRus, FIo and DMRca, and DMRus and DMRca were 0.98, 0.98, and 0.99, respectively. Sire BV rank correlations were calculated for FIo, DMRus and DMRca. BV rank correlations among FIo, DMRus and DMRca were all 0.99. Residual feed intake (RFI) was calculated using FIo, metabolic MW and ADG. Phenotypic correlations between RFI and FIo, MW, and ADG were 0.21, 0.09, and 0.01, respectively. Heritability for RFI was 0.18. Genetic correlations between RFI and FIo, MW, and ADG were -0.88, -0.48, and 0.00, respectively. The strong genetic relationships between FIo, DMRus and DMRca and minimal re-ranking of sires suggested that predicted DMRus and DMRca may be used in place of FIo in genetic evaluations.

Key Words: Feed intake, Mathematical models, Beef cattle

Source: J. Anim. Sci. Vol. 84, Suppl. 1/J. Dairy Sci. Vol. 89, Suppl. 1
D. P. Kirschten, E. J. Pollak1, L. O. Tedeschi2, D. G. Fox1, B. Bourg2, and G. E. Carstens2,
1Cornell University, Ithaca, NY, 2Texas A&M University, College Station.

Genetic Prediction of Efficiency in the Future:

A U.S. Perspective
As an example, we will use the Cornell Value Discovery System (CVDS) model for prediction of individual feed consumption (Tedeschi et al., 2002). Models require inputs. In the case of the CVDS, the inputs are gain on test and carcass measures, ration ingredient analysis, and environmental factors (i.e., temperature, windspeed, lot conditions, and description of facilities).

This modeling approach can also be used for performance tests of potential breeding bulls using ultrasound in place of carcass measures (see Appendix A). The CVDS model has been validated in experimental environments, (Fox et al., current proceedings), and the NBCEC is now running a pilot study with the American International Charolais Association to collect information on performance-tested bulls (see Appendix B).

This pilot will be used to estimate parameters for genetic evaluation. The EPDs produced by the genetic evaluation will then be validated. This would include studies similar to those undertaken to validate EPDs for traits such as maternal milk. This process is potentially iterative. Shortcomings in the models that are identified by the genetic studies could be addressed to improve the predictability of the phenotypic prediction models. Obviously other technologies will come along that may enhance data capture. It is just difficult to envision those technologies addressing the ERT directly. They will most likely enhance the predictability of the phenotypes.

In conclusion, we believe that the future for genetic prediction of efficiency will necessitate a closer relationship between those working with biologically based models for predicting phenotypes and those implementing genetic evaluation programs. The industry would be required to provide the second basic component, which is pedigree information. For performance-tested breeding males and females, this is not an issue, but if we want to evaluate animals closer to the end product, strategies for parent identification in commercial herds will need to be developed.

Read full report: Click Here

Source: E. John Pollak and David Kirschten
Cornell University

Using simulation models to predict feed intake:
Phenotypic and genetic relationships between observed and predicted values in cattle

RESULTS AND DISCUSSION
Summary of Observed and Predicted Data
The mean, CV, minimum and maximum values for OFI of steers, and predicted measures of average daily required DMI are shown in Table 3. Compared with the OFI mean, the DFRmcg mean was very similar, but the CFRmcg mean was about 3.5% lower. Mean values and CV for DFRg and CFRg were about the same, suggesting that the CVDS and DECI performed about the same in predicting the average daily required DMI for gain. The difference in means between DFRmcg and CFRmcg is mainly due to the difference in means for DFRm and CFRm because the difference in requirements for cold stress (DFRmc -?DFRm; CFRmc -?CFRm) were small. These results suggest that the CVDS model may be underpredicting maintenance requirements compared with the DECI model. Predicted average daily required DMI variables showed a similar variability compared with OFI, except for average daily required DMI for gain, and this agrees with a CV of 17% for ADG in the observed data.

Figure 1. Relationship between observed average daily DMI (OFI) and average daily required DMI for maintenance, cold stress, and ADG of steers predicted using the Cornell Value Discovery System. Dashed line y = x indicates the position of the perfect fit between observed and model predicted values. The data used were from 502 individually fed steers produced in a heterosis experiment (Olson et al., 1978a,b). Read full report: Click Here

Source: C. B. Williams, G. L. Bennett, T. G. Jenkins, L. V. Cundiff, and C. L. Ferrell
USDA, ARS, US Meat Animal Research Center, Clay Center, NE

939 Use of a mathematical computer model to predict feed intake in Angus cattle:

Genetic parameters between observed and predicted values, and relationships with other traits.

The objectives of this study were to investigate the suitability of using dry matter required (DMR) as predicted by the Cornell Value Discovery System (CVDS) in genetic evaluations and to determine relationships between model predicted and individual DMI and other traits. Group 1 (659 ?nishing steers) and group 2 (309 yearling bulls) had observed feed intake (FI) records. Group 3 (1586 yearling bulls and heifers) had pedigree ties to the other datasets, but did not have FI data. The data also contained records of ADG and body weight (BW).

Two predictions of DMR were made with the CVDS model iterating on BW (DMR-W) and ADG (DMR-G). For purposes of parameter estimation, CVDS DMR predictions were considered surrogates for FI. Genetic parameters were estimated with MTDFREML, using an animal model with ?xed effects of weaning weight contemporary group and pen. Phenotypic correlations between FI and DMR-W and DMR-G were 0.69 and 0.71. The phenotypic correlation between DMR-W and DMR-G was 0.98. Genotypic correlations between FI and DMR-W and DMR-G were 0.79 and 0.85. The genetic correlation between DMR-W and DMR-G was 0.97. Heritabilities for FI, DMR-W and DMR-G were 0.42, 0.31 and 0.33, respectively.

Genetic correlations between ADG and FI, DMR-W and DMR-G were 0.45, 0.80, and 0.83, respectively. Genetic correlations between mean body weight (MW) and FI, DMR-W, and DMR-G were 0.50, 0.64 and 0.58, respectively. Residual feed intake (RFI) was calculated using FI, metabolic MW (MW0.75) and ADG. The heritability of RFI was 0.36. The phenotypic correlation between RFI and FI was 0.57. Phenotypic correlations between RFI and MW and ADG were not estimated. Genetic correlations between RFI and FI, MW, and ADG were 0.77, 0.09, and 0.01, respectively. Heritabilities of ADG and MW were 0.27 and 0.48. Standard errors for all genetic correlations were less than 0.06.

The genetic relationships between FI, DMR-W and DMR-G suggest that CVDS predictions of FI may be used as surrogates for actual FI in genetic evaluations.

Key Words: Feed Intake, Mathematical Models, Beef Cattle

Source: J. Anim. Sci. Vol. 85, Suppl. 1/J. Dairy Sci. Vol. 90, Suppl. 1/Poult. Sci. Vol. 86, Suppl. 1
D. P. Kirschten, E. J. Pollak, and D. G. Fox, Cornell University, Ithaca, NY.

Back to Top