Genetic diversity and population structure of American Red Angus cattle 1

Transcription

1 Published December 4, 2014 Genetic diversity and population structure of American Red Angus cattle 1 G. C. Márquez,* S. E. Speidel,* R. M. Enns,* and D. J. Garrick 2 *Department of Animal Sciences, Colorado State University, Fort Collins 80523; Department of Animal Science, Iowa State University, Ames 50011; and Institute of Veterinary, Animal and Biomedical Sciences, Massey University, Palmerston North, 4442 New Zealand ABSTRACT: The objective of this study was to characterize the population structure and genetic diversity of registered American Red Angus cattle. Inbreeding and average relationship coefficients, effective population size, effective number of founders, and effective number of herds supplying grandparents to the population were calculated from the recorded pedigree. Inbreeding in 1960 was 10.7% and decreased until 1974 at a rate of 0.2% per year, whereas in 1975 inbreeding was 3.2% and increased until 2005 at a rate of 0.02% per year. The numerator relationship coefficients of the 10 individual paternal grandsires (PGS; sires of sires), paternal granddams (PGD; dams of sires), maternal grandsires (MGS; sires of dams), and maternal granddams (MGD; dams of dams) that had the greatest number of registered grandprogeny, with all other registered animals, increased with their birth year from 1960 on. Average numerator relationships of these with all other PGS, PGD, MGS, MGD, bulls, and sires were greater for paternal (PGS, PGD) than maternal (MGS, MGD) pathways. The effective population size was 445, with 649 effective founders. The effective numbers of herds supplying PGS, PGD, MGS, and MGD were 435, 369, 453, and 459, respectively. Inbreeding is at a low level and the effective population size is large. The effective number of founders and effective number of herds supplying grandparents is small in relation to the total number of animals and herds, indicating the disproportionate influence of a few founders and herds on the genetics of the breed. The calculated parameters indicate satisfactory genetic diversity in American Red Angus cattle. Key words: genetic diversity, inbreeding, population structure, Red Angus cattle 2010 American Society of Animal Science. All rights reserved. J. Anim. Sci :59 68 doi: /jas INTRODUCTION 1 The authors thank the Red Angus Association of America for providing the data used in this study. 2 Corresponding author: dorian@iastate.edu Received July 15, Accepted September 15, Inbreeding is a measure of genetic diversity, quantifying the probability that 2 genes in an individual are identical by descent (Wright, 1922). The inbreeding coefficient of an offspring is one-half the additive numerator relationship between its parents; therefore, the average numerator relationships among candidate parents are useful for predicting future inbreeding (Wright, 1922). Other descriptive parameters of population structure are the effective population size (N e ), effective number of founders (f e ), and effective number of herds (h e ), which are indicative of genetic diversity levels in a population. Inbred animals exhibit inbreeding depression, which adversely affects traits such as reproduction, conformation, and growth, among others (Burrow, 1993). Maintaining genetic diversity in a population is important because its loss limits mating choices and has adverse effects on economically (Golden et al., 2000) and biologically relevant traits. The Red Angus Association of America (RAAA) was established in 1954 with a focus on performance testing. The Red Angus is a widely used breed of beef cattle in the United States, yet little analysis of its population structure or genetic diversity has been published. The objective of this study was to investigate the population structure of registered American Red Angus cattle by estimating descriptive parameters, and to explore genetic diversity through inbreeding levels and numerator relationship coefficients. The information can be applied to develop mating strategies that maintain diversity. The study offers a novel view of the American Red Angus breed structure by analyzing the influence of popular herds and different selection pathways. The maintenance of diversity in American Red Angus cattle will preserve future opportunities for selection to improve the breed and adapt it to different production environments. 59

2 60 Márquez et al. Table 1. Numbers of cows, bulls, steers, paternal grandsires (PGS), paternal granddams (PGD), maternal grandsires (MGS), and maternal granddams (MGD) registered by decade for birth year, 1920 to 2002 Birth yr Cows Bulls Steers PGS PGD MGS MGD 1920 to to to to , , to ,701 12, ,095 2,510 2,121 11, to ,430 74,518 4,137 2,565 7,331 6,426 40, to , ,975 7,917 3,721 13,099 8,860 62, to , ,445 19,312 6,885 17,917 17, , to , ,986 24, ,503 3,582 14,185 MATERIALS AND METHODS Animal Care Animal Care and Use Committee approval was not obtained because the data came from an existing RAAA database. Data Description Data were obtained from RAAA for 2,141,506 registered Red Angus animals born between 1927 and Data included animal, sire, and dam identification numbers, birth year (BY), sex (cow, bull, or steer), percentage of Red Angus by breed composition (as calculated by the RAAA), and herd of origin. The number of animals with known BY was 2,051,804. Commercial animals and their progeny were included in the analysis only if they were registered with the breed association and appeared in the pedigree file. Pedigree information in the decades before 1960 was limited (Table 1) because the breed did not have an official association until 1954 (Red Angus Association of America, 2006). In 1995, the RAAA implemented total herd reporting, requiring the fate of all animals born to be reported. Inbreeding Coefficients Individual inbreeding coefficients (Wright, 1922) for the entire pedigree were calculated with the Animal Breeder s ToolKit (ABTK; Golden et al., 1992), using the algorithm described by Meuwissen and Luo (1992). Founders, defined as animals with an unknown sire, dam, or both, were considered to be the base population. The number of founders in the pedigree influences inbreeding coefficients, and a greater number of founders in the pedigree leads to reduced inbreeding because these are assumed not to be inbred and to be unrelated to each other for computing purposes. Depth and level of completeness of the pedigree were investigated by identifying the percentage of animals with at least 3 generations of known ancestry because these factors affect inbreeding calculations (Boichard et al., 1997). Mean inbreeding by BY was calculated from 1960 to 2005 to investigate trends and rates of change in inbreeding. The full calf crop of 2006 had not been entered into the database; therefore, the year 2006 was excluded from the analysis. Inbreeding coefficients for animals with BY before 1959 were not taken into account in analyses of trends because of incomplete pedigree information (Table 1). During preliminary analysis, 2 distinct periods showing apparent differences in inbreeding were identified, one from 1960 to 1974, representing the formative years of the breed, and another from 1975 to 2005, the period corresponding to the majority of registrations in the breed. Changes in inbreeding during these 2 time periods were calculated as the linear regression of individual inbreeding on BY: F i = b 0 + b 1 time + b 2 (BY i 1960) + b 3 time(by i 1960) + e i, where F i is the inbreeding coefficient of the ith individual; b 0 is the intercept for 1960; time is an indicator variable with values of 0 or 1, depending on the BY group of the animals (1960 to 1974 or 1975 to 2005); b 1 is the partial regression coefficient of time, which can be interpreted as the difference in intercept between 1960 and 1975; BY i is the BY of the ith individual; b 2 is the partial regression coefficient of BY, which can be interpreted as the slope of the linear regression of inbreeding on BY for 1960 to 1974; b 3 is the partial regression coefficient of the interaction term, which can be interpreted as the difference in regression of inbreeding between 1960 to 1974 and 1975 to 2005; and e i is the residual component of the model. Differences between rates of inbreeding by time period were tested using the significance of the interaction (b 3 ) term. A test of whether b 1 is equal to or different from zero was performed to test whether inbreeding was different in the 2 time periods. Bootstrap Estimates of Inbreeding Founders that lack complete pedigree information appear in the RAAA pedigree over time, influencing inbreeding calculations because they are assumed to be unrelated to each other and not inbred. Bootstrapping allows determination of the effect of these unknown

3 Red Angus genetic diversity 61 pedigrees on the inbreeding calculations (M. D. Mac- Neil, USDA Agricultural Research Service, Miles City, MT, personal communication). A bootstrap estimate of inbreeding using subsamples of the pedigree was obtained to determine if there were differences in inbreeding levels when taking into account complete vs. incomplete pedigrees. A complete 3-generation pedigree was assembled for 10,000 animals (with no missing parents, grandparents, or great-grandparents) from animals with BY later than Individual inbreeding coefficients were calculated on this sample by using the ABTK (Golden et al., 1992). To approximate true inbreeding in the population, parents of some of the sample of animals were set to unknown in the same frequency as in the original pedigree, and inbreeding was again calculated for this subset of animals. This process was repeated 10,000 times. Differences between inbreeding of all bootstrap samples before and after randomly removing parents were tested with a t-test (df = 9,999) to quantify the significance of differences between inbreeding in complete and incomplete pedigrees. Relationship Coefficients The inverse numerator relationship matrix (A 1 ) was obtained for all animals in the pedigree with the ABTK, using the method described by Quaas (1976), with inbreeding coefficients computed according to the method of Meuwissen and Luo (1992). The inverse of A 1, A, is a matrix containing numerator coefficients or additive relationships between all animals. These coefficients are of interest in quantifying relatedness in a population. The order of A 1 is large (2,141,506 2,141,506), so direct inversion was not feasible. Additive relationship coefficients for the ith animal with all the other animals in the pedigree were obtained by taking advantage of the fact that A 1 A = I, and therefore A 1 A i = e i, where A i is the ith column of A and e i is the ith column of the identity matrix I, a vector of zeros except for a 1 in position i. The linear equations A 1 A i = e i were solved indirectly by using conjugate gradient methodology, as implemented by Garrick et al. (2007). Grandparents in the pedigree were segregated into 4 pathways of selection as paternal grandsires (PGS; sires of sires), paternal granddams (PGD; dams of sires), maternal grandsires (MGS; sires of dams), and maternal granddams (MGD; dams of dams). The 10 most influential animals (producing the greatest number of registered grandprogeny), representing each of the PGS, PGD, MGS, and MGD pathways, were identified according to all the following criteria: 1) recorded BY after 1960, 2) known sire and dam, 3) the herds of origin of the animal, its sire, and its dam registered with RAAA for at least 10 yr (consecutive or not), and 4) 100% Red Angus, as computed by the RAAA. These animals were collectively grandparents of 1 to 10% of all registered animals. Relatedness in the pedigree was investigated by calculating additive relationship coefficients of the 10 most widely represented grandparents with all other animals in the pedigree. Average additive relationship coefficients each year for PGS, PGD, MGS, and MGD with all other animals in the pedigree were calculated from 1960 to 2002 to investigate changes in relatedness over time. Years subsequent to 2002 were not investigated, because animals born after 2002 would not have had the opportunity to become grandparents. To investigate relatedness in the different pathways of selection, average additive relationship coefficients were evaluated between the top 10 progeny-producing grandparents and all other PGS, PGD, MGS, MGD, sires, and bulls (parents and nonparents included) in the pedigree. A selection pathway was considered more related to all other animals in the pedigree than another if the mean additive relationship coefficient was greater than in the others. Differences in the mean additive relationship coefficients were tested using t-tests with a Bonferroni adjustment for multiple comparisons. N e The N e is the number of individuals that would account for the observed rate of inbreeding in the population if they were randomly mating. The N e reflects the rate at which genetic diversity will be lost in a population because loss of heterozygozity (or increases in inbreeding) is inversely related to N e (Falconer and Mackay, 1996). The N e was calculated for animals with BY after 1974 (because inbreeding trends stabilized after 1974) and for animals with BY after 1983 as N e = 1 2( F L), where L is the generation interval, calculated as the average age of parents when their offspring were born, and ΔF is the annual change in mean inbreeding, calculated as a linear regression of individual inbreeding on BY, using the R statistical language (R Development Core Team, 2007). Further, minimum, maximum, mean, and variance of family sizes were calculated because these are indicative of population structure. Large variation in family sizes leads to greater inbreeding levels because it indicates that some animals are used very widely and others are not. This increase in inbreeding leads to decreased N e. The number of offspring produced by sires and dams was quantified. The number of selected offspring (those used as parents in the next generation) of PGS, PGD, MGS, and MGD was also quantified to calculate family size statistics. f e and h e The f e is the number of equally contributing founders that would be expected to produce the same genetic

4 62 Márquez et al. Figure 1. Mean inbreeding coefficients for the registered American Red Angus population from 1960 to diversity as the population under study (Lacy, 1989). It was calculated as f e = n 1 å i= 1 where p i is the proportion of genes in the current population contributed by the ith founder, and n is the total number of founders. This can be used as a descriptor of the number of influential founders in the pedigree. The h e is the number of equally contributing herds expected to have produced the animals in the pedigree, given that all herds have an equal chance of contributing animals to the pedigree (Robertson, 1953). It was calculated for the 4 pathways of selection as the chance that 2 animals in the pedigree would have their grandsires (or granddams) bred in the same herd if grandsires (or granddams) were being equally contributed to the pedigree as h e = h å é n ëê i i-1 i-1 h å i-1 én ëê h å i p 2 i, ù ( n -1) i ûú, n -1 ù i ûú ( ) where n i is the number of grandsires (or granddams) contributed by the ith herd, summed over the total number of herds, h. The sum h ån i i-1 is therefore the total number of grandsires (or granddams) in the pedigree. The h e quantifies how many herds in the pedigree contribute to the observed levels of diversity and was calculated for PGS, PGD, MGS, and MGD as a way to describe the number of influential herds in each of these pathways of selection. RESULTS AND DISCUSSION Inbreeding Coefficients Mean inbreeding coefficients by BY are shown in Figure 1. Information before 1960 was excluded because pedigree information was limited (Table 1). The number of animals with a computed inbreeding coefficient and known BY after 1959 was 1,791,950 (83.6% of the animals in the pedigree). Inbreeding changed from 1960 up to 1974 at a rate of 0.2% per year (P < 0.001), and from 1975 to 2005, it increased slightly at a rate of 0.02% per year; the rates of change in inbreeding over these 2 periods were different (P < 0.001). The mean inbreeding in 1960 was 10.7%, whereas the mean in 1975 was 3.2%. Although inbreeding still decreased after 1975, an estimate of inbreeding in the breed as a whole was desired; therefore, 1975 was chosen as the cutoff year for calculating changes in inbreeding in the 2 different periods. This may cause downward bias in the estimation of inbreeding but reflects the entire breed after its formation. Before 1975, inbreeding decreased because new registrations resulted from the mating of parents that were less related to each other than the population average. New registrations had a proportionately large increase in this early 1960 to 1974 period, which represented only 5.8% of animals with known BY. Founders introduced into the population contributed to the reduction in inbreeding, with founders as a percentage of all registered animals averaging 20% registrations per year from 1960 to 1974, with a maximum value of 41%

5 Red Angus genetic diversity 63 Figure 2. Percentage of founders within each birth year from 1960 to in 1960 and a minimum of 4.6% in 1974 (Figure 2). The RAAA registration policy allows the offspring of animals that are not fully Red Angus to be registered and to increase their breed percentage by a process of grading up, and it allows Black Angus animals to be registered as Red Angus (Red Angus Association of America, 2006). A large number of founders, including Black Angus animals, were introduced into the pedigree in this way, many of which may have ancestors recorded by other breed associations. After 1974, inbreeding increased only slightly (0.02% per year) because a large number of unrelated individuals were available for mating. The percentage of founders (out of animals registered each year) after 1974 was on average 4% per year, thereby decreasing the immediate influence of founder registrations on inbreeding calculations and eroding the impact of founders on inbreeding in the next generations. With fewer nonrelated founders available, matings between related individuals increased, albeit not by much. The estimated inbreeding levels in American Red Angus cattle are less than estimates in most other breeds. Inbreeding of Danish dairy cattle populations was reported as 0.9 to 1.1% (Sorensen et al., 2005), that of Japanese Black cattle was reported as 5% (Nomura et al., 2001), and that of American Herefords was reported as 9.8% (Cleveland et al., 2005). Red Angus inbreeding was also less than in Irish (0.54 to 2.19%; McParland et al., 2007), Italian (1.8 to 2.15%; Bozzi et al., 2006), and Spanish (0.25 to 3.13%; Gutiérrez et al., 2003) cattle populations. All these populations had fewer animals registered than RAAA except for the Herefords, with 20,624,418 animals registered. A summary of inbreeding of Red Angus cattle from the 1990 genetic evaluation was published by Golden et al. (1991), in which 73% of the animals analyzed had inbreeding coefficients of 5% or less. These levels are low and comparable with the current estimate, which indicates that inbreeding has not been accumulating in the population since their analysis was done. Computation of inbreeding coefficients depends on the level of pedigree completeness (Figure 3) and on the base year or population used for calculations. Inbreeding levels are biased downward when pedigree information is incomplete, and the N e is overestimated (Boichard et al., 1997). Figure 3 shows the percentage of animals with ancestors going back at least 3 generations, an indicator of the depth of pedigree reporting in RAAA. More than 94% of all the animals in the pedigree have sires and dams recorded, and more than 88% have recorded ancestors going back at least 3 generations. The rate of change in inbreeding may be more informative than the actual inbreeding levels because the definition of base population is less relevant to that statistic. The base population is assumed to be mutually unrelated and not inbred for computing purposes, but in practice these assumptions do not hold. For the present computations, the base population was considered to be the founders. Young and Seykora (1996) found that estimates of inbreeding decreased when a more recent base was defined, ignoring relationships before the base, but that the rate of change in inbreeding was not significantly affected. In the present analysis, the founders that represent the base were predominantly born before Bootstrap Estimates of Inbreeding Bootstrapping allows for nonparametric estimates of a population parameter drawn from the population distribution itself. If the bootstrap samples provide an appropriate estimate of the true population parameters,

6 64 Márquez et al. Figure 3. Percentage of known ancestors out of the entire pedigree going back 3 generations. PGS = paternal grandsires; PGD = paternal granddams; MGS = maternal grandsires; MGD = maternal granddams; PGSS = sires of paternal grandsires; PGSD = dams of paternal grandsires; PGDS = sires of paternal granddams; PGDD = dams of paternal granddams; MGSS = sires of maternal grandsires; MGSD = dams of maternal grandsires; MGDS = sires of maternal granddams; MGDD = dams of maternal granddams. then they will be a good approximation of the distribution of true inbreeding in the population (Kutner et al., 2005). The mean bootstrap estimate of inbreeding in the complete 3-generation pedigree of animals born later than 1974 with no parents removed was 2.13%. When parents were randomly removed at the same frequency as in the original pedigree, the estimate was 2.10%. There were differences between samples with and without randomly removed parents, as indicated by t-tests (P < 0.001). These are statistically different because of the large degrees of freedom, but are not biologically important because of the relatively low magnitude of the differences. Mean inbreeding after parents were randomly removed was less than estimates using all parents because, by removing some parents, genetic links between individuals are broken. The results indicate that in the Red Angus pedigree, estimates of inbreeding using the current pedigree vs. a more complete pedigree (with more information on founders) would not be much different, assuming that the simulation of missing parents is similar to that in the true pedigree. There may be recording errors in the Red Angus pedigree resulting from human error or other sources. Bootstrapping can also be used as a way to investigate the effects such errors have on calculations. Because removing parents did not have a great effect on inbreeding calculations, pedigree errors would not be expected to have a great effect either. Relationship Coefficients Mean additive relationship coefficients by BY for the 10 most progeny-producing PGS, PGD, MGS, and MGD with all the other animals in the pedigree, from 1960 to 2002, are shown in Figure 4. Birth year of each of the individual top-ranked animals in each pathway is shown in Table 2. Most of these animals were born between 1988 and 1990; thus, additive relationship coefficients showed the greatest increases after this time period, which can be observed in Figure 4. The mean additive relationships for PGS, PGD, MGS, and MGD with the rest of the animals in the pedigree were 3.4, 2.2, 0.1, and 0.009%, respectively, from 1960 to There were differences in additive relationship coefficients between the 4 groups, as indicated by t-tests [P < for all comparisons except between MGS and MGD (P = 0.01) and between MGS and PGD (P = 0.01)]. Mean additive relationship coefficients increased over time for all pathways, with PGS being more closely related to the population than the rest, followed by PGD, MGS, and MGD. Average additive relationships between the most used grandparents and all other PGS, PGD, MGS, MGD, bulls, and sires in the pedigree are summarized in Table 3. Of these, the paternal pathways (PGS, PGD) are more closely related to the rest of the groups than are the maternal pathways (MGS, MGD) because fewer bulls are used as PGS than animals in other pathways (Table 1). The 10 most progeny-producing PGS are more closely related to the rest of the groups than are any of the other pathways. Conversely, the top 10 progeny-producing MGD are the least closely related to the rest of the groups because many more MGD are used than are any of the other pathways owing to replacement considerations and reproductive physiology. Females usually have 1 or 2 offspring per year (except when using artificial reproductive technologies, such as multiple-ovulation embryo transfer). Males can have multiple offspring per year through natural service and can have hundreds of offspring through AI. In the Red Angus pedigree, PGS have the least proportion of animals selected and therefore have the greatest selection intensity (2.07 SD) of the 4 pathways of selection,

7 Red Angus genetic diversity 65 Table 2. Individual birth years of the 10 most widely represented grandparents 1 Rank Birth yr PGS PGD MGS MGD Rank 1 was the most widely represented. PGS = paternal grandsires; PGD = paternal granddams; MGS = maternal grandsires; MGD = maternal granddams. whereas MGD have the greatest proportion selected and have the least selection intensity (0.96 SD) of the 4 pathways of selection (Márquez and Garrick, 2007), corroborating the current results. These results differ from a similar study by Cleveland et al. (2005) investigating the American Hereford population. In their study, the mean additive relationship between the 25 most progeny-producing Hereford sires was shown to have decreased over time, ranging from 10 to 20%. The most widely used Red Angus animals are less related to the population than is the case in Herefords. These differences in additive relatedness are partly due to Hereford cattle being more inbred (Cleveland et al., 2005). Estimates of average relationship coefficients in Spanish beef cattle populations ranged from 0.10 to 1.7% (Gutiérrez et al., 2003), and in UK dairy cattle populations ranged from 0.16 to 1.34% by Table 3. Average relationship between the 10 most progeny-producing paternal grandsires (PGS), paternal granddams (PGD), maternal grandsires (MGS), and maternal granddams (MGD), and all PGS, PGD, MGS, MGD, sires, and bulls Item PGS, % PGD, % MGS, % MGD, % All PGS All PGD All MGS All MGD All bulls All sires year (Roughsedge et al., 1999). Average additive relationships can be used to predict the average inbreeding in subsequent generations because the inbreeding coefficient of the offspring is one-half the additive relationship between its parents (Wright, 1922). The additive relationships observed in this study are similar to the inbreeding coefficients observed, so theoretically, average inbreeding for the whole pedigree may remain at stable levels in future generations because of the association between additive relationship coefficients and inbreeding. N e Calculation of Ne is dependent on the range of years used for calculating generation intervals and changes in inbreeding coefficients. In the Red Angus pedigree, N e was calculated as 445 when using a generation interval of 4.72 yr for all animals in the 1975 to 2005 population subset. The generation intervals of PGS, PGD, MGS, and MGD at this time were 4.5, 5.1, 4.4, and 5 Figure 4. Mean numerator relationships of the 10 most represented paternal grandsires ( ), paternal granddams ( ), maternal grandsires ( ), and maternal granddams ( ), with the rest of the pedigree, by birth year.

8 66 Table 4. Minimum, maximum, and average family sizes 1 with variance Pathway 2 Minimum Maximum Avg. SD Márquez et al. some sires are used widely, relatedness between individuals increases and the probability of matings between related individuals increases. Sires 1 9, Dams PGS PGD MGS 1 2, MGD Family size = number of offspring of sires and dams, and number of selected offspring of grandparents. 2 PGS = paternal grandsires; PGD = paternal granddams; MGS = maternal grandsires; MGD = maternal granddams. yr, respectively. The N e for the 1984 to 2005 population subset was calculated as 384. This latter range of years may remove some of the bias from the calculation of N e, because from 1984 onward, inbreeding increased. The 1975 base estimate of N e reflects the breed as a whole after its formation, and the 1984 base estimate reflects the N e of the breed when inbreeding began to increase steadily. The level of pedigree completeness in this study (Figure 3), compared with other studies (e.g., Gutiérrez et al., 2003; Cleveland et al., 2005), suggests that N e is not overestimated because at least 88% of animals have a complete 3-generation pedigree recorded (Boichard et al., 1997). The small rate of change in inbreeding in the Red Angus population (0.02%) contributes to the increased N e observed, which is greater compared with other breeds. Cleveland et al. (2005) found an N e of 85 in American Hereford cattle, whereas McParland et al. (2007) found N e for Irish Hereford, Simmental, and Holstein-Friesian cattle of 64, 127, and 75, respectively. In Danish dairy cattle, population estimates ranged from 49 to 157 (Sorensen et al., 2005). The N e of Italian beef cattle populations ranged from 122 to 138 (Bozzi et al., 2006), and the N e of Spanish beef cattle populations ranged from 21 to 123 (Gutiérrez et al., 2003). The greater N e indicates that there may be a wider genetic base in Red Angus for these years as compared with the other breeds. The number of offspring of sires and dams, and the number of selected offspring of grandparents were quantified, and these family sizes are detailed in Table 4. They vary widely in male pathways, but not in female pathways. Some males are used intensively, as evidenced by the maximum family size for males. This is in accord with previous results of male pathways being more related to the population than female pathways. Family sizes are indicative of population structure and reflect mating decisions made by breeders. Large variances in family size indicate that not all animals contribute their genes equally to subsequent generations of the population because some animals (males) are used very intensively, whereas others are not. This leads to increases in inbreeding (and decreased N e ), because as f e and h e Studying founders is problematic because they are unrealistically assumed not to be inbred and to be unrelated to each other. Some founders are used widely and others may seldom be used. To account for the overrepresentation of some founders in the pedigree, the f e can be used. This parameter accounts for potential gene loss in the population and is less affected than N e by pedigree completeness (Boichard et al., 1997). The number of founders in the population was 40,679, which resulted in an f e of 649. The difference between the actual founders and f e suggests that some founders were used widely, whereas others contributed little to the population (Lacy, 1989). Many of the founders in the RAAA were females that had calves that were not widely used (Comstock et al., 1998). Comstock et al. (1998) found an f e of 74 in the Red Angus population in 1996, with a total of 11,432 founders. The wide difference between the current and previous estimates is due to the large number of new animals added to the pedigree after their analysis in 1996 (50% of registrations) because the same definition of f e was used in both cases. The current estimate of f e is greater compared with reported estimates in other breeds. In Italian beef cattle populations, f e ranged from 71 to 152 (Bozzi et al., 2006), that in Irish cattle populations ranged from 55 to 357 (McParland et al., 2007), and that in Spanish beef cattle populations ranged from 48 to 846 (Gutiérrez et al., 2003). Estimates of f e are dependent on the size of the population and the number of actual founders, so smaller populations are expected to have a smaller f e, which is the case in most of the previous literature estimates. A small f e would suggest that genetic drift has been occurring since the founder generation (Sorensen et al., 2005), which is not the case in the American Red Angus population. The f e in Red Angus is large enough so that matings between unrelated individuals are possible, which also explains the small inbreeding rates. The number of herds registering animals per decade has grown (Table 5), with some herds making a larger genetic contribution to the breed than others. The total number of herds in the entire American Red Angus pedigree was 15,465, with 3,106 of these herds appearing in the pedigree for at least 10 yr (consecutive or not). The h e is a way to determine the number of influential herds in the pedigree; h e for PGS, PGD, MGS, and MGD, as well as the percentages out of total herds, are shown in Table 6. The small h e values suggest that a few herds make a disproportionate contribution to the genetics of the breed. Results are similar to those reported by Gutiérrez et al. (2003) in Spanish beef cattle, in which a small number of herds act as the selection nucleus for

9 Red Angus genetic diversity 67 Table 5. Number of herds registering at least 1 animal in a decade Birth yr Herds, n 1960 to to , to , to , to ,198 the rest of the population. Márquez and Garrick (2007) found that 4 to 5% of Red Angus herds supply 50% of the animals used as PGS, PGD, MGS, and MGD. The majority of animals selected for breeding come from a minority of herds, causing the genetic composition of the whole breed to reflect that of these animals and herds (Harris, 1998). The loss of genetic diversity in a breed can have economic consequences. McCurley et al. (1984) found a 2-kg decrease in mature BW associated with a 1% increase in inbreeding in beef cattle. In Limousin cattle, postweaning BW gain was estimated to decrease by 0.24 kg for every 1% increase in inbreeding (Gengler et al., 1998), and in dairy cattle, effects of inbreeding were found to be cumulative, decreasing lifetime profit (Smith et al., 1998). Inbreeding levels in American Red Angus cattle are well below a 1% increase per year on average. Therefore, the effects of inbreeding depression should not be as prominent as in breeds with greater levels of inbreeding. Inbreeding in American Red Angus cattle has been increasing only slightly since Although increasing, it is below critical levels for management of genetic diversity. The N e is also large, indicating a sustainable gene pool for the breed. Most grandsires and granddams come from a small number of herds. Therefore, these herds have a disproportionate influence on the future genetics of the breed. Greater rates of inbreeding, with deleterious consequences for the breed, could be observed if too much emphasis is placed on this selection nucleus. The results of this study serve as a baseline estimate of levels of genetic diversity in Red Angus cattle, and can be used to design breeding programs Table 6. Effective number of herds (h e ), and percentage of total herds in the pedigree, supplying animals represented as grandparents in the pedigree Animals 1 h e Percentage of total PGS PGD MGS MGD PGS = paternal grandsires; PGD = paternal granddams; MGS = maternal grandsires; MGD = maternal granddams. that maintain this genetic diversity and sustainability in the breed. LITERATURE CITED Boichard, D., L. Maignel, and É. Verrier The value of using probabilities of gene origin to measure genetic variability in a population. Genet. Sel. Evol. 29:5 23. Bozzi, R., O. Franci, F. Forabosco, C. Pugliese, A. Crovetti, and F. Filippini Genetic variability in three Italian beef cattle breeds derived from pedigree information. Ital. J. Anim. Sci. 5: Burrow, H. M The effects of inbreeding in beef cattle. Anim. Breed. Abstr. 61: Cleveland, M. A., H. D. Blackburn, R. M. Enns, and D. J. Garrick Changes in inbreeding of U.S. Herefords during the twentieth century. J. Anim. Sci. 83: Comstock, S., B. Golden, and R. Bourdon Exploring the RAAA s pedigree. Am. Red Angus 34: Falconer, D., and T. F. C. Mackay Introduction to Quantitative Genetics. 4th ed. Longman Group Ltd., Essex, U.K. Garrick, D. J., B. W. Brigham, and S. E. Speidel Iterative solution of linear equations from national beef cattle evaluation. Proc. West. Sec. Am. Soc. Anim. Sci. 58: Gengler, N., I. Misztal, J. K. Bertrand, and M. S. Culbertson Estimation of the dominance variance for postweaning gain in the U.S. Limousin population. J. Anim. Sci. 76: Golden, B. L., J. S. Brinks, and R. M. Bourdon A performance programmed method for computing inbreeding coefficients from large data sets for use in mixed-model analyses. J. Anim. Sci. 69: Golden, B. L., D. J. Garrick, S. Newman, and R. M. Enns Economically relevant traits: A framework for the next generation of EPDs. Pages 2 13 in Proc. Beef Improv. Fed. 32nd Annu. Res. Symp. Annu. Meet., Wichita, KS. Beef Improv. Fed., Athens, GA. Golden, B. L., W. M. Snelling, and C. H. Mallinckrodt Animal Breeder s ToolKit User s Guide and Reference Manual. Tech. Bull. LTB92-2. Colorado State Univ. Agric. Exp. Stn., Fort Collins. Gutiérrez, J. P., J. Atarriba, C. Díaz, R. Quintanilla, J. Cañón, and J. Piedrafita Pedigree analysis of eight Spanish beef cattle breeds. Genet. Sel. Evol. 35: Harris, D. L Livestock improvement: Art, science, or industry? J. Anim. Sci. 76: Kutner, M. H., C. J. Nachtsheim, J. Neter, and L. William Applied Linear Statistical Models. 5th ed. McGraw-Hill/Irwin, New York, NY. Lacy, R. C Analysis of founder representation in pedigrees: Founder equivalents and founder genome equivalents. Zoo Biol. 8: Márquez, G. C., and D. J. Garrick Selection intensities, generation intervals and population structure of Red Angus cattle. Proc. West. Sec. Am. Soc. Anim. Sci. 58: McCurley, J. R., W. T. Butts Jr., and K. P. Bovard Growth patterns of Angus, Hereford and Shorthorn Cattle. I. Comparison of inbred and noninbred lines, changes in patterns over time and effects of level of inbreeding and reproductive performance. J. Anim. Sci. 59: Mc Parland, S., J. F. Kearney, M. Rath, and D. P. Berry Inbreeding trends and pedigree analysis of Irish dairy and beef cattle populations. J. Anim. Sci. 85: Meuwissen, T. H. E., and Z. Luo Computing inbreeding coefficients in large populations. Genet. Sel. Evol. 24: Nomura, T., T. Honda, and F. Mukai Inbreeding and effective population size of Japanese Black cattle. J. Anim. Sci. 79:

10 68 Márquez et al. Quaas, R. L Computing the diagonal elements and inverse of a large numerator relationship matrix. Biometrics 32: R Development Core Team R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Accessed Oct. 22, Red Angus Association of America The History of Red Angus. Accessed Oct. 22, Robertson, A A numerical description of breed structure. J. Agric. Sci. 43: Roughsedge, T., S. Brotherstone, and P. M. Visscher Quantifying genetic contributions to a dairy cattle population using pedigree analysis. Livest. Prod. Sci. 60: Smith, L. A., B. G. Cassell, and R. E. Pearson The effects of inbreeding on the lifetime performance of dairy cattle. J. Dairy Sci. 81: Sorensen, A. C., M. K. Sorensen, and P. Berg Inbreeding in Danish dairy cattle breeds. J. Dairy Sci. 88: Wright, S Coefficients of inbreeding and relationship. Am. Nat. 56: Young, C. W., and A. J. Seykora Estimates of inbreeding and relationship among registered Holstein females in the United States. J. Dairy Sci. 79: