Search
Author
Title
Vol.
Issue
Year
1st Page

Journal of Animal Science - Animal Genetics

Calving day and age at first calving in Angus heifers1

 

This article in JAS

  1. Vol. 88 No. 6, p. 1947-1956
     
    Received: June 25, 2009
    Accepted: Feb 02, 2010
    Published: December 4, 2014


    2 Corresponding author(s): jbormann@ksu.edu
 View
 Download
 Share

doi:10.2527/jas.2009-2249
  1. J. Minick Bormann 2 and
  2. D. E. Wilson
  1. Department of Animal Sciences and Industry, Kansas State University, Manhattan 66506; and
    Department of Animal Science, Iowa State University, Ames 50010

ABSTRACT

Because of difficulties in data collection and analysis, in most breeds there have been limited ways to evaluate fertility in females on a between-herd basis. The objective of this study was to determine the heritabilities (direct and maternal) for CD (calving day) and AFC (age at first calving) in American Angus heifers and to evaluate the potential for using these traits in genetic improvement of female fertility. Records (n = 2,082) from 2 herds were used. Calving day was defined as the calving date of a heifer minus the first calving date in her contemporary group. To avoid bias, noncalving heifers were assigned a penalty record CD of 30, 60, and 90 d after the last CD in that contemporary group. These assigned CD were also used to give open heifers a predicted AFC. Data were analyzed by MTDFREML using a general linear animal model. Fixed effects included herd-year, service sire of the heifer, and age of dam, and a covariate of age of the heifer at the start of the breeding season (for CD only). A model including a maternal effect was also analyzed. Heritabilities for CD using a direct model were 0.07 ± 0.04, 0.10 ± 0.05, and 0.11 ± 0.05, for each penalty adjustment, respectively. Average, minimum, and maximum estimated breeding values (in days) for sires of heifers for the 3 adjustments were −0.7, −10.6, and 9.8; −1.1, −17.2, and 16.5; and −1.6, −22.6, and 19.5. The estimates of heritability for AFC using a direct model did not differ for the different adjustments for penalty records and were 0.28 ± 0.06. Average, minimum, and maximum estimated breeding values (in days) for sires of heifers for the 3 adjustments were −0.6, −46.6, and 45.9; −1.2, −50.1, and 51.6; and −1.7, −52.9, and 56.7. In a direct-maternal model, direct heritabilities for CD decreased slightly, and for AFC increased to 0.66 ± 0.14. The maternal heritabilities and direct-maternal genetic correlations were 0.08 ± 0.05 and −0.18 ± 0.58 for CD, and 0.32 ± 0.08 and –0.85 ± 0.06 for AFC. Although AFC had a greater heritability and a wider range of breeding values than CD, the negative direct-maternal genetic correlation indicated that selecting on AFC may favor heifers that are themselves born later in the season. Therefore, CD may be more useful than AFC in selecting for female fertility in beef cattle.



INTRODUCTION

Reproduction is a very economically important complex of traits in beef production. Ponzoni (1992) showed that with an index of reproductive, growth, intake, and composition traits, reproduction made the largest difference in genetic improvement expressed in dollars. As in other livestock species, reproductive traits tend to be lowly heritable. Due to difficulties in data collection and analysis, in most breeds there are limited ways to evaluate fertility in females on a between-herd basis other than heifer pregnancy.

Some researchers have advocated using calving day (CD) as a measure of female fertility (Meyer et al., 1990; Johnston and Bunter, 1996). One advantage to using CD is that the data are easy to collect. Because birth date of the calf is all that is needed, if producers are regularly collecting and reporting calf performance information, there are no additional data required. Calving day makes sense economically because early calving cows have more than 1 breeding attempt in a season and therefore have more opportunity to breed back and stay in the herd.

Age at first calving (AFC) is another trait that is simple to collect because birth dates are all that is required. Age at first calving encompasses puberty and ability to conceive, gestate, and deliver a calf. However, expression of AFC is limited by the breeding season, both the season in which the heifers are born and the season in which they are bred. Heifers that are born later are younger and have more opportunity to be younger at calving than heifers that are born earlier. Heifers that are born earlier are not bred until they are relatively older than their herdmates and therefore do not have a chance to have a very young AFC.

The maternal effects of CD and AFC have not been examined in the literature. It could be that the AFC or CD of a dam within her calving season has an impact on the subsequent AFC or CD of her heifer. The objective of this study was to determine the heritabilities (direct and maternal) for CD and AFC in American Angus heifers and to evaluate the potential for using these traits in genetic improvement of female fertility.


MATERIALS AND METHODS

All data on university-owned animals were collected under the guidelines of the Institutional Animal Care and Use Committee. Data provided by the seedstock producer were collected under normal industry management conditions.

Records (n = 2,082) from 2 purebred Black Angus herds were used in this study. Results from analysis of pregnancy rate and first service pregnancy rate from these herds were included in Bormann et al. (2006). There were a total of 763 heifers born from 1996 to 2001 from herd 1 and a total of 1,319 heifers born from 1994 to 2000 in herd 2. All heifers were born in the spring and were born in calving seasons that ranged from 61 to 106 d, with an average of 78 d. Herd 2 synchronized estrus, and herd 1 did not. Herd 1 performed estrus detection both visually and by the HeatWatch system (Manalapan, NJ) and inseminated 12 h after observed estrus. Herd 2 had a longer breeding season, performing AI up to 6 times on their heifers (based on observed estrus). Herd 1 had a shorter breeding season and bred by AI up to 3 times. Heifers that were inseminated multiple times were bred to the same bull each time. Clean-up bulls were turned out after the AI breeding season for approximately 60 d. Heifers from both herds were fed and managed according to industry standards for developing heifers. In each herd, heifers born within a season were managed together as a contemporary group. There were 147 sires with heifers represented in the data. Seven sires were used across herds, and 43 sires were used in multiple years. The distribution of daughters by sire is shown in Table 1. Three generations of pedigree were used in the genetic analysis, and there were 5,123 animals in the pedigree. Calving day was calculated for each heifer by subtracting the calving date of the first heifer to calve in that contemporary group from the calving date of the heifer. For example, within each contemporary group of heifers, heifers that calved on the first day of the calving season were given a CD of 1. Heifers that calved the next day had CD of 2, heifers that calved the next day had CD of 3, until all heifers had been assigned a CD. Open heifers in this data set were assigned a CD of 30, 60, and 90 d after the last heifer in that contemporary group calved, based on the method of Johnston and Bunter (1996). These assigned CD were also used to give open heifers a predicted AFC.

Table 1.

Please see the pdf to view this table.

 

Summary statistics and t-tests for herd differences were calculated, and fixed effects and covariates were tested using SAS (Cary, NC). Fixed effects tested included herd-year (HY), service sire of the heifer (SS), and age of dam. Age of dam was divided into 2, 3, 4, 5 to 10, and 11 or greater years of age. Within each herd, heifers were developed, bred, and managed together, so contemporary group was defined as herd × year. Covariates that were tested included adjusted yearling weight and age of the heifer at d 1 of the breeding season (AD1). Age at d 1 of the breeding season was only used in CD models. If records for AFC were adjusted for AD1, then it became the same trait as CD. The final model for CD included HY, SS, and age of the heifer at the start of breeding. The final model for AFC included HY, SS, and age of dam.

Variance component analysis was performed by MTDFREML (Boldman et al., 1993), using a general linear mixed animal model. The model equation waswhere y = vector of phenotypic records, X = incidence matrix relating fixed effects to records, β = vector of fixed effects, Z = incidence matrix relating animals to records, u = vector of random additive direct genetic effects, and e = vector of residuals.

Expectations, variances, and covariances, respectively, were where σ2u = additive genetic variance, A = Wright’s numerator relationship matrix, G = genetic variance/covariance matrix, Z′ = the inverse of Z, and R = residual variance matrix;where σ2e = residual error variance, and I = identity matrix; and

First, only the data from herd 1 were used to obtain preliminary estimates. After that, the entire data set was analyzed.

Using the entire data set, a model with a maternal effect was considered. This model equation waswhere y = vector of phenotypic records, X = incidence matrix relating fixed effects to records, β = vector of fixed effects, Z = incidence matrix relating animals to records, u = vector of random additive direct genetic effects, W = incidence matrix relating maternal genetic effects to records, m = vector of random maternal genetic effects, S = incidence matrix relating permanent environmental effects to records, pe = vector of permanent environmental effects to records, and e = vector of residuals.

Expectations, variances, and covariances, respectively, were where g11 = additive genetic variance for direct effects, and S′ = the inverse of S;where g22 = additive genetic variance for maternal effects;where σ2pe = variance due to permanent environmental effects;where σ2e = residual error variance;where g12 = additive genetic covariance between direct and maternal effects; and

The phenotypic relationships between dam and daughter CD and AFC were further examined by calculating the phenotypic correlation between heifer and dam CD and AFC, the regression of heifer CD and AFC on dam CD and AFC, least squares means for heifer CD and AFC by dam CD or AFC category, and chi-squared values for heifer CD and AFC categories by dam CD and AFC categories. These analyses were performed by SAS.

To further understand the nature of the maternal effects, dam-daughter combinations that existed in the data were followed. Within herd, 2 data sets were made. These were all heifers in the data that had a mother with a record in the data. The second data set was a subset that included all heifers in the data that were born to first calf heifers with a record in the data. There were 369 dam-daughter pairs in herd 1 and 457 in herd 2. There were 132 heifers born to heifers in herd 1 and 239 in herd 2. The CD and AFC of the dams were divided into categories of early, middle, and late for CD and young, middle, and old for AFC, to see if the CD and AFC of the daughters was different by maternal CD or AFC category. Heifers were also divided into categories of early, middle, or late CD and young middle, or old AFC. A chi-squared test was performed to determine the relationships between mother and daughter CD categories and AFC categories.


RESULTS AND DISCUSSION

The only animals that have a true CD are those heifers that were able to breed, gestate, and deliver a calf. Assigning CD in this way creates a bias because heifers that did not get pregnant are eliminated from the data set. These are presumably the poorest fertility animals and should be included in the analysis. Notter (1988) argued that eliminating open cows would cause the genetic parameters to be biased downward, or underestimated. Also, if there are large differences between sires in the number of open daughters, eliminating those records will give the poorer sires an advantage and inflate their breeding values (Notter, 1988). Notter (1988) proposed assigning open cows a calving date based on the assumption that open cows would have calved eventually if the breeding season were long enough (Notter and Johnson, 1988). Notter and Johnson (1988) also suggesting transforming the data to normalize the distribution. However, Meyer et al. (1990) showed that normalizing the data did not change estimates. Johnston and Bunter (1996) then showed that a simpler method, which would be more appropriate to national cattle evaluations, was to assign open cows a CD that was some arbitrary number of days after the last cow had calved (Johnston and Bunter, 1996). Donoghue et al. (2004a,b) compared the methods of Meyer et al. (1990) and Johnston and Bunter (1996) and showed no difference in heritability estimates or rankings of sires, indicating either method was appropriate for genetic evaluation. Both methods were superior to eliminating records on open cows (Donoghue et al., 2004a). Because of the findings of Johnston and Bunter (1996), open heifers in this data set were assigned a value for CD, instead of predicting one from the normal distribution. Further, data were not normalized because Meyer et al. (1990) showed that it was not necessary.

Summary statistics for age of heifers at the start of the breeding season, CD, and AFC for each of the assigned days for open heifers are shown in Table 2. Both herds were virtually identical in the age at which they started breeding heifers, at approximately 14 mo. Although the actual calving season for herd 2 was longer than for herd 1, herd 2 had more of their heifers concentrated early in the breeding and calving season than herd 1. Therefore, the average CD for herd 2 were much less than for herd 1 (P < 0.05). However, the ranking of animals within their HY contemporary groups should not be affected by length of the calving season. Heifers from herd 1 were older than heifers from herd 2 when they had their first calf (P < 0.05). Like CD, average AFC were longer with the larger adjustments for open heifers.

Table 2.

Please see the pdf to view this table.

 

Table 3 shows the pregnancy data for the heifers. The overall conception rate was 87.6% for herd 1 and 96.5% for herd 2. Likewise, the first service conception rate was 52.6% for herd 1 and 75.4% for herd 2. The proportion of natural calves was inversely related to the number of inseminations performed.

Table 3.

Please see the pdf to view this table.

 

Table 4 includes the P-values for the fixed effects and covariates in the model. Herd-year was highly significant for all measures of CD and AFC. Service sire of the heifer approached significance (P < 0.23) for CD and AFC when the 30-d adjustment for open heifers was used, but it was significant (P < 0.09) for the 60- and 90-d adjustment, and therefore was left in the model. Age of dam of the heifer had no effect (P > 0.48) on CD but was highly significant (P < 0.01) for AFC. Heifers born to 2-yr-old dams were older at first calving than heifers born to 3- to 10-yr-old dams (P < 0.05). This could be because the day that the heifer was born (within the calving season of her dam) has a large affect on her subsequent AFC. If the dam of a heifer calves early, that heifer will be older at the beginning of her own breeding season, not giving her the opportunity to have a good, short AFC herself. The parity of her dam affects when that heifer was born. In this study, yearling heifers were bred so they will calve before mature cows. Daughters of yearling heifers were born earlier, causing their AFC to be longer. The covariate of adjusted yearling weight was not significant for CD or AFC. The covariate of age of the heifer at the start of breeding (AD1) approached significance (P < 0.25) for CD.

Table 4.

Please see the pdf to view this table.

 

Calving Day

Records from herd 1 were analyzed first. Additive genetic variances, error variances, and heritabilities ± SE for the 3 different adjustments for CD in herd 1 and for the entire data set are shown in Table 5. The heritability of CD in herd 1, with open heifers given a penalty record of 30, 60, and 90 d more than the last day of the calving season, was 0.07 ± 0.06, 0.08 ± 0.06, and 0.09 ± 0.07, respectively. When the entire data set was analyzed together, results were similar, with heritabilities of 0.07 ± 0.04 for CD30, 0.10 ± 0.05 for CD60, and 0.11 ± 0.05 for CD90. Genetic and error variances were larger using all the data than when using just herd 1. These estimates agreed with those in the literature of 0.06 (MacNeil and Newman, 1994), 0.06 ± 0.01 (Donoghue et al., 2004c), 0.07 (Johnston and Bunter, 1996), 0.08 (Meyer et al., 1990, 1991), 0.09 (Smith et al., 1989; Morris et al., 2000), and 0.13 (Bailey et al., 1987). Average, minimum, and maximum estimated breeding values for CD30, CD60, and CD90 for heifers and sires of heifers are presented in Table 6.

Table 5.

Please see the pdf to view this table.

 
Table 6.

Please see the pdf to view this table.

 

AFC

The same procedure was followed to analyze AFC. First herd 1 was analyzed and then the whole data set together. Table 7 shows the genetic and error variances and the heritabilities from the analysis of the herd 1 alone and for the entire data set analyzed together. Both the genetic and the error variance were less when all the data were analyzed together. In herd 1, the heritabilities of AFC30, AFC60, and AFC90 were 0.35 ± 0.09, 0.31 ± 0.09, and 0.27 ± 0.08, respectively. Unlike CD, the estimates slightly decreased as the adjustment for open heifers increased. The estimates for AFC30, AFC60, and AFC90 using the whole data set were the same, at 0.28 ± 0.06. This was comparable with literature estimates of 0.22 (Frazier et al., 1999) and 0.24 (Toelle and Robison, 1985). Both of these studies eliminated open heifers and used only records from cows that calved as 2-yr-olds (Frazier et al., 1999) or as 2- or 3-yr-olds (Toelle and Robison, 1985). Average, minimum, and maximum EBV for AFC30, AFC60, and AFC90 for heifers and sires of heifers are presented in Table 8.

Table 7.

Please see the pdf to view this table.

 
Table 8.

Please see the pdf to view this table.

 

Maternal Effects

Table 9 shows the direct genetic, maternal genetic, and error variances, the direct and maternal heritabilities, and the direct-maternal covariances and correlations for CD30, CD60, and CD90 for a maternal effect model. When a maternal effect was added to the model, direct heritability for CD30 decreased to 0.04 ± 0.04, and the maternal heritability was 0.08 ± 0.05. The direct-maternal genetic correlation was lowly negative with a large SE (−0.18 ± 0.58). The trend was very similar for CD60 and CD90; however, the direct-maternal correlation became stronger for each successive adjustment for open heifers. None of the studies that have looked at CD as a fertility trait have incorporated maternal effects into the model.

Table 9.

Please see the pdf to view this table.

 

Direct genetic, maternal genetic, and error variances, the direct and maternal heritabilities, and the direct-maternal covariances and correlations for AFC30, AFC60, and AFC90 for a maternal effect model are shown in Table 10. For AFC30, direct heritability increased from 0.28 ± 0.06 without a maternal effect to 0.66 ± 0.14 with a maternal effect in the model. The direct heritability was slightly less for the other adjustments, at 0.59 ± 0.14 and 0.54 ± 0.14, respectively. The heritabilities for each adjustment were much greater with the maternal effect in the model than without it. This is because of an increase in the genetic variances. The explanation for this is unknown. It could be that there is confounding between sires and observations. The youngest heifers in the data are likely to be sired by clean-up bulls. By definition, those clean-up bulls had no opportunity to sire heifers early in the breeding season. This should be examined in a larger data set that has heifers sired only by AI or only by natural service.

Table 10.

Please see the pdf to view this table.

 

The maternal heritability was 0.32 ± 0.08 for AFC30. Maternal heritability was slightly less for AFC60 and AFC90. The direct-maternal genetic correlation for AFC30 was large and negative, at about –0.85 ± 0.06. The direct-maternal correlation was a little less strong for the larger adjustments for open heifers. One possible explanation for the large negative direct-maternal genetic correlation was the fixed breeding season. For a heifer to have a good, short AFC, she must have been bred early in the breeding season, but also have been born late herself, meaning her mother (if she was a first calf heifer) had a longer AFC. Conversely, those heifers with the longest AFC calved late themselves, but were also born early, meaning their mothers (if they were first calf heifers) had a shorter AFC. It also could be the maternal heritability and direct-maternal correlation are an artifact of the data caused by a fixed breeding season and the birth date of a heifer in relation to her contemporary group. This is the first report of these parameters in the literature and should be considered preliminary. The relationship between direct and maternal AFC should be examined in larger data sets with more generations of dams and daughters included.

Table 11 is the phenotypic correlations between the CD and AFC of a heifer and the CD and AFC of her mother (MCD and MAFC, respectively) within farm and data set. As expected, there was a large correlation between the CD and AFC of each animal. The 0.25 correlation between CD and MCD indicated that in herd 1, first calf heifers that calved earlier in the season had daughters that calved earlier in the season, which was what would be expected, but not what the genetic analysis showed. The correlations of 0.25 and 0.11 between CD and MAFC showed that heifers that calved younger had heifers that calved earlier in the season. This makes sense because a heifer with a young AFC probably calved early in her season, giving her daughter an age advantage to calve early. The negative correlations between AFC and MCD, which were only significant in herd 2, indicated that mothers that calved later had daughters that calved younger, which was logical, because AFC depended on when the heifer was born. Regression coefficients from the regression of the CD and AFC of a heifer on the MCD and MAFC of her dam are shown in Table 12. These results were very similar to those of the phenotypic correlations. There were significant relationships (P < 0.10) between CD and MCD in herd 1 heifers, between CD and MAFC in heifers from herd 1 and herd 2, and between AFC and MCD for all cows from both herds.

Table 11.

Please see the pdf to view this table.

 
Table 12.

Please see the pdf to view this table.

 

Tables 13 and 14 are the least squares means for daughter CD and AFC by the CD or AFC category of the dam. There was a positive effect of the CD of the mother on the CD of the daughter. Cows that calved early tended to have daughters that calved earlier (Table 13). There were differences between the phenotypic and the genetic analysis. Conversely, dams that calved late had daughters with a shorter AFC (Table 13). This was because a short AFC was partially caused by the heifer being born late. Here, the phenotypic results were in agreement with the genetic analysis. When the dams were categorized by their AFC (MAFC) into young, medium, and old, results showed that first calf heifers that were younger at calving produced daughters that calved earlier in the season (Table 14). The results were more ambiguous with all cows included, but there still was a trend (P < 0.10) for older-calving cows to have daughters that calved later in the season. There was no difference (P > 0.10) in heifer AFC between dam AFC category (Table 14). This was in disagreement with the large negative direct-maternal correlation for AFC. It could be that there were other factors compensating for the negative genetic relationship. Possibilities include culling younger heifers or younger heifers experiencing some compensatory growth during development.

Table 13.

Please see the pdf to view this table.

 
Table 14.

Please see the pdf to view this table.

 

Heifers were then also divided into categories of early, middle, or late CD and young middle, or old AFC. Comparisons between mother and daughter CD categories and AFC categories are presented in Tables 15 and 16. For most of the data sets, there was a significant difference (P < 0.05) in heifer CD category due to the MCD category of her mother. Early-calving cows and heifers tended to have heifers that also calved earlier (Table 15). This was in disagreement with the negative direct-maternal genetic correlation. There was less relationship between MAFC category and heifer CD category. In herd 1 heifers, younger-calving heifers had heifers that became calved earlier, and older-calving heifers had heifers that calved later. This relationship was not significant (P > 0.30) in herd 2 heifers, or in either herd with all cows included (Table 15). There was a highly significant relationship (P < 0.05) in most data sets between the AFC category of a heifer and the MCD category of her dam. In general, dams that calved early had heifers that were older when they had their first calf (Table 16). Conversely, later-calving dams had heifers that were younger at first calving (Table 16). As discussed previously, this was because those heifers that were born later (their dams calved later) had more of an opportunity to be younger when they themselves calved. Heifers born early (their dams were early calves) were older when they themselves calved because they were given no opportunity to conceive earlier. There was a significant relationship (P < 0.01) between the MAFC category of the dam and the AFC category of her daughter for herd 2, but not herd 1 (Table 16). In this case, younger-calving cows tended to have older-calving daughters, and older-calving cows tended to have younger calving daughters (Table 16). This was in agreement with the negative direct-maternal genetic correlation for AFC.

Table 15.

Please see the pdf to view this table.

 
Table 16.

Please see the pdf to view this table.

 

The large negative direct-maternal genetic correlation for AFC and the results from the chi-squared analysis indicate that selection for AFC could result in choosing heifers that are born late and discriminating against heifers that are born early. In this case, selection is not on the inherent genetic merit for fertility of the heifer, but on when she was born. If the young heifer that calved early (indicating good fertility) had a heifer calf, her heifer calf would not be selected because she was older and had no opportunity to have a very young AFC, in spite of her possibly good fertility.

In conclusion, whereas AFC has a greater heritability than CD, selection on AFC could have unwanted consequences, such as selecting heifers that happen to be born later with respect to the calving season. Therefore, CD may be a more appropriate trait for selection in spite of its decreased heritability.

 

References

Footnotes


Comments
Be the first to comment.



Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.