Evaluation of bull prolificacy on commercial beef cattle ranches using DNA paternity analysis 1,2

Transcription

1 Published November 21, 2014 Evaluation of bull prolificacy on commercial beef cattle ranches using DNA paternity analysis 1,2 A. L. Van Eenennaam,* 3 K. L. Weber,* and D. J. Drake *Department of Animal Science, University of California, Davis 95616; and University of California Cooperative Extension, Yreka Abstract: SNP-based DNA testing was used to assign paternity to 5,052 calves conceived in natural service multisire breeding pastures from 3 commercial ranches in northern California representing 15 calf crops over 3 yr. Bulls present for 60 to 120 d at a 25:1 cow to bull ratio in both fall and spring breeding seasons in ~40 ha or smaller fenced breeding pastures sired a highly variable (P < 0.001) number of calves (Ncalf), ranging from 0 (4.4% of bulls present in any given breeding season) to 64 calves per bull per breeding season, with an average of 18.9 ± There was little variation in Ncalf among ranches (P = 0.90), years (P = 0.96), and seasons (P = 0.94). Bulls varied widely (P < 0.01) in the average individual 205-d adjusted weaning weight (I205) of progeny, and I205 varied between years (P < 0.01) and seasons (P < 0.01) but not ranches (P = 0.29). The pattern for cumulative total 205-d adjusted weaning weight of all progeny sired by a bull (T205) was highly correlated to Ncalf, with small differences between ranches (P = 0.35), years (P = 0.66), and seasons (P = 0.20) but large differences (P < 0.01) between bulls, ranging from an average of 676 to 8,838 kg per bull per calf crop. The peak Ncalf occurred at about 5 yr of age for bulls ranging from 2 to 11 yr of age. Weekly conception rates as assessed by date of calving varied significantly and peaked at wk 3 of the calving season. The distribution of calves born early in the calving season was disproportionately skewed toward the highly prolific bulls. The DNA paternity testing of the subset of those calves born in wk 3 of the calving season was highly predictive of overall bull prolificacy and may offer a reduced-cost DNA-based option for assessing prolificacy. Prolificacy of young bulls in their first breeding season was positively linearly related (P < 0.05) to subsequent breeding seasons, explaining about 20% of the subsequent variation. Prolificacy was also positively linearly related (P < 0.05) to scrotal circumference (SC) EPD for Angus bulls that had SC EPD Beef Improvement Federation accuracies greater than Varying prolificacy of herd bulls has implications for the genetic composition of replacement heifers, with the genetics of those bulls siring an increased number of calves being disproportionately represented in the early-born replacement heifer pool. Key words: beef cattle, bulls, DNA, markers, paternity, prolificacy 2014 American Society of Animal Science. All rights reserved. J. Anim. Sci : doi: /jas Introduction In the commercial cow-calf sector, the principal determinants of income are the number of sale animals and the value per sale animal (Garrick and Golden, 2009). In 1 This work was supported by National Research Initiative Competitive Grant No ( Integrating DNA information into beef cattle production systems ) from the USDA National Institute of Food and Agriculture. 2 The authors gratefully acknowledge the cooperation and labor provided by the 3 collaborating ranches (Cowley Family Ranch, Kuck Ranch, and Mole Richardson Farms). 3 Corresponding author: alvaneenennaam@ucdavis.edu Received October 3, Accepted March 22, that regard, a herd bull has 2 qualities of value to commercial producers. One is his ability to impregnate as many cows as possible, and the other is the ability to pass genes for superior performance on to his offspring. In the absence of the former, the latter is moot. Natural service breeding is the predominant practice for beef cattle operations in the United States, but few studies have examined the variation in the number of calves sired in multiple-sire breeding pastures and the consistency of an individual bull s performance over time. Few genetic tools exist for selecting bulls with superior breeding performance. Holroyd et al. (2002) found that there were breed differences in a variety of traits related to calf output (e.g., scrotal circumference,

2 2694 Van Eenennaam et al. testicular tone, dominance, libido score, and semen quality), but that those traits explained only 35 57% of the phenotypic variation in the number of progeny sired. The objectives of this study were to use SNP-based DNA testing to quantify variation in prolificacy of bulls in multisire breeding groups using data from 3 large commercial beef ranches in northern California, and to investigate the potential uses of this information for management decisions in commercial beef cattle production systems. Bull age and breed association EPD were examined to determine their relationship with bull prolificacy. Additionally, various calf subsampling strategies were examined to determine their accuracy as as an approach to avoid the costs associated with sampling DNA from the entire calf crop. Materials and Methods Three commercial ranches, designated A, B, and C and located in the Shasta Valley of northern California, were evaluated for bull performance using DNA testing to determine parentage. Registered Angus sires had been used extensively during the previous 10 yr on these 3 ranches, making the cow herds primarily Angus. Ranches raised their own replacement heifers. Ranch A had a spring calving herd of 550 cows and a fall calving herd of 350 cows. Breeding seasons were 60 d in length and included several breeding pastures typically involving 2 to 5 bulls and a cow to bull ratio of approximately 25:1. Breeding pastures were fenced and generally less than 40 ha in size. Bulls used included predominately Angus (n = 64) plus a small number of South Devon (n = 2) and South Devon Angus cross (n = 6) bulls bred and raised on the ranch. Ranch B had a 200-cow spring calving herd and a fall calving herd of about 300 cows. On this ranch breeding seasons were 90 d in length with several breeding pastures made up of 2 to 5 bulls and a cow to bull ratio of approximately 25:1. Breeding pastures were less than 40 ha in size. Bulls used include predominately Angus (n = 19) and a small number of Horned Hereford (n = 2). Ranch C had a fall calving herd of about 700 cows. The breeding season was approximately 120 d. Breeding pastures tended to be bigger than on the other 2 ranches, with a larger number of bulls, 5 to 9, in each pasture, but a cow to bull ratio of approximately 25:1 was maintained. Bulls used included predominately Angus (n = 37) and a small number of Red Angus (n = 3), Horned Hereford (n = 1), and Polled Hereford (n = 1). Before being joined with the cows, a breeding soundness examination (BSE) was conducted on all bulls, and only bulls passing the exam were used. Breeding groups consisted of replacement heifers as a single group and mature (all other) cows. The cows were not assigned to the same breeding group each year, but rather were assigned on the basis of practical considerations (e.g., amount of feed available in different pastures) at the judgment of the ranch manager. Cows in the various fall or spring herds generally stayed with those breeding herds. Bulls were observed on a daily basis or several times per week during the breeding season and were removed from the breeding pasture for injury or poor body condition. When bulls were removed, replacement bulls were most frequently obtained from other breeding groups as idle substitute bulls were not typically available. The replacement bull selection decision was based on the judgment of the rancher when considering a variety of practical factors, including reassigning bulls from breeding pastures that had a slightly lower cow to bull ratio, selecting bulls that were observed to be very actively breeding cows, and making selections to avoid known bull dominance issues such as a history of observed aggressive behavior between specific individuals and avoiding the comingling of older bulls with young inexperienced bulls. Replacement heifers were bred to younger (generally less than 3 yr of age), lighter-weight bulls that had typically been purchased for calving ease. These bulls were shifted from replacement heifer to mature cow breeding pastures as they became older and heavier after a couple of years. Birth dates and dam identification were obtained at calving, and calves were individually identified (n = 5,052). Birth weights were taken only on Ranch B. At marking time, electronic ear tags (Destron Fearing, St. Paul, MN) were applied to calves, and hair samples were obtained for DNA testing. Calf genotypes (~100 SNP) were obtained as part of the Bovine SeekSire genotyping service (GeneSeek Inc., Lincoln, NE). Parentage SNP genotypes for bulls were extracted from Bovine SNP50 BeadChip genotypes (Illumina, San Diego, CA), which had been acquired for these bulls in the course of a previous project. Bull genotypes were compared with calf genotypes using the SireMatch program (J. Pollak, U.S. Meat Animal Research Center, Clay Center, NE). Bull prolificacy was defined as the number of calves assigned to each bull by DNA paternity testing (Ncalf) for an entire calf crop. Individual calf weights were obtained at approximately 205 d of age. Weights were adjusted using Beef Improvement Federation (BIF) linear adjustments for cow age and calf age (BIF, 2010), although some of the calves fell outside the recommended age range on weigh day because of practical constraints associated with calves going to summer pastures such that they were not accessible for weighing during the BIF prescribed age range. Weights for each calving group were obtained on consecutive days when it was not possible to weigh the entire group on 1 d. Individual 205-d weights (I205)

3 Commercial beef bull prolificacy 2695 were also adjusted for sex differences by least squares using a model that included fixed effects of ranch, year, season, and calf sex. Total adjusted 205-d weight of progeny (T205) for each bull was defined as the sum of the I205 of all his progeny for each calf crop. Prolificacy of young bulls (age less than 3 yr) was defined as that observed for their first breeding season and was compared to the average of their prolificacy recorded in subsequent breeding seasons Prolificacy and EPD Relationships To determine the relationships between American Angus Association EPD and bull means for Ncalf, I205, or T205, these variables were evaluated using scatterplots for each EPD. Because of the use of some young low-accuracy bulls, all analyses were restricted to bulls with BIF EPD accuracies greater than 0.05 for the trait being evaluated. Average phenotypic values that showed a probable linear relationship with EPD (P < 0.20) using linear regression were selected for further evaluation. Prolificacy and Calving Distribution The beginning of the calving season (d 1) for each calf crop was assigned as the date the fifth calf was born on a given ranch. Birthdates were grouped into 7-d intervals representing weeks of the calving season. Bulls were categorized into 3 equal-sized prolificacy groups, high (HP), middle (MP), and low (LP) prolificacy, based on the total number of progeny produced in each calf crop. Additionally, the prolificacy grouping of young bulls (age less than 3 yr) for their first breeding season was compared to their average prolificacy grouping in subsequent breeding seasons. Alternative Prolificacy Assessment Methods As a less costly alternative to parentage-testing all calves for assessment of prolificacy, smaller subsets of calves were used to sort bulls into prolificacy groups HP, MP, and LP. Alternative subsets included calves from a single week (wk 3) of the calving season or the sum of calves from 3 wk (wk 2, 3, and 4). The original prolificacy assessments based on the entire calf crop were compared to the assessments based on these subsets using a chi-squared test. Statistical Analyses Analyses of Ncalf were restricted to bulls present for the full duration of the breeding season and were analyzed using a model that included fixed effects of ranch (R), year (Yr), and season (Sn). The contributions of Ncalf and I205 to T205 were estimated by comparing R 2 from regressions with T205 as the dependent variable and inclusion of Ncalf and I205 together or singularly (Systat). Correlations were estimated with Pearson correlations. Ncalf repeatability for bulls that were present for more than two seasons was calculated for all bulls, and for mature bulls older than 3 years of age using the method of Lessells and Boag (1987), which accounts for among-group and within-group variance for unequal group size. Effect of bull age on Ncalf was evaluated using the linear mixed effects model (LMER) function of R and a model that included fixed effects R, Yr, Sn, and bull age and random effect of individual bull. Linear regression of selected breed association EPD and average prolificacy were evaluated for P levels. Individual bulls were examined in scatterplots, and suspect outliers based on a very high Cook s distance (Di) and studentized residual (P < 0.05) were removed from the analyses if their inclusion alone obfuscated a statistically significant linear relationship that was present in the absence of that data point. The mean number of calves born each week for bulls with 2 or more calf crops was calculated by least squares using a model that included fixed effects of R, Yr, Sn, and week (W), and mean separation significance was corrected by Bonferroni adjustments. Mean weekly calving rate for each group was calculated by least squares using a model that included fixed effects of R, Yr, Sn, W, prolificacy group, and prolificacy group by week. Prolificacy for the first breeding season for young bulls was compared to the average value in subsequent breeding seasons by regression. First season prolificacy grouping (HP, MP, or LP) was compared to the average grouping in subsequent breeding seasons by Pearson, rank order correlation, and chi-squared analyses. Because more than 20% of the cells in the chi-squared analysis were below 5, a likelihood ratio chi-squared analysis was also conducted. RESULTS Paternity Assignments A total of 5,052 calves were assigned paternity on the basis of DNA from 15 calf crops and 275 bull natural breeding opportunities (Table 1). Reproductive failure, meaning that no calves were produced, occurred in 4.4% of the bull seasons (12 out of 275 bull breeding season opportunities). In 40% of the calf crops at least 1 bull sired only 1 calf and at least 1 bull sired more than 50 calves. DNA information was unable to uniquely assign paternity to an average of 3.8% of progeny across all ranches, with 2.6% on Ranch A, 3.0% on Ranch B, and 5.7% on Ranch C.

4 2696 Van Eenennaam et al. Table 1. Average bull age at the beginning of the breeding season and number of calves produced per natural service bull in multisire breeding pastures on 3 commercial ranches (A, B, C) in northern California in Per bull No. of Minimum bull Maximum bull Mean bull age Total Minimum no. Maximum no. Mean no. calves Ranch Year Calf crop sires age, yr age, yr ± SEM, yr no. calves calves 1 calves ± SEM A 2009 Spring ± ± 3.8 A Fall ± ± 18.2 A 2010 Spring ± ± 3.8 A Fall ± ± 22.1 A 2011 Spring ± ± 4.7 A Fall ± ± 14.5 B 2009 Spring ± ± 10.0 B Fall ± ± 9.3 B 2010 Spring ± ± 7.4 B Fall ± ± 12.9 B 2011 Spring ± ± 14.4 B Fall ± ± 6.2 C 2009 Fall ± ± 3.0 C 2010 Fall ± ± 3.8 C 2011 Fall ± ± 5.7 A ± 0.7 2, ± 6.0 B ± 0.5 1, ± 3.8 C ± 1.6 1, ± 13.2 A, B, C ± 0.3 1, ± 3.8 A, B, C ± 1.0 1, ± 1.9 A, B, C ± 0.6 1, ± 7.7 A, B, C Spring ± 0.7 1, ± 12.2 A, B, C Fall ± 0.2 3, ± 5.0 A, B, C ± 1.7 5, ± Where bulls produced at least 1 calf. In 4.6% of the breeding seasons (12 out of 275) bulls produced no progeny. Prolificacy The overall mean Ncalf was 18.9 ± 13.1 progeny, with little variation among ranches, 18.6 ± 6.0, 19.9 ± 3.8, and 21.1 ± 13.2 (P = 0.90) for Ranches A, B, and C, respectively. Similarly, differences between years (19.9 ± 3.8, 20.1 ± 1.9, and 19.7 ± 7.7; P = 0.96) and seasons (spring, 20.5 ± 12.2; fall, 19.2 ± 5.0; P = 0.94) for Ncalf were small. Additionally, Ncalf across the 15 calf crops showed little variation (P = 0.51), ranging from 14.4 ± 5.7 to 26.5 ± 14.4 (Table 1). However, the mean Ncalf per bull varied widely (P < 0.01), ranging from a mean of 3.3 ± 6.3 to 39.1 ± 10.9 (Fig. 1). Repeatability of Ncalf for bulls 3 or more years of age was 0.37, and it was 0.33 when all bulls were included in the analysis. I205. Individual calf 205-d weight did not vary significantly between ranches (232 ± 10.1, 235 ± 6.4, and 279 ± 18.0 kg; P = 0.29) but showed significant year (225 ± 2.4, 233 ± 2.3, and 227 ± 2.4 kg; P < 0.01) and season (spring, 237 ± 2.5 kg; fall, 220 ± 2.0 kg; P < 0.01) differences. Again, bulls varied widely in mean progeny I205 (P < 0.01), ranging from means of 196 to 262 kg (Fig. 1). T205. Total contribution of calf weight (sum of individual sex-adjusted 205-d wt) per bull showed a pattern similar to that of Ncalf alone (Fig. 1). Total adjusted 205- d weight of progeny per bull was similar across ranch (4,687 ± 1,890, 5,713 ± 1,077, and 3,425 ± 1,404 kg; P = 0.35), year (4,762 ± 388, 4,410 ± 368, and 4,654 ± 386 kg; P = 0.66), and season (spring, 4,882 ± 414 kg; fall, 4,336 ± 304 kg; P = 0.20) but varied widely between bulls (P < 0.01), ranging from mean totals of 676 to 8,838 kg. The mean number of calves per bull was highly related (P < 0.01) to total production, explaining 96.9% of the variation (R 2 = 0.969), with each calf contributing 220 ± 2.6 kg. Mean I205 was also related to total production (P < 0.05) but by itself explained only 2% of the variation (R 2 = 0.02). Similarly, Ncalf was highly correlated (r = 0.98) to T205 per bull compared to the lower correlation (r = 0.15) to I205. Prolificacy and Bull Age. There was no significant relationship between Ncalf and bull age. As shown in Fig. 2, bulls of increasing age tended to have a higher maximum Ncalf, resulting in a higher variance in Ncalf. Young bulls in their first breeding season (N = 24) ranged in age from 1.4 to 2.9 yr (mean of 2.4 ± 0.3 yr). The mean number of calves per young bull ranged from 1 to 40 (mean of 14.9 ± 9.8). Prolificacy in subsequent breeding seasons was positively linearly related (P < 0.05) to first breeding season prolificacy, explaining about 20% of the subsequent varia-

5 Commercial beef bull prolificacy 2697 Figure 1. Mean number of calves (Ncalf; left axis), calf 205-d sexadjusted weight (I205), and total 205-d sex-adjusted weight (T205)/20 (right axis) per natural service bull present in a multisire breeding pasture. Only bulls that were present for the entire length of the breeding season and that were in use for more than a single breeding season are included. tion. The correlation between first breeding season prolificacy and mean subsequent prolificacy was 0.45 with a rank order correlation of Young bulls categorized into 3 equal groups, HP, MP, and LP, based on their firstyear prolificacy were not related to subsequent categorization by chi-squared analysis (P = 0.20) or likelihood ratio chi-squared analysis (P = 0.20), although the sample size of this young sire group was relatively small (n = 24). Young bulls categorized by their first breeding season as HP tended to remain HP (3/8 or 4/8), and only 12.5% (1/8) fell to LP. Similarly, young bulls initially categorized as LP mostly remained as LP (62.5%, 5/8), with 25% (2/8) changing to MP but only 12.5% (1/8) improving to HP. Prolificacy and EPD. Scrotal circumference (SC) EPD was not significantly related (P = 0.16) to prolificacy all bulls were included in the data set. However, a single outlier bull (from Ranch C) had a large Cook s distance (Di) and studentized residual. When that bull was removed from the analysis, SC was related (P = 0.04) to prolificacy (Fig. 3), where Ncalf = (±3.0) SC (R 2 = 0.13, SE = 8.7). The equation for Ranches A and B combined (without Ranch C) was within the confidence intervals for the combined equation of Ranches A, B, and C (without the single outlier bull). Carcass weight (CW) was negatively related (P = 0.03) to prolificacy, where Ncalf = (±0.158) CW (R 2 = 0.09, SE = 10.1). Figure 2. Calves per bull (Ncalf) per calf crop vs. age of the bull for natural service bulls present in multisire breeding pastures on 3 northern California commercial beef ranches (A, B, and C). Carcass weight and the $Beef (BE) index value were also both negatively related (P < 0.05) to prolificacy, likely because CW is a component in BE and is therefore highly correlated (r = 0.81) to it. Additionally, when CW and BE were regressed together, only CW was significant. When both SC and CW were regressed together against prolificacy, only SC (P < 0.05) remained significant compared to CW (P = 0.28). Scrotal circumference and CW were not correlated (r = 0.01). Regression responses similar to those seen for Ncalf were seen for SC and CW when regressed on T205, where T205 = 7, ,189 (±1919) SC (P = 0.04, R 2 = 0.13, SE = 4,345), T205 = 14, (±77.7) CW (P = 0.03, R 2 = 0.10, SE = 4,987). The similar regression responses for T205 and Ncalf were not surprising because of the high correlation (r = 0.98) between these 2 variables. No other EPD examined were related to prolificacy. Individual 205-d weight was positively correlated with weaning weight (P < 0.01), yearling weight (P < 0.01), CW (P < 0.01), $Feedlot index value (P < 0.01), and BE (P < 0.01) EPD but not to any other EPD. Individual 205-d weight was not highly correlated with either Ncalf (r = 0.22) or T205 (r = 0.26). The importance of Ncalf compared to that of I205 on T205 is demonstrated by comparing the R 2 (0.9793) for the regression of Ncalf (P < 0.01) on T205 to the R 2 (0.9812) for the combined regression of Ncalf (P < 0.01) and I205 (P = 0.09) on T205, showing the very small improvement resulting from the inclusion of I205 in the regression.

6 2698 Van Eenennaam et al. Figure 4. Adjusted calves born per week of the calving season across all 15 calf crops show that peak calving and implied peak conception occur during wk 3 of the calving (and breeding) season. Means that differ (P < 0.05) are noted with different letters. Error bars represent SEM. Figure 3. Mean calves per bull (Ncalf) vs. scrotal circumference (SC) EPD of Angus natural service bulls present in multisire breeding pastures on 3 northern California commercial beef ranches (A, B, and C). Bull 648 was classified as an outlier on the basis of a high Cook s distance (Di) and studentized residual (P < 0.05). Calving Distribution Calving distribution showed the preponderance of calves being born early in the calving season (Fig. 4). The largest number of calves born in a single week occurred in wk 3 in 12 of the 15 (80%) calving seasons evaluated. Pooled across ranches and adjusted for ranch, year, and season, peak calving occurred in wk 3. In the 3 seasons where peak calving was not during wk 3, it occurred in wk 2 twice and wk 1 once. The HP bulls sired more calf births per week during the early part of the calving season than the MP or LP bulls (P < 0.01; Fig. 5). Bulls siring more progeny (HP) had a disproportionately higher percentage of calves born early in the calving season. Low prolificacy bulls tended to have a consistently low number of calves born throughout the calving season. These data suggest that high prolificacy is associated with the breeding of a greater than expected number of cows early in the breeding season, ultimately leading to a larger total number of progeny for the calf crop. Alternative Prolificacy Assessment Methods As an alternative to assessing bull prolificacy by determining parentage of all calves, progeny from only a single week (wk 3) or a subset of weeks (wk 2, 3, and 4) were modeled and compared to DNA sampling and testing all calves. Prolificacy assessments based on either a single week or the sum of several weeks were closely related (P < 0.01) to prolificacy assessment based on the total calf crop. No HP bulls were reassessed to LP using the subset of calves born in wk 2, 3, and 4, and only 1.4% were reassessed using only those born in wk 3 (Table 2). Between 17% and 19% of the HP bulls were reassessed to MP using the subsets. Nonetheless, either of these reduced assessments (wk 2, 3, and 4 or only wk 3) offers a reduced sampling (and thus cost) method of determining prolificacy. Reassessment of LP bulls to HP also occurred at a low rate ( 3% for both subsets). DISCUSSION DNA testing can be used to accurately assign paternity and has been documented here and by others (Holroyd et al., 2002; Van Eenennaam et al., 2007; Gomez-Raya et al., 2008), providing an opportunity to investigate bull performance in commercial multisire breeding pastures. We have shown that prolificacy is the main driver of bull productivity as measured by total weight of calves weaned per bull. The results of bull prolificacy studies conducted in varying environments and management systems consistently reveal that prolificacy varies considerably among individual bulls. Holroyd et al. (2002) also used DNA paternity in multisire commercial ranches but with Bos indicus or Bos indicus cross bulls in extensive conditions in northern Australia and found very similar results to our temperate intensive production systems, including a 6% frequency of reproductive failure, i.e., bulls that sired no calves in a given breeding season, compared to the 4.4% observed in the current study. They found several phenotypic traits related to bull performance with little practical application because of the limited amount

7 Commercial beef bull prolificacy 2699 Table 2. Percent concordance between alternative prolificacy assessment methods comparing assignments of bulls into equal-sized top (high), middle (medium), and bottom (low) third prolificacy groups based on sampling only those calves born in a 3-wk period (wk 2, 3, and 4) or a single week (wk 3) of the calving season to prolificacy assessments based on sampling the entire calf crop Original Reassessed using only prolificacy calves born in wk 2, 3, and 4, % assessment 1 Reassessed using only calves born in wk 3, % High Medium Low High Medium Low High Medium Low Based on sampling entire calf crop. Figure 5. Bulls categorized into 3 equal-sized groups (high, medium, and low prolificacy) based on their total number of progeny in each calf crop had different calving distributions with implied differences in breeding and conception distribution. Error bars represent SEM. of variation explained by these traits. Bamualim et al. (1984) found bull age (between 2 and 5 yr of age) was not significant on pregnancy rate of Bos indicus cross bulls; however, they may not have had sufficient cow numbers to adequately challenge bull fertility capacity. A significant relationship was found between SC EPD and bull prolificacy in the current study, but no other breeding group management activities explained a significant amount of variation in calf output. Scrotal circumference measurements have been previously associated with Ncalf (Coulter and Kozub, 1989), although Holroyd et al. (2002) generally found no relationship between actual SC measurements and prolificacy with the exception of 5/8 Brahman bulls. This relationship between SC EPD and prolificacy during a natural service breeding season has not been previously reported. Scrotal circumference EPD have been positively associated with sperm motility and total BSE score (Moser et al., 1996). Favorable influence of scrotal circumference and SC EPD on heifer maturity has been reported (Brinks et al., 1978; Toelle and Robison 1985; Smith et al., 1989; Moser et al., 1996; Martínez-Velázquez et al., 2003). Scrotal circumference is also a component of the breeding soundness examination that has been related to bull fertility (Kealey et al., 2006), and SC estimates testicular tissue volume, which impacts semen quantity. The most valuable SC measurements are those taken at approximately 1 yr of age. Difficulties in assessing fertility in both male and female cattle are well documented, and genetic improvement is further hindered by the low heritability of fertility traits (<0.15; Mackinnon et al., 1990). Bull and cow fertility are low but positively correlated (r = 0.16), suggesting indirect trait selection in males for fertility traits could be an approach to improve female fertility. Scrotal circumference EPD provide an early indication of potential sire prolificacy and avoid environmental influences associated with actual SC measurements. Tropical cattle selected for high or low pregnancy rate breeding value resulted in bull progeny with larger scrotal circumference at 18 mo of age for the higher pregnancy rate group (Mackinnon et al., 1987). Bamualim et al. (1984) found actual SC was correlated to pregnancy rate (r = 0.15) in the year of measurement but was negatively correlated to lifetime pregnancy rate (r = -0.10) and that these correlations were much lower than those for breeding soundness scores, which were r = 0.36 and 0.47, respectively. Although the relationship between SC EPD and bull prolificacy was not strong, SC EPD may serve as an indicator trait for the economically relevant traits of heifer pregnancy and stayability. Others have reported moderate positive genetic correlations between SC EPD and these traits in Nellore cattle (Eler et al., 2006; Van Melis et al., 2010), although the correlation was not found to be as strong in Bos taurus breeds (Evans et al., 1999; Martínez- Velázquez et al., 2003). The relationship between SC EPD and male reproduction as measured by prolificacy during a natural service breeding season reported in this study does not appear to be among those published elsewhere. In a post hoc sense bulls appear to be somewhat consistent in their success or lack of success in breeding large numbers of cows. Repeatability of DNA paternity determined prolificacy was 0.33 to 0.37 over 3 yr under our intensive conditions with Bos taurus bulls on different ranches with different breeding seasons and 0.43 to 0.69 under extensive Australian conditions with mixed Bos indicus bulls (Holroyd et al., 2002), suggesting applicability to wide conditions. The variation in prolificacy can partly be explained by moderate repeatability, but other underlying reasons remain elusive. By paternity testing only calves born in wk 2, 3, and 4 or only from wk 3, sampling costs would be reduced to approximately 50% and 20%, respectively, of

8 2700 Van Eenennaam et al. that required to sample the entire calf crop in this study. Because of age and source verification marketing, birthdates are often known, and samples could be collected at marking time from a designated group of calves. Even without whole-herd birthdate records, strategic planning during the third week of calving could include recording or marking calves born during that time period for later sampling. Given the moderate repeatability of prolificacy, this type of reduced sampling could provide an approach to identify bulls with the greatest likelihood of being either highly or lowly prolific. The costs involved in DNA collection include not only the costs of the tests (currently ~$15/head for parentage testing; accessed January 14, 2014), which are likely to continue to decrease in the future, but also the costs associated with unique animal identification, labor to process each animal and collect the DNA sample, and costs required to manage the records and integrate the DNA information back into herd management decisions. The value of parentage information needs to outweigh the costs of genotyping. One study examined the value of DNA paternity identification on commercial beef cattle operations. The assumption of their model, based on 15 microsatellite loci, was that the information would be used to cull bulls that were producing low weaning weight calves (Gomez-Raya et al., 2008). Although this might be important if all bulls are producing an equal number of progeny, as can be seen from the data in this study, some of the bulls that produced bulls with the lowest weaning weight were siring a large number of calves and hence were not the least profitable bulls. The benefit that could be derived from these parentage data would conceptually be the removal of low prolificacy bulls and increasing the cow:bull ratio of the more prolific bulls, thereby saving on the costs associated with maintaining an inactive or low prolificacy bull. Using the data obtained in this study, it can be estimated that if the entire calf crop were sampled to obtain prolificacy estimates, then the cost per bull to obtain prolificacy data would be approximately 20 times the cost of the test (1 bull + 19 offspring), or $300/bull in the case of a $15 test. Less expensive alternative sampling strategies could be envisioned, including sampling all bulls and only those calves born in wk 3 (~20% of the calf crop) or, alternatively, sampling only those offspring produced by young sires in their first breeding season based on the observation that young bulls categorized as either HP or LP tended to remain in those categories in subsequent breeding seasons. However, given the small number of young bulls involved in this study (8 each initially categorized as HP and LP), care should be taken in over interpreting these results. In addition to prolificacy data, the DNA information could also be used to calculate genetic merit estimates of these commercial bulls and identify those producing superior or problematic classes of calf (e.g., high birth weight calves). These results reveal that highly prolific bulls sire a disproportionately large number of the preferred more valuable, early born calves (Funston, 2012) in well-managed herds that have a large numbers of females cycling early in the breeding season. In self-replacing beef systems, replacement heifers are often selected on the basis of age to enhance the potential for conception early in their first breeding season. This study showed that only a small percentage of such replacement heifers would likely be sired from LP bulls, providing indirect selection on male fertility. Using DNA paternity assignment to evaluate the relationship between heifer fertility and sire prolificacy would provide information of economic interest given the high costs of raising replacement heifers and the overriding importance of fertility to the beef enterprise (Melton et al., 1979; Melton 1995). Implications The DNA paternity testing of the calf crop can be used to determine sire prolificacy, and testing of fewer calves offers opportunities for reduced costs. The number of calves sired by natural service bulls in multisire breeding pastures is highly variable between bulls but similar across ranches and was positively associated with SC EPD. Varying prolificacy of herd bulls has implications for the genetic composition of replacement heifers, with the genetics of those highly prolific bulls siring a lot of calves likely to be disproportionately represented in the replacement heifer pool. To be cost-effective, the costs of parentage testing need to be recouped by the value derived from this resulting information. One such use might be the cost savings associated with the removal of low prolificacy bulls, although the feasibility of this approach would depend on the continued ability of the more prolific bulls in one year to be able to successfully service an increased cow:bull ratio in the following year. Prolificacy was found to be moderately repeatable (0.33) in this field study of commercial herd sires. literature cited Bamualim, W. R. B., K. W. Entwistle, and M. E. Goddard Variation in fertility in Bos indicus cross bulls. Proc. Aust. Soc. Anim. Prod. 15: BIF Guidelines for uniform beef improvement programs. 9th ed. Beef Improv. Fed., Raleigh, NC. Brinks, J., M. J. McInerney, and P. J. Chenoweth Relationship of age at puberty in heifers to reproductive traits in young bulls. Proc. West. Sect. Am. Soc. Anim. Sci. 29: Coulter, G. H., and G. C. Kozub Efficacy of methods used to test fertility of beef bulls used for multiple-sire breeding under range conditions. J. Anim. Sci. 67:

9 Commercial beef bull prolificacy 2701 Eler, J. P., J. B. S. Ferraz, J. C. C. Balieiro, E. C. Mattos, and G. B. Mourão Genetic correlation between heifer pregnancy and scrotal circumference measured at 15 and 18 months of age in Nellore cattle. Genet. Mol. Res. 5: Evans, J. L., B. L. Golden, R. M. Bourdon, and K. L. Long Additive genetic relationships between heifer pregnancy and scrotal circumference in Hereford cattle. J. Anim. Sci. 77: Funston, R. N., J. A. Musgrave, T. L. Meyer, and D. M. Larson Effect of calving distribution on beef cattle progeny performance. J. Anim. Sci. 90: Garrick, D. J., and B. L. Golden Producing and using genetic evaluations in the United States beef industry of today. J. Anim. Sci. 87:E11 E18. Gomez-Raya, L., K. Priest, W. M. Rauw, M. Olomo-Adhiambo, D. Thain, B. Bruce, R. Torell, L. Grellman, R. Narayanan, and C. W. Beattie The value of DNA paternity identification in beef cattle: Examples from Nevada s free-range ranches. J. Anim. Sci. 86: Holroyd, R. G., V. J. Doogan, J. De Faveri, G. Fordyce, M. R. McGowan, J. D. Bertram, D. M. Vankan, L. A. Fitzpatrick, G. A. Jayawardhana, and R. G. Miller Bull selection and use in northern Australia: 4. Calf output and predictors of fertility of bulls in multiple-sire herds. Anim. Reprod. Sci. 71: Kealey, C. G., M. D. MacNeil, M. W. Tess, T. W. Geary, and R. A. Bellows Genetic parameter estimates for scrotal circumference and semen characteristics of Line 1 Hereford bulls. J. Anim. Sci. 84: Lessells, C. M., and P. T. Boag Unrepeatable repeatabilities: A common mistake. Auk 104: Mackinnon, M. J., D. J. S. Hetzel, K. W. Entwistle, and R. Dixon Correlated responses to selection for fertility in Droughtmaster cattle. Proc. Aust. Assoc. Anim. Breed. Genet. 6: Mackinnon, M. J., J. F. Taylor, and D. J. S. Hetzel Genetic variation and covariation in beef cow and bull fertility. J. Anim. Sci. 68: Martínez-Velázquez, G., K. E. Gregory, G. L. Bennett, and L. D. Van Vleck Genetic relationships between scrotal circumference and female reproductive traits. J. Anim. Sci. 81: Melton, B. E Conception to consumption: The economics of genetic improvement. In: Proc. Beef Improv. Fed. 27th Res. Symp. Annu. Meet., Sheridan, WY. p Melton, B. E., E. O. Heady, and R. L. Willham Estimation of economic values for selection indices. Anim. Prod. 28: Moser, D. W., J. K. Bertrand, L. L. Benyshek, M. A. McCann, and T. E. Kiser Effects of selection for scrotal circumference in Limousin bulls on reproductive and growth traits of progeny. J. Anim. Sci. 74: Smith, B. A., J. S. Brinks, and G. V. Richardson Relationships of sire scrotal circumference to offspring reproduction and growth. J. Anim. Sci. 67: Toelle, V. D., and O. W. Robison Estimates of genetic correlations between testicular measurements and female reproductive traits in cattle. J. Anim. Sci. 60: Van Eenennaam, A. L., R. L. Weaber, D. J. Drake, M. C. Penedo, R. L. Quaas, D. J. Garrick, and E. J. Pollak DNA-based paternity analysis and genetic evaluation in a large, commercial cattle ranch setting. J. Anim. Sci. 85: Van Melis, M. H., J. P. Eler, G. J. Rosa, J. B. Ferraz, L. G. Figueiredo, E. C. Mattos, and H. N. Oliveira Additive genetic relationships between scrotal circumference, heifer pregnancy, and stayability in Nellore cattle. J. Anim. Sci. 88: