J. Anim. Breed. Genet.

ISSN 0931-2668

ORIGINAL ARTICLE

Estimates of genetic diversity in the brown cattle population of Switzerland obtained from pedigree information
C. Hagger
Institute of Animal Science, Swiss Federal Institute of Technology, ETH Zurich, Switzerland

Correspondence Christian Hagger, Institute of Animal Science, ETH-Zentrum TAN, CH-8092 Zurich, Switzerland. Tel: +41 (0)44 632 3320; Fax: +41 (0)44 632 1167; E-mail: hagger@inw.agrl.ethz.ch Received: 21 February 2005; accepted: 20 July 2005

Summary The study investigates the genetic diversity present as well as its development in the Brown Cattle population of Switzerland from pedigree information. The population consisted of three subpopulations, the Braunvieh (BV), the original Braunvieh (OB) and the US-Brown Swiss (BS). The BV is a cross of OB with BS where crossing still continues. The OB is without any genetic inuence of BS. The diversity measures effective population size, effective number of ancestors (explaining 99% of reference genome) and founder genome equivalents were calculated for 11 reference populations of animals born in a single year from 1992 onwards. The BS-subpopulation consisted of animals and their known ancestors which were used in the crossing scheme and was, therefore, quite small. The youngest animals were born in 2002, the oldest ones in the 1920s. Average inbreeding was by far the highest in BS, in spite of the lowest quality of pedigrees, and lowest in OB. Effective population size obtained from the difference between average inbreeding of offspring and their parents was, mostly due to the heavy use of few highly inbred BS-sires, strongly overestimated in some BV-reference populations. If this parameter was calculated from the yearly rate of inbreeding and a generation interval of 5 years, no bias was observed and ranking of populations from high to low was OB BV BS, i.e. equal to the other diversity parameters. The high genetic diversity found in OB was a consequence of the use of many natural service sires. Rate of decrease of effective number of ancestors was steeper in BV than OB was, however, equal for founder genome equivalents. Founder genome equivalents were more stable than effective population sizes calculated from the difference between average inbreeding of offspring and parents. The ve most important ancestors contributed one-third of the 2002-reference genomes of BV and OB, in BV all were BS-sires. The relative amount of BS-genes in the BV-genome increased from 59.2% to 78.5% during the 11 years considered.

Introduction Heavy use of articial insemination and the global trade of semen have a strong inuence on the genJ. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

etic diversity of dairy cattle breeds. The easy availability of many sires can help to increase or at least keep the genetic diversity of national populations. The global use of only a few outstanding (for
405

Genetic diversity in brown cattle population of Switzerland

C. Hagger

whatsoever reasons) sires could, however, lead to a decrease of the genetic diversity of a whole breed. Knowledge of the genetic dynamics in a breed is necessary to initiate actions if possible unfavourable trends, like severe reduction of genetic diversity could threaten a population. Such knowledge can be obtained by a closer look at the population pedigree, if available. The rate of inbreeding is a common means to monitor changes of genetic diversity through its transformation to the effective size of a population. This measure has some limitations, because it is dependent on the completeness and the depth of the available pedigrees which often vary greatly between different livestock populations. Preservation of genetic diversity is of primary importance in breeding programmes for rare breeds and for zoo animal populations. Appropriate population parameters to measure genetic diversity were developed for such applications, see, e.g. MacCluer et al. (1986) or Lacy (1989). They measure the contributions of founders (no known ancestors) or of important ancestors to the genome of a reference population, e.g. all animals born in a certain period and were rst adapted and applied to cattle populations by Boichard et al. (1997). Three of these parameters were used in the present investigation to characterize the genetic diversity in the Brown Cattle population of Switzerland, which is a composition of three related subpopulations, the Swiss Brown, the US-Brown Swiss and their cross. The genetic relationship between the USBrown Swiss and the mixed population, where crossing is still going on, was also of interest.

Materials and methods

Data

The Swiss Brown Cattle is a traditional breed in Switzerland. It is the origin of the US-Brown Swiss breed (BS) which dates back to animals imported to the USA between 1869 and 1910 (Engeler 1947). Crossing between BS and several European Braunvieh populations started with the importation of semen in the late 1960s. This practice is still going on today. A subpopulation of the Swiss Brown Cattle without any such crossings was kept in Switzerland and is called the original Braunvieh population (OB), whereas the animals with any amount of BSgenes belong to the much bigger Braunvieh population (BV). These codes were on all records of a set of over 2.5 million individual pedigrees made available by the Swiss Braunvieh Association for this investigation. The youngest animals were born in 2002, the oldest ones in the 1920s. The pedigrees of BS-animals were only available for animals related in some way to the crossing scheme and, thus, stored by the Swiss Braunvieh Association. Therefore, only a limited part of the whole BS-population could be taken into account which explains the small number of these animals listed in Table 1. Only this subpopulation was, however, of interest for the study. Analyses were done for 11 reference populations of each subpopulation according to year of birth, beginning with 1992. The pedigrees for members of each of these reference populations were assembled in an animal list. Each record contained the animals

Table 1 Size, % males and number of sires of reference populations, % males and females from BS bulls in BV populations BV % From BS bulls Reference population 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Animals 74 74 74 74 74 69 72 68 67 112 123 082 319 257 037 315 280 729 285 734 302 061 % Males 3.59 3.45 3.08 2.55 1.87 2.16 3.41 3.55 3.71 38.66 47.66 Sires1 3270 3993 3379 3222 3131 3033 2753 2331 2170 2357 2420 Females 13.0 17.6 24.6 28.4 29.3 38.3 48.5 50.6 48.2 41.4 37.9 Males 29.9 39.1 51.1 60.6 60.3 67.1 62.6 57.2 56.5 39.3 38.0 Animals 2079 2020 1916 1924 2027 2089 2439 2476 2561 3790 4342 % Males 10.39 11.58 10.80 10.40 8.63 11.15 20.91 23.30 21.12 42.37 47.03 Sires1 355 339 314 301 305 307 304 329 328 357 356 Animals 40 46 57 64 86 65 99 137 106 120 106 Sires1 15 18 18 20 30 23 34 34 30 33 37 OB BS

OB, original Swiss Brown; BS, US Brown Swiss; BV, crosses of various degree between OB and BS. Sires of animals in reference populations.

406

J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

C. Hagger

Genetic diversity in brown cattle population of Switzerland

identication, the identication of its parents (if known), year of birth (approximated for some old animals with incomplete records), sex and subpopulation code. Characteristics of reference populations are summarized in Table 1. For reference populations of 1992 until 2000 the gures represent the number of animals entering the herd book of BV and OB. The entries for 2001 and 2002 approach the total number of calves born in BV and OB and, thus, include calves which were not registered later in the herd book. The increase in numbers was due to the introduction of a national identication system for all animals of the cattle species, and, consisted mainly of additional males. The size of the OB-population was about 3.2% the size of the BV-population. The numbers in Table 1 further demonstrate the different breeding structures of the two subpopulations. Until 2000 the number of registered calves of an OB-sire was between 6 and 8 per year, whereas the corresponding numbers for a sire contributing offspring to the BV-population (BV-, OB- and BS-sires) was between 22 and 32. In the year 2000 41.8% of the calves entering the herd book of the OB-population were sired by bulls with one to 10 offspring (80% of the bulls in BV and OB), but only 9.3% in the BVpopulation. The percentage of offspring of the sire with the highest number of offspring in 2000 was, however, very similar in both populations, i.e. 6.3% and 6.1% in OB and BV, respectively. The difference between the breeding structures of the populations is a consequence of the popularity of natural service sires among OB-breeders. The gures in Table 1 also show that in BV (until 2000) registered males were to a larger part sired by BS-bulls than females pointing, thus, to a selection preference of breeders for BS-genes. Methods The relative number of known ancestors per generation was used to indicate the quality of pedigrees of the three populations. The average number of known ancestors of an animal was also deduced. Known ancestors were used to calculate the comPN Pmj 1 1 plete generation equivalent as N j 1 i1 2nij , with mj the total number of known ancestors of animal j, nij the number of generations between j and its ancestor i and N the number of animals in the reference population (Boichard et al. 1997). Three different measures of genetic diversity were investigated from the pedigrees of the reference populations.
J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

The effective population size, Ne

This parameter was obtained from the rate of inbreeding (Falconer 1982). Inbreeding coefcients of parent (Ft)1) and offspring (Ft) populations were required to calculate Ne (1 ) Ft)1)/2(Ft ) Ft)1). Average inbreeding was computed for the members of a reference population and their parents with the method of Meuwissen & Luo (1992). One additional estimate for each subpopulation was obtained by approximating the necessary differences between average inbreeding of two succeeding generations from the regression of average inbreeding of all reference populations on the year and a certain generation interval.
The effective number of ancestors, fa

This number is, as outlined by Boichard et al. (1997), the minimum number of ancestors (founders or not) which are necessary to explain the complete genetic diversity of a population if equal contributions are supposed. Care has to be taken with this approach that only the marginal genetic contribution of an ancestor k, i.e. the contribution, pk, to the genome of the respective reference population, not already explained by related animals considered before, is taken into account. Boichard et al. (1997) presented an algorithm to approximate the marginal contributions needed. The effective number of ancestors is Pf Pf m 2 1 given by fa km 1 p k ; k1 pk is the proportion of the reference genome to be covered and fm is the number of real ancestors over which summation is necessary to obtain the desired proportion. In the present study the minimum number of ancestors to explain 99% of such a genome was determined. It was pointed out by Boichard et al. (1997) that fa does account for the inuence of bottlenecks, a population went through, on genetic diversity. These effects are the major cause of allele loss in modern dairy cattle populations with heavy use of few sires through AI. The method also allowed the estimation of the portion of BS-genes present in the BV-reference populations.
Founder genome equivalents, Ng

This measure of genetic diversity (also called effective number of founder genomes) is a transformation of the frequency of a founders alleles still present in the reference population, gk, where each founder (no known ancestors) bears two distinct alleles, i.e. 2fo alleles altogether in the population with fo
407

Genetic diversity in brown cattle population of Switzerland

C. Hagger

total number of founders. The effective number of P2f o 2 1 founder genomes is Ng 2 k 1 gk . It represents the number of equally contributing founders to the genome of the extant reference population. The parameter does account for the effects of bottlenecks and of random drift on genetic diversity and, thus, will be smaller than fa. A Monte-Carlo method was proposed by MacCluer et al. (1986) to follow the Mendelian path of each founder allele through the generations in a population pedigree to obtain the sought frequencies of the founder alleles. The procedure has to be repeated many times, 1000 were chosen in this study, to get a reliable average for Ng. The outcomes of the repetitions can also be used to compute a variance measure for the parameter and, thus, provide additional information about the range of the most probable values for a given reference population. It was pointed out by Ballou & Lacy (1995) that Ng is the inverse of the average additive relationship coefcient of a population. Using this latter denition would, however, not provide the information about the variance of the parameter. It should be remembered, that the parameters used are all dependent, to different degrees, on the quality (completeness and depth) of the pedigrees available see, e.g. Boichard et al. (1997). Results and discussion The quality of the pedigrees of the three populations considered is visualized by the percentage of known members of ancestor generations for the female (male for BS) reference populations of the year 2002 in Figure 1. Both quality criterions, i.e. completeness

100

Known ancestor (%)

80

60 BV OB BS

40

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Ancestor generation
Figure 1 Percentage of known ancestors per generation of the 2002reference population of three subpopulations of the Brown Cattle of Switzerland.

and depth, are much higher for the OB-population than the BS-population, with the cross (BV) slightly lower until the third ancestor generation and in between in later generations. The curves for older reference populations of all subpopulations (not given) decreased earlier than the ones shown. The curves exhibit the characteristic shape also found for pedigrees of other species, e.g. horse (Zechner et al. 2002) or sheep (Hagger 2002). The complete generation equivalents for the populations considered increased from 6.9, 7.6 and 5.9 of the 1992-reference populations to 8.1, 9.6 and 7.4 in the 2002-populations of BV, OB and BS, respectively. The relative increase was considerably stronger in OB and BS than in BV. The differences in pedigree structure between the three populations were also expressed by the average number of known ancestors. For the 2002-populations these were 1518, 3576 and 914 for BV, OB and BS, respectively. The values obtained might be compared with the numbers in case of all ancestors known up to generation l, which is 2l+1 ) 2, e.g. 2046 for l 10. The numbers obtained for BV and OB, therefore, would result from 9 and 10 completely known and fractions of the 10th and 11th ancestor generations, respectively. The increase in complete generation equivalents and in average number of known ancestors, reect the improvement in pedigree quality over the years. Both of these parameters were much higher than given by So lkner et al. (1998) for the Austrian Braunvieh and three other cattle breeds. Honda et al. (2004) found an increase in complete generation equivalents for the Japanese Black cattle, a beef breed with a considerable amount of AI, from 9.2 in 1990 to 10.2 in 2000, a smaller relative increase than observed here for an equal time interval. The development of average inbreeding from 1992 to 2002 is given in Figure 2 for BV, OB and BS. The highest level of inbreeding was found for the BSreference populations relevant for the development of the present and future BV-population in Switzerland. This is somewhat unexpected, because the depth and after the third ancestor generation also the completeness of the pedigrees was lowest for the BS-subpopulation (Figure 1). The reason for the high inbreeding must be due to relatively few but famous old ancestors which had a great inuence on the relatively small breed, i.e. they appeared in a large part of the pedigrees, which increased inbreeding in spite of many gaps therein. The small number of the available BS animals must be the reason for the considerable variation of the inbreeding averages over the years. The curves of average inbreeding of
J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

408

C. Hagger

Genetic diversity in brown cattle population of Switzerland

6.0 5.0

Table 2 Effective population sizes, Ne, calculated from difference between average inbreeding of reference and their parent population for BV, OB and BS1 Reference population 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
1

Average inbreeding (%)

4.0 3.0 2.0 1.0 0.0 BV OB BS

BV 100.4 106.4 137.0 108.3 81.6 88.1 449.5 180.6 117.6 126.2 113.8

OB 82.2 80.1 82.3 83.0 74.7 90.6 98.8 87.6 83.0 78.3 73.0

BS 42.5 34.5 34.6 20.2 30.6 40.8 44.5 24.0 16.3 28.2 45.9

97

96

98

01

93

95

19 92

Reference year
Figure 2 Development of average inbreeding of reference populations.

99 20 00

94

02

For descriptions see Table 1.

BV and OB are almost parallel and show that in the period considered, crossing had reached a certain state of equilibrium. The decrease of inbreeding, inevitably occurring at the beginning of a crossing scheme (Kearney et al. 2004) had already been regained and the transfer of the higher inbreeding of BS to BV had taken place and is still in progress. From Figure 2 it seems that the rate of inbreeding in BV became slightly steeper for the last 2 years of reference. The regression of average inbreeding on year revealed rates of yearly increases of 0.1364, 0.1154 and 0.2387% for BV, OB and BS, respectively. In the study of Casanova et al. (1992) the development of mean inbreeding of earlier generations of the Swiss Braunvieh appeared not linear and on a considerable lower level. A rate of 0.17% per year after 1992 was observed by Kearney et al. (2004) for the UK Holstein population. Effective population sizes calculated from differences of average inbreeding of the members of reference populations and of their parents are given in Table 2. The least variable estimates were obtained for the OB-population, which is a closed population of still reasonable size and where the number of sires used is, because of natural service, relatively high (Table 1). The structure of this population comes, therefore, closer to the structure of the ideal population, on which the relevant theory is based, than the BV-population. The estimates of Ne for BV show a very distinct outlier in 1998 compared with the other reference populations of this subpopulation. A closer look at sires with many offspring in this population revealed that three BS-sires with 2313, 2480
J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

and 6332 offspring and inbreeding coefcients of 0.1039, 0.0836 and 0.0462, respectively, increased the average inbreeding of the parents substantially without a corresponding increase in their offspring. The result was a very small difference between generations that lead to an unrealistic large estimate of Ne. This clearly demonstrates that the denition of the effective population size as used for the entries in Table 2 does not t cases of crossbred populations too well, as also pointed out by Roughsedge et al. (1999). A more reliable estimate of the difference in average inbreeding between offspring and parent generations and, thus, for Ne of BV might be obtained from the rate of yearly inbreeding and a given generation interval. The reduction of inbreeding at the beginning of the crossing system and the following regain were not visible anymore in the curves of Figure 2 and, thus, should not inuence the approach. An idea about a realistic generation interval could be obtained from OB-results. For this subpopulation the average over all differences in inbreeding between offspring and parents was 0.6061%. Dividing this value by the rate of the yearly increase in average inbreeding of 0.1154% (see above), resulted in a realized generation interval of 5.25 years. Using, thus, an interval of 5 years and the yearly increases mentioned, provided effective population sizes of 70.6, 84.4 and 39.8 for BV, OB and BS, respectively. With this approach the effective size of the OB-subpopulation was higher than of the BV-subpopulation, in spite of the much larger census size of the latter (Table 1). A higher generation interval than supposed would result in a lower effective size. The effective sizes of the BS-reference
409

Genetic diversity in brown cattle population of Switzerland

C. Hagger

populations are, probably due to the limited data available, quite variable. So lkner et al. (1998) reported for the Austrian Braunvieh a similar effective size as given in Table 2 for BV. Their value could, however, be biased upwards for a similar reason as was found here, because crossing with BS took place also in the Austrian Braunvieh. The development of effective number of ancestors, fa, over time, which explained 99% of the genome of the reference populations of BV, OB and BS are given in Figure 3. They depict a gradual decrease from 1992 until 2002 for BV and OB, with a slightly steeper rate in BV. Higher numbers were observed for OB than for BV as already seen for the effective size obtained from the yearly rate of inbreeding. The curves for the two populations diverged only after 1993. A decrease of fa to 64.8% and 75.1% of the 1992 values occurred until 2002 in BV and BS, respectively (Figure 3). This difference must be a consequence of the different breeding structures in the two subpopulations. The effective number of ancestors found by So lkner et al. (1998) for the Austrian Braunvieh was higher than in BV. The shorter pedigrees available there and the observed reduction over time here suggest, that the depth of the pedigrees does inuence fa. The effective number of ancestors of the BS-subpopulation seemed to stabilize around a value of 14. Calculation of fa requires ordering ancestors according to their marginal genetic contribution to the genome of a reference population. The ve ancestors with the largest marginal contributions to the reference genome are listed in Table 3 for BV and OB for the rst and last reference populations considered. Between 1992 and 2002 only 10 different animals, partly shown in Table 3,

Table 3 Identication (ID) and relative genetic contributions (%) of the ve most important ancestors of reference populations 1992 and 2002 for the BV and OB populations BV1 1992 ID BS1 BS6 BS2 BS5 BS10 Total
1

OB 2002 % 7.31 6.74 6.24 5.17 3.28 28.72 ID BS1 BS8 BS7 BS6 BS2 % 10.41 9.63 5.71 5.15 4.18 35.08 1992 ID OB1 OB2 OB3 OB9 OB5 % 8.40 7.26 5.50 4.64 3.77 28.53 2002 ID OB1 OB2 OB3 OB4 OB8 % 10.53 6.40 5.78 5.34 4.59 32.64

For descriptions see Table 1.

45.0

40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 1992 93 94 95 96 97 98 99 2000 01 02 BV OB BS

Reference year
Figure 3 Effective number of ancestors explaining 99% of reference genomes of subpopulations.

were found among the ve most important ancestors in all BV reference populations. All of them were BS-bulls and BS2 was the sire of BS8, which means that these ancestors represented <10 unrelated individual genomes. Within reference population, the ve most important ancestors contributed between 28.7% and 35.1% to the genome. The 10 most important ancestors in the Austrian Braunvieh (So lkner et al. 1998) explained the same amount of the reference genome as the ve most important ancestors in BV. The most important ancestor contributed roughly 10% of the reference genome in the BV and the Austrian Braunvieh. It is interesting to note in Table 3 that ancestors BS1, BS2 and BS6 were among the ve most important ancestors in reference populations 1992 and 2002. The total number and the genetic contributions of the most important ancestors to the genome of the reference populations for OB is similar to the pattern observed in BV. In OB, nine different animals were found among all the ve most important ones of the reference populations. The three ranked rst in 1992 occupied the same ranks in 2002 (Table 3). The genome of the OB-subpopulation contains quite a stable contribution of these three animals. The rst ve ancestors explained between 28.5% and 32.6% of the reference genomes. The largest single contribution was 10.5% from animal OB1 in 2002. Summation of the marginal genetic contribution of BS-ancestors in the BV reference populations provided the percentage of BS-genes in the crossed population. Figure 4 shows still an increase of this share during the period considered. The genome of the BV-2002 reference population contained more than 78% BS genes. The percentage of animals descending from BS-sires in the reference populations listed in Table 1 suggest, that the total amount of BS-genes
J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

410

Effective number of ancestors (fa)

C. Hagger

Genetic diversity in brown cattle population of Switzerland

100.0

80.0

60.0

40.0

20.0

0.0 1992 93 94 95 96 97 98 99 2000 01 02

Reference year
Figure 4 Percentage of BS-genes in BV-reference genomes.

in the BV-genome will increase further. The spread of these genes through the BV-males was accentuated by the preference of breeders for sons of BSsires. The Monte-Carlo estimates of founder genome equivalents and their standard deviations are in Table 4 for BV, OB and BS. The gures are averages of 1000 replicates. The values of this parameter were about half the corresponding effective number of ancestors and they decreased steadily to 62.8% and 63.1% of the 1992 values in 2002 in BV and OB, respectively. The relative decrease was not different between BV and OB and, thus, contrary to the relative decrease in effective number of ancestors observed during the same period. The numbers for BV and OB show a nearly parallel course over the

Table 4 Number of founder genome equivalents, Ng, and standard deviations present in reference populations (1000 replications) of BV, OB and BS1 Reference population 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
1

BV 19.9 19.2 18.4 16.6 15.9 15.0 13.5 13.5 13.9 13.4 12.5 (3.96) (4.00) (3.73) (3.37) (3.01) (2.86) (2.57) (2.69) (2.75) (2.51) (2.32)

OB 25.5 23.8 23.8 23.1 21.9 20.2 19.3 18.8 18.3 17.4 16.1 (3.29) (3.01) (2.97) (3.06) (3.01) (2.76) (2.79) (2.53) (2.57) (2.56) (2.50)

BS 7.1 7.1 7.0 5.6 7.1 5.7 6.4 6.3 6.0 6.4 6.0 (1.40) (1.37) (1.40) (1.06) (1.39) (1.15) (1.32) (1.30) (1.16) (1.25) (1.24)

For descriptions see Table 1.

years similar to, but in the opposite direction than the course of inbreeding (Figure 2) and, thus, support the connection between this parameter and the average relationship coefcient derived by Ballou & Lacy (1995). The corresponding estimates for BS varied only slightly around a value of 6.4. So lkner et al. (1998) reported a considerably higher Ng for the Austrian Braunvieh and two other dairy breeds. The Ng-diversity measure of a fourth breed with deeper pedigrees available was of roughly the same size as the value for the OB 1997-reference population. Boichard et al. (1997) reported similar Ng values for two French dairy breeds as found here. Honda et al. (2004) observed in the Japanese Black cattle a reduction of the Ng-diversity from 12.6 to 7.3 or to 57.9% in an 11-year period, which was above the reduction observed here. The Ng-diversity of that breed was of similar size as found for the BS, the TuxZillertal cattle breed described by Baumung & So lkner (2002) or the Lipizzan horse breed (Zechner et al. 2002), despite a much larger census size. The heavy use of few males possible through AI always counteracts the effect of a large census size. The standard deviations of the estimated numbers of founder genome equivalents in the reference populations are given in Table 4. If the Ng values were normally distributed, 95% of the Monte-Carlo outcomes for the 2002-reference genomes would be contained in the intervals 7.917.1, 11.121.1 and 4.58.5 for BV, OB and BS, respectively. The variation of the estimated Ng-diversity is, thus, considerable and makes inferences about its real size for a population perhaps slightly less precise than anticipated. It is to be supposed that this fact is also valid for the Ne-diversity measure. It was found, that the time span covered by a reference population did not affect decisively the diversity measures used. If only two successive reference populations of 5 years each were considered in the OB-subpopulation, estimates of Ne, fa and Ng were 81.9, 81.4, 39.7, 34.4, 23.4 and 18.5, respectively. The values were close to the averages of the corresponding 1-year reference populations did, however, not provide the development of the parameters. Comparison with the studies cited, where different time spans for reference populations were applied will, therefore, be valid. Rate of inbreeding might, because of several reasons, not be the most appropriate tool to measure genetic diversity in livestock populations. Overlapping generations, articial selection or heavy use of few sires through AI, render the search for a reliable estimate of the increase in inbreeding between
411

J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

% BS-genes

Genetic diversity in brown cattle population of Switzerland

C. Hagger

successive generations rather difcult. This quantity was approximated in most publications by taking only a single difference between the average inbreeding of offspring and their parents. If, e.g. few sires were heavily used and these sires were either very lowly or very highly inbred, the difference between inbreeding of parents and offspring would be biased. This effect was detected in the crossed BV-subpopulation at a stage, where the yearly increase of average inbreeding had reached linearity again. The effect had been accentuated by the considerable higher inbreeding in the BS-parent population. It was shown by Boichard et al. (1997) that Ne was also more affected by incomplete pedigrees than the number of effective ancestors and the number of founder genome equivalents, both of which are based on the origin of alleles. If the rate of inbreeding of a population could be estimated from a series of mean inbreeding coefcients of animals born in successive years, then it seemed that some of the drawbacks just mentioned would be mitigated and more reliable estimates for the effective population size could be obtained. The diversity measures used depend not only on the quality of the pedigrees but also on the breeding structure as shown by Boichard et al. (1997), So lkner et al. (1998) and by the differences observed between BV and OB. Comparison of the diversity parameters between breeds or between countries needs, thus, some cautions. In most investigations it was found that the genetic diversity of dairy cattle populations was, due to the heavy use of few sires quite low and, a further decrease was expected. New selection strategies have been proposed in the last few years with the aim to slow down this undesirable trend. One approach was to maximize selection response under a constraint on the rate of inbreeding, e.g. Meuwissen (1997) or Grundy et al. (2000). Subdivision of the Austrian Simmental population into more or less distinct subpopulations also led to higher genetic diversity compared to breeding the population as an entity (So lkner et al. 1998). A larger diversity was observed for populations where natural service sires were of some genetic importance as, e.g. in beef cattle (Boichard et al. (1997) or in the OB population, which had a larger estimated diversity than BV in spite of a much smaller census size. Continuous use of substantial amounts of BSsemen in the BV-population has resulted in a strong displacement of the original genes of the Braunvieh population with BS-genes. Beginning with birth year 1999 the BV-genome nowadays consists of more
412

than three quarters of BS-genes. Similar displacements took place in the other Braunvieh populations in Europe. They had been relatively independent before, with only a limited exchange of genetic material among them. Crossing with US-Brown Swiss, a population of relatively small effective size, must now have led to strong genetic links between these populations and, therefore, to a considerable reduction of their combined precrossing genetic diversity. Conclusions Completeness and depth of pedigrees inuence effective population size, effective number of ancestors and founder genome equivalents, measures of genetic diversity of livestock populations. Level of inbreeding in the BS-population was, however, much higher than in the OB-population in spite of the higher percentage of unknown ancestors in the former. The use of quite many natural service sires in the OB-population was the reason for the considerable higher values of the diversity parameters compared with the BV-population with a much larger census size. The effective population size obtained directly from the inbreeding difference between offspring and parents has drawbacks if few sires are heavily used through AI or crossing occurs in a population. Founder genome equivalent appeared more robust against deviations from the ideal case than effective population size. Comparisons of genetic diversity between populations and between countries should be done cautiously, because completeness and depth of pedigrees can vary strongly. Genetic diversity of dairy cattle populations should be monitored carefully to avoid a possible, but mostly undesired, strong reduction of it. Acknowledgements The continuous support of G. Stranzinger, the interest of M. Schneeberger for this topic, the provision of the data by the Swiss Braunvieh Association and of programs by D. Boichard are greatly appreciated. References
Ballou J.D., Lacy R.C. (1995) Identifying genetically important individuals for management of genetic variation in pedigreed populations. In: J.D. Ballou, M. Gilpin, T.J. Foose (eds), Population Management for Survival and Recovery. Columbia University Press, New York, pp. 76111.

J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

C. Hagger

Genetic diversity in brown cattle population of Switzerland

Baumung R., So lkner J. (2002) Analysis of pedigrees of Tux-Zillertal, Carinthian Blond and Original Pinzgau cattle populations in Austria. J. Anim. Breed. Genet., 119, 175181. Boichard D., Maignel L., Verrier E. (1997) The value of using probabilities of gene origin to measure genetic variability in a population. Genet. Sel. Evol., 29, 523. Casanova L., Hagger C., Kuenzi N., Schneeberger M. (1992) Inbreeding in Swiss Braunvieh and its inuence on breeding values predicted from a repeatability animal model. J. Dairy Sci., 75, 11191126. Engeler W. (1947) Das Schweizer Braunvieh. Verlag Huber & Co., Frauenfeld, Switzerland. Falconer D.S. (1982) Introduction to Quantitative Genetics, 2nd edn. Longman, London and New York. Grundy B., Villanueva B., Woolliams J.A. (2000) Dynamic selection for maximizing response with constrained inbreeding in schemes with overlapping generations. Anim. Sci., 70, 373382. Hagger C. (2002) Genetic variability of two Swiss sheep breeds derived from pedigree information. In: Proceedings of 7th World Congress on Genetics Applied to Livestock Production. CD-ROM communication no. 26-21, INRA and CIRAD, Montpellier, France. Honda T., Nomura T., Yamaguchi Y., Mukai F. (2004) Monitoring of genetic diversity in the Japanese Black cattle population by the use of pedigree information. J. Anim. Breed. Genet., 121, 242252.

Kearney J.F., Wall E., Villanueva B., Coffey M.P. (2004) Inbreeding trends and application of optimized selection in the UK Holstein population. J. Dairy Sci., 87, 35033509. Lacy R.C. (1989) Analysis of founder representation in pedigrees: founder equivalents and founder genome equivalents. Zoo Biol., 8, 111123. MacCluer J.W., VandeBerg J.L., Read B., Ryder O.A. (1986) Pedigree analysis by computer simulation. Zoo Biol., 5, 147160. Meuwissen T.H.E. (1997) Maximizing the response of selection with predened rate of inbreeding. J. Anim. Sci., 75, 934940. Meuwissen T.H.E., Luo Z. (1992) Computing inbreeding coefcients in large populations. Genet. Sel. Evol., 24, 305313. Roughsedge T., Brotherstone S., Visscher P.M. (1999) Quantifying genetic contributions to a dairy cattle population using pedigree analysis. Livest. Prod. Sci., 60, 359369. So lkner J., Filipcic L., Hampshire N. (1998) Genetic variability of populations and similarity of subpopulations in Austrian cattle breeds determined by analysis of pedigrees. Aim. Sci., 67, 249256. Zechner P., So lkner J., Bodo I., Druml T., Baumung R., Achmann R., Marti E., Habe F., Brem G. (2002) Analysis of diversity and population structure in the Lipizzan horse breed based on pedigree information. Livest. Prod. Sci., 77, 137146.

J. Anim. Breed. Genet. 122 (2005) 405413 2005 Blackwell Verlag, Berlin

413