Limits...
Strategic test-day recording regimes to estimate lactation yield in tropical dairy animals.

McGill DM, Thomson PC, Mulder HA, Lievaart JJ - Genet. Sel. Evol. (2014)

Bottom Line: Comparisons between the different test-day sampling regimes showed that, with the same number of records per lactation (for example, quarterly and four test-days), strategically timed test-days can result in more accurate estimates of lactation yield than test-days at equal intervals.Furthermore, calculations show that although using fewer test-days results in a decrease in the accuracy of selection, it does provide an opportunity to progeny-test more sires.Thus, using strategically timed test-days and Wood's model to estimate lactation yield, can lead to a more efficient use of the allocated resources.

View Article: PubMed Central - PubMed

Affiliation: Graham Centre for Agricultural Innovation, Locked Bag 588, Wagga Wagga 2678, NSW, Australia. dmcgill@csu.edu.au.

ABSTRACT

Background: In developing dairy sectors, genetic improvement programs have limited resources and recording of herds is minimal. This study evaluated different methods to estimate lactation yield and sampling schedules with fewer test-day records per lactation to determine recording regimes that (1) estimate lactation yield with a minimal impact on the accuracy of selection and (2) optimise the available resources.

Methods: Using Sahiwal cattle as a tropical dairy breed example, weekly milk records from 464 cows were used in a simulation study to generate different shaped lactation curves. The daily milk yields from these simulated lactation curves were subset to equally spaced (weekly, monthly and quarterly) and unequally spaced (with four, five or six records per lactation) test-day intervals. Lactation yield estimates were calculated from these subsets using two methods: the test-interval method and Wood's (Nature 216:164-165, 1967) lactation curve model. Using the resulting lactation yields, breeding values were predicted and comparisons were made between the sampling regimes and estimation methods.

Results: The results show that, based on the mean square error of prediction, use of Wood's lactation curve model to estimate total yield was more accurate than use of the test-interval method. However, the differences in the ranking of animals were small, i.e. a 1 to 5% difference in accuracy. Comparisons between the different test-day sampling regimes showed that, with the same number of records per lactation (for example, quarterly and four test-days), strategically timed test-days can result in more accurate estimates of lactation yield than test-days at equal intervals.

Conclusions: An important outcome of these results is that combining Wood's model for lactation yield estimation and as few as four, five or six strategically placed test-day records can produce estimates of lactation yield that are comparable with estimates based on monthly test-day records using the test-interval method. Furthermore, calculations show that although using fewer test-days results in a decrease in the accuracy of selection, it does provide an opportunity to progeny-test more sires. Thus, using strategically timed test-days and Wood's model to estimate lactation yield, can lead to a more efficient use of the allocated resources.

Show MeSH
Probability density plots of the mean square error of prediction from Wood model lactation estimates. The probability density plots are shown separately for random selections of test-day (TD) sampling regimes of 4, 5 and 6 TD per lactation; plot (a) shows the trend for the simulated lactations with an average peak and high persistency (APHP) and (b) for the simulated lactations with a high peak and low persistency (HPLP); the line type represents the number (m) of TD per lactation used to estimate the lactation yield, where dotted (•••) is for m = 4, solid (―) for m = 5 and dashed (−−−) for m = 6.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4248470&req=5

Fig4: Probability density plots of the mean square error of prediction from Wood model lactation estimates. The probability density plots are shown separately for random selections of test-day (TD) sampling regimes of 4, 5 and 6 TD per lactation; plot (a) shows the trend for the simulated lactations with an average peak and high persistency (APHP) and (b) for the simulated lactations with a high peak and low persistency (HPLP); the line type represents the number (m) of TD per lactation used to estimate the lactation yield, where dotted (•••) is for m = 4, solid (―) for m = 5 and dashed (−−−) for m = 6.

Mentions: The key question in this study was how lactation yields estimated with Wood’s model [12] using fewer TD records compare with estimates from the recommended TIM method [9]. The plots in Figure 4 show the distribution of the MSEP values for the TDSR using four, five or six TD to estimate lactation yield. Lower values of MSEP indicate more accurate estimates of lactation yield.Figure 4


Strategic test-day recording regimes to estimate lactation yield in tropical dairy animals.

McGill DM, Thomson PC, Mulder HA, Lievaart JJ - Genet. Sel. Evol. (2014)

Probability density plots of the mean square error of prediction from Wood model lactation estimates. The probability density plots are shown separately for random selections of test-day (TD) sampling regimes of 4, 5 and 6 TD per lactation; plot (a) shows the trend for the simulated lactations with an average peak and high persistency (APHP) and (b) for the simulated lactations with a high peak and low persistency (HPLP); the line type represents the number (m) of TD per lactation used to estimate the lactation yield, where dotted (•••) is for m = 4, solid (―) for m = 5 and dashed (−−−) for m = 6.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4248470&req=5

Fig4: Probability density plots of the mean square error of prediction from Wood model lactation estimates. The probability density plots are shown separately for random selections of test-day (TD) sampling regimes of 4, 5 and 6 TD per lactation; plot (a) shows the trend for the simulated lactations with an average peak and high persistency (APHP) and (b) for the simulated lactations with a high peak and low persistency (HPLP); the line type represents the number (m) of TD per lactation used to estimate the lactation yield, where dotted (•••) is for m = 4, solid (―) for m = 5 and dashed (−−−) for m = 6.
Mentions: The key question in this study was how lactation yields estimated with Wood’s model [12] using fewer TD records compare with estimates from the recommended TIM method [9]. The plots in Figure 4 show the distribution of the MSEP values for the TDSR using four, five or six TD to estimate lactation yield. Lower values of MSEP indicate more accurate estimates of lactation yield.Figure 4

Bottom Line: Comparisons between the different test-day sampling regimes showed that, with the same number of records per lactation (for example, quarterly and four test-days), strategically timed test-days can result in more accurate estimates of lactation yield than test-days at equal intervals.Furthermore, calculations show that although using fewer test-days results in a decrease in the accuracy of selection, it does provide an opportunity to progeny-test more sires.Thus, using strategically timed test-days and Wood's model to estimate lactation yield, can lead to a more efficient use of the allocated resources.

View Article: PubMed Central - PubMed

Affiliation: Graham Centre for Agricultural Innovation, Locked Bag 588, Wagga Wagga 2678, NSW, Australia. dmcgill@csu.edu.au.

ABSTRACT

Background: In developing dairy sectors, genetic improvement programs have limited resources and recording of herds is minimal. This study evaluated different methods to estimate lactation yield and sampling schedules with fewer test-day records per lactation to determine recording regimes that (1) estimate lactation yield with a minimal impact on the accuracy of selection and (2) optimise the available resources.

Methods: Using Sahiwal cattle as a tropical dairy breed example, weekly milk records from 464 cows were used in a simulation study to generate different shaped lactation curves. The daily milk yields from these simulated lactation curves were subset to equally spaced (weekly, monthly and quarterly) and unequally spaced (with four, five or six records per lactation) test-day intervals. Lactation yield estimates were calculated from these subsets using two methods: the test-interval method and Wood's (Nature 216:164-165, 1967) lactation curve model. Using the resulting lactation yields, breeding values were predicted and comparisons were made between the sampling regimes and estimation methods.

Results: The results show that, based on the mean square error of prediction, use of Wood's lactation curve model to estimate total yield was more accurate than use of the test-interval method. However, the differences in the ranking of animals were small, i.e. a 1 to 5% difference in accuracy. Comparisons between the different test-day sampling regimes showed that, with the same number of records per lactation (for example, quarterly and four test-days), strategically timed test-days can result in more accurate estimates of lactation yield than test-days at equal intervals.

Conclusions: An important outcome of these results is that combining Wood's model for lactation yield estimation and as few as four, five or six strategically placed test-day records can produce estimates of lactation yield that are comparable with estimates based on monthly test-day records using the test-interval method. Furthermore, calculations show that although using fewer test-days results in a decrease in the accuracy of selection, it does provide an opportunity to progeny-test more sires. Thus, using strategically timed test-days and Wood's model to estimate lactation yield, can lead to a more efficient use of the allocated resources.

Show MeSH