Population structure and genealogical analysis of the Brazilian Crioula Horse

Animal Genetic Resources, page 1 of 11. Food and Agriculture Organization of the United Nations, 2014 doi:10.1017/s2078633613000489 Population structure and genealogical analysis of the Brazilian Crioula Horse F.C. Maciel 1, C.D. Bertoli 2, J. Braccini Neto 2, J.A. Cobuci 2, S.R. Paiva 3 and C.M. McManus 2,4 1 Faculdade de Veterinária, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil; 2 Departamento de Zootecnia, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil; 3 EMBRAPA Recursos Genéticos e Biotecnologia, Brasília, DF 70770-970, Brazil; 4 Universidade de Brasília, Brasília, DF 70910-900, Brazil Summary A genealogical analysis of Crioula horses registered by the Brazilian Association of Crioula Horse Breeders was performed. The state of Rio Grande do Sul is the largest producer of animals with 89.85 percent of registered animals, of which 45.29 percent were males and 54.71 percent females. The inbreeding coefficient was calculated at 0.88 percent and the average relatedness was 0.65 percent in the total population (animals born in Brazil and imported). Inbreeding increased over the generations. An increase was seen in the average genetic conservation index in animals born after the year 1927 (0) until 2010 (8.67 percent). The average generation interval was 10.3 years and the average age of parents when offspring born were 10.5 years, falling in recent years. The effective population size of founders was 95.19 animals, the number of ancestors contributing to this population was 5086 where 56 ancestors explained 50 percent of the genetic diversity of the breed. Inbreeding is under control in the Crioula horse. The increase in registrations reflects the increased interest from farmers in this breed. Keywords: average relatedness, founding, ancestors, genetic conservation index, inbreeding, statistics of Wright Résumé Une analyse généalogique des chevaux Créoles inscrits à l Association Brésilienne d Éleveurs de Chevaux Créoles a été réalisée. L état du Rio Grande do Sul est le plus grand producteur d animaux avec 89,85 pour cent des animaux inscrits, dont le 45,29 pour cent sont mâles et le 54,71 pour cent femelles. Le coefficient de consanguinité a été estimé à 0,88 pour cent, avec une parenté moyenne de 0,65 pour cent dans la population totale, y compris les animaux nés au Brésil et ceux importés. La consanguinité a augmenté au fil des générations. Un accroissement de l indice moyen de conservation génétique a été observé pour les animaux nés entre 1927 (0) et 2010 (8,67 pour cent). L intervalle générationnel moyen a été de 10,3 ans et l âge auquel les parents ont en moyenne leur premier descendant a été de 10,5 ans, cet âge ayant diminué au cours des dernières années. La taille effective de la population des fondateurs a été de 95,19 animaux. Le nombre d ancêtres ayant contribué à cette population a été de 5086, dont 56 sont responsables du 50 pour cent de la diversité génétique de la race. La consanguinité est sous contrôle dans la race équine Créole au Brésil. L augmentation des inscriptions reflète l intérêt croissant des éleveurs pour cette race. Mots-clés: consanguinité, parenté moyenne, indice de conservation génétique, statistiques de Wright, ancêtres fondateurs, généalogie, cheval créole, Brésil Resumen Se realizó un análisis genealógico de los caballos Criollos registrados en la Asociación Brasileña de Criadores de Caballo Criollo. El estado de Río Grande del Sur es el mayor productor de animales con el 89,85 por ciento de los animales registrados, de los cuales el 45,29 por ciento son machos y el 54,71 por ciento hembras. El coeficiente de endogamia se estimó en un 0,88 por ciento, con un parentesco medio de 0,65 por ciento en la población total, contando tanto con animales nacidos en Brasil como con animales importados. La endogamia aumentó de generación en generación. Se detectó un incremento en el índice medio de conservación genética en los animales nacidos entre 1927 (0) y 2010 (8,67 por ciento). El intervalo generacional medio fue de 10,3 años y la edad media de los progenitores al nacimiento del primer descendiente fue de 10,5 años, habiéndose reducido esta edad en los últimos años. El tamaño efectivo de la población de fundadores fue de 95,19 animales. El número de ancestros que han contribuido a esta población fue de 5086, de los cuales 56 eran responsables del 50 por ciento de la diversidad genética de la raza. La endogamia está bajo control en la raza equina Criolla en Brasil. El aumento de registros refleja el interés creciente de los ganaderos por esta raza. Palabras clave: endogamia, parentesco medio, índice de conservación genética, estadísticos de Wright, ancestros fundadores, genealogía, caballo criollo, Brasil Submitted 19 April 2013; accepted 17 October 2013 Correspondence to: C.M. McManus, Universidade de Brasília, Brasília, DF 70910-900, Brazil. email: concepta@unb.br 1

2 F.C. Maciel et al. Introduction Brazil has the third largest herd of horses in the world with ~5.9 million animals, second only to China with 7.9 million and Mexico with 6.2 million. The equine generates about 640 000 direct jobs and if we consider indirect jobs about 3.2 million, according to ESALQ/USP (2006). The Crioula horse of southern Latin America is a direct descendant of the horses brought to the New World by the Spanish and Portuguese conquistadores during the sixteenth century (Rodero, Delgado and Rodero, 1992). Many horses escaped or were abandoned, and returned to a wild state. These were Portugese, Barbe and Spanish (particularly Andaluse) horses. Studies with Uruguayan (Kelly et al., 2002) and Argentinian Crioulas (Mirol et al., 2002) using blood group and protein polymorphism variants that are considered to be breed markers of Spanish Pure-bred and Barb horses were detected in the Crioula breed but some microsatellites and protein polymorphisms alleles were found uniquely in the Crioula horse. Argentinean Crioula horses shared two haplotypes with the Peruvian Paso from Argentina, and the commonest haplotype of the Crioula horses is identical to one of the Andalusian horses. Even when there was substantial subdivision between breeds with highly significant Wright s Fixation Index (F ST ), the parsimony and distance-based phylogenetic analyses failed to show monophyletic groups and there was no clear relationship in the trees between the South American and any of the other horses analysed. Although this result could be interpreted as mixed ancestry of the South American breeds with respect to the Spanish breeds, it is probably indicating the retention of very ancient maternal lineages in the breeds analysed. The South American horse breeds form two main groups, the Crioula types (including Brazilian, Argentinean and Uruguayan) and the other Brazilian breeds (Cothran et al., 1998) and the Crioula is most closely related to Iberian Sorraia, Andalusian, Paso Fino and Lusitano breeds. The pairing of the Sorraia with the Argentine Crioula was probably due to the combination of low variability of the Sorraia, small sample size of the Argentine Crioula and the fact that an Argentine Crioula stallion was used in the regeneration of the Sorraia. According to the Brazilian Association of Crioula Horse Breeders (ABCCC), from 2001 until the end of 2010 there was an increase of 143.7 percent in applications for registration of temporary records, jumping from 10 271 in 2001 to 25 020 in 2010. Based on this growth, the Association estimates that this number could reach 57 502 in the year 2020, an increase of approximately 129.82 percent over 2010. Trade of animals increased 985 percent between 2001 and 2010, rising from $9.2 to $100 million, respectively, and this number could exceed $300 million in the year 2020, according to data from the ABCCC. According Laat (2001), population aspects of horse herds, especially those related to inbreeding and effective size, present peculiarities since the horse herds are generally small, selection does not have well-defined goals and economic characteristics to evaluate progress, and sports prizes, morphology and associated functionality are not objectively measured. Furthermore, the low uses of reproductive biotechnologies such as artificial insemination, embryo transfer and IVF which are prohibited by the Association of Crioula breeders, affect population parameters. Despite its importance, no studies on the level of inbreeding and population structure of the Crioula breed and, therefore, the objective of this study was to analyse the current situation of the breed using population genetics parameters, focusing especially on some aspects of genetic variability and population structure. Material and methods The analysis was based on data from the studbook where the population was 341 616 animals that included all registered animals (final and interim) by the Brazilian Association of Crioula Horse Breeders (ABCCC) since its foundation up to 8th April 2011 (154 663 males and 186 843 females). Data were analysed using ENDOG programme to calculate population parameters for the breed. The following parameters were calculated: the pedigree completeness level was computed as the proportion of ancestors known per parental generation (MacCluer et al., 1983); The number of equivalent to discrete generations (t) for each individual in a pedigree (Boichard, Maignel and Verrier, 1997); The inbreeding coefficient (F) (Malécot, 1948); average relatedness coefficient (AR) (Goyache et al., 2003; Gutiérrez, Goyache and Cervantes, 2009). The probability of gene origin was characterized by computing the following parameters: effective number of founders ( f e ) (James, 1977) computed from the genetic contribution of founders to the descendant gene pool of the population (Lacy, 1989); effective number of ancestors ( f a ), defined as the minimum number of ancestors, not necessarily founders, explaining the complete genetic diversity of a population (Boichard, Maignel and Verrier, 1997); the founder genome equivalents (Ballou and Lacy, 1995) obtained by the inverse of twice the average coancestry of the individuals within the population (Caballero and Toro, 2000); Effective population size (N e ) was computed following Gutiérrez, Goyache and Cervantes (2009); F-statistics where a population has a population structure of two levels; one from the individual (I ) to the subpopulation (S) and one from the subpopulation to the total (T ). F-statistics describe the amount inbreeding-like effects within subpopulations F IS, among subpopulations F ST, and within the entire population F IT. In this case Wright s F-statistics are obtained according to Caballero and Toro (2000, 2002); The within-breed coancestry ( f ii ) and the between-breeds coancestry matrix

Population structure of the Brazilian Crioula Horse 3 ( f ij ) were computed averaging all pairwise coancestry coefficients of the individuals belonging, respectively, to a given breed i or to two different breeds i and j. Following Caballero and Toro (2000, 2002) the betweenbreeds Nei s minimum distance (D m ) matrix was also computed as D m =((f ii + f jj )/2) f ij where f ii and f jj are the average coancestry within two breeds i and j and f ij the coancestry between two breeds i and j. In the present study, municipalities, farms and sires were considered separate populations; the effective number of founders in an animal s pedigree was calculated using the genetic conservation index, genetic conservation index (GCI; Alderson, 1991): GCI =1/ΣP 2 i where P i is the proportion of genes of founder animal i in the pedigree. Subsets were also analysed using sires with more than 100 offspring (803) giving 163 817 foals, the fathers of these animals (196) giving 53 632 foals. Farms (106) and municipalities with more than 500 horses registered were also investigated. Two further reference populations were created families of stallions with more than 500 offspring (184 005 animals FAMILY) and animals registered since 1998 (184 669 animals RECENT). In the latter two populations 135 449 animals appeared in both. Patterns of inbreeding in the three populations studied (ALL, RECENT, FAMILY). Levels of inbreeding were grouped (0.00, 0.001 0.05 (0.05), 0.051 0.10 (0.10), 0.101 0.15 (0.15), 0.151 0.20 (0.20), 0.201 0.25 (0.25) and >0.25 (0.30). Figure 1. Geographical distribution of the population (Pop) of Crioula horses in Brazil. Results The Crioula horse is present in 23 of the 27 states of the Federative Republic of Brazil (Figure 1). The Crioula horse does not have animals born in states of Alagoas, Maranhão, Rio Grande do Norte and Amapá. The largest producers are the three southern states of the country that together add up to 96.34 percent, and Rio Grande do Sul, the largest producer with 89.85 percent, 3.81 percent in Paraná and Santa Catarina with 2.67 percent of the population total. Horses were also imported from Argentina (866), Chile (868) and Uruguay (852) of which ~80 percent were females. While horses from Argentina and Uruguay have been imported since the start of breed registration, those from Chile have been more recent becoming the most important source of importation (Figure 2). Of all registered animals, 97.98 percent have their pedigrees fully known. The number of known pedigrees has increased over the generations with emphasis on the growth of the ninth generation. In the following, six generations this increased from 4.95 percent of the known pedigrees to 97.98 percent in the current 15th generation (Figure 3). The increase in the number of records per year is evident (Figure 4). Since the 1970s there has been a steady growth in the breed until the mid-1990s, more precisely in 1992. Before 1980, 9.3 percent of all animals were registered. Figure 2. Number of Crioula horses imported per year from Uruguay, Argentina and Chile. In this year, there was a decrease in the number of records of animals to 1996. From 1996 to 1998 the number of records was almost stable, and from 1999 there was a Figure 3. Percentage of known generation pedigrees of the Crioula horse in Brazil.

4 F.C. Maciel et al. Figure 4. Percentage of records by year of birth for the Crioula horse in Brazil. higher growth than seen in the 1970s and 1980s. From the year 2005, growth has intensified, with an increase in 3042 animals from 2006 to 2007. The numbers of records for the years 2010 and 2011 should be disregarded, since the data were collected in April 2011 and most of the animals born in the spring of 2010 onwards had not been through technical inspection on the date on which the data were gathered. Of the animals recorded by the ABCCC, 54.71 percent are female and 45.29 percent are males, in a total of 341 497 observations. As for the month of birth, 90.71 percent of registered animals are born between September and January, the largest concentration in the months of October and November where we find, respectively, 25.85 and 26.44 percent of births (Figure 5). Farms with the largest number of animals were registered for Estância Nazareth (3930), Cabanha Tupambaé (3114), Figure 5. Percentage of records by month of birth for the Crioula horse in Brazil. Cabanha Paineiras (2994), Cabanha St. Angelo (2370) and Estância San Francisco (2252), and these five farms, from a total of 7949, were responsible for 4.5 percent of registered animals. When you consider the 50 largest farms, they, together, represent 19.67 percent of the animals reared in the country, reaching 63 977 for a total of 325 301 animals. The inbreeding coefficient was calculated at 0.88 percent and the AR was 0.65 percent in the total population (animals born in Brazil and imported). With the increase of generations, an increase in inbreeding was observed (Table 1). This increase is evident from the first generation, and especially observed in the fourth generation full generation. 96.58 percent of the animals which have some degree of inbreeding and inbreeding coefficient in this generation was 2.78 percent. There is an increasing trend up to 100 percent of animals in the 15th generation being inbred, Table 1. Inbreeding (F), average relatedness (AR) and effective population size (Ne) per generation in Brazilian Crioula Horse. Generation No. of animals F Mean (%) % Endogamic animals F Mean for endogamic animals % Mean AR Ne Endogamy per generation 1 0 5882 0.00 0.01 1 7157 0.00 0.16 2 7794 0.68 3.77 18.15 0.57 73.0 3 10 997 0.93 8.11 11.41 0.88 206.5 4 15 147 1.08 14.72 7.33 0.90 323.2 5 23 019 0.94 18.73 5.04 1.04 6 38 604 0.93 27.46 3.40 1.35 7 59 015 1.10 41.55 2.65 1.60 2117.7 8 67 583 1.30 57.17 2.28 1.79 243.7 9 56 469 1.44 68.72 2.10 1.96 353.2 10 34 183 1.72 78.87 2.18 2.14 180.6 11 13 532 1.85 84.85 2.18 2.24 370.4 12 2004 2.16 85.83 2.52 2.24 155.9 13 186 1.99 87.10 2.29 2.18 14 42 2.27 95.24 2.38 2.13 461.7 15 2 13.10 100.00 13.10 1.30 4.5 Endogamy per complete Generation 2 0 7901 0.00 0.08 1 64 559 0.27 2.39 11.12 0.83 188.0 2 18 1704 1.25 48.33 2.59 1.77 50.4 3 84 471 1.89 80.99 2.33 1.89 77.7 4 2981 2.78 96.58 2.88 1.69 55.0 1 Maximum number of generations traced. 2 Complete generations traced as defined by Gutiérrez, Goyache and Cervantes (2009).

Population structure of the Brazilian Crioula Horse 5 Table 2. Average number of generations, inbreeding per generation and effective population size for the Brazilian Crioula horse. Mean generations Increase in endogamy (%) Effective population size All Maximum 12.09 0.05 994.64 Complete 1.85 0.50 99.18 Equivalent 4.62 0.23 214.46 Family Maximum 7,15 0.14 348.53 Complete 2.03 0.76 65.45 Equivalent 3.40 0.54 93.28 which is not complete since the animals can be registered definitively up to 4 years after birth. The number of generations, complete and equivalent, shows that the breed is expanding (Table 2). Large variations in effective herd size per generation are an indication of animals being imported into the herd (for example, from Uruguay, Chile or Argentina). The highest level of inbreeding was 0.41 with 6755 animals with a value >0.1. These animals were on average in the seventh generation, but there were animals in the second generation with high levels of inbreeding. Analysing the full herd, 72 (0.02 percent) animals were from crosses between full-sibs, 3047 (0.89 percent) between half-sibs and 1784 (0.52 percent) between parents and offspring. The mean coefficient of individual relatedness showed an increase in population over the years, especially after the1950s, reaching 2.09 percent in animals born in the year 2010 (Figure 6). An increase in average individual relatedness was found from the 1940s until 1970, the largest observation in 1955 was 3.11 percent. In recent years, there was an average increase of inbreeding about 0.50 percent (Figure 7). An increase in the average individual coefficient of inbreeding was found in animals born after the 1940s, and in 1955 an average rate of 2.95 percent was seen. Figure 7. Genetic conservation index (GCI) and average relatedness (AR) average per year of birth for the Brazilian Crioula horse. After 1970 there was a reduction in the coefficient of average individual inbreeding, and the animals born in the past decade had rates of ~1.50 percent (Figure 7). An increase in the average rate of genetic conservation was seen in animals born after the year 1927 until 2010, which had the highest average GCI, 8.67 percent (Figure 7). The interval between generations (Table 3) was higher for mares than stallions. This indicates that mares are kept longer for reproduction or start reproductive life later. There was no significant genetic differentiation between the states of the federation where animals are raised when considered sub-populations (not shown). Average coancestry within subpopulations was 0.0084 and average coancestry in the metapopulation was 0.0079. Wright parameters show that the inbreeding in the population is greater than expected if matings were performed at random and which occur between the mating subpopulations. The effective population size of founders was 95.19 animals, expected inbreeding due to an imbalance in the contribution of founders was calculated as 0.53 percent and the mean inbreeding coefficient computed was 1.21 Table 3. Generation interval and the average age of parents when their offspring born for the Brazilian Crioula horse. Relation N L Se Generation interval Sire Son 10.580 9.986 0.051 Sire daughter 74.037 9.737 0.018 Dam son 10.647 10.611 0.045 Dam daughter 74.006 10.781 0.018 Total 169.270 10.264 0.012 Mean age of parents when offspring born Sire Son 153.294 9.880 0.013 Sire daughter 180.964 9.983 0.012 Dam son 153.901 11.042 0.012 Dam daughter 181.173 11.097 0.011 Total 669.332 10.504 0.006 Figure 6. Mean individual inbreeding (F) and Average individual change in inbreeding (AF) by year of birth for the Brazilian Crioula horse. N is the number of animals, L is the generation interval and SE is the standard error.

6 F.C. Maciel et al. Table 4. Measures of genetic variation and Wright s statistics (F IS, F ST and F IT ) for Brazilian Crioula Horse. Parameter All Recent Family Mean coancestry within 0.008414 0.009137 0.010384 Subpopulations Autocoancestry 0.506043 0.506043 0.506043 Endogamy 0.012087 0.012087 0.012087 Nei s Distance 0.000516 0.001238 0.002486 Mean Coancestry in Metapopulation 0.007898 0.007898 0.007898 F IS 0.003704 0.002977 0.001721 F ST 0.000520 0.001249 0.002505 F IT 0.004222 0.004222 0.004222 percent for the population of 341 616 animals and the base population (animals with one or more parents unknown) was 7901 (Table 4). The ratio of founders to ancestors was 1.21 (91/75) meaning that in general no specific lineages were created within the Crioula population. The reference population (both parents known) was smaller than that used for the analysis of founders. The number of ancestors contributing to this population was 5086 for a reference population of 333 715 animals, where 56 ancestors explained 50 percent of the genetic diversity of the breed. Animals 9838, 5174, 18 182, 5178, 5532, 5497, 5208, 5418, 5439 and 5449 had the largest contribution to the herd, reaching ~25 percent of the total genetic variation. The largest number of offspring per stallion was 1428 (sire 63 592) with four animals (9838, 97 059 and 5208) with more than 1000 offspring. Animals 63 592 and 97 059 are offspring of 9838, and eight stallions had more than 500 offspring (forming the second reference population FAMILY). Animal 9838 had the highest average relatedness. Mare 49 893 had the most offspring (20), but more than 12 292 mares had more than ten offspring, showing the prolificacy and longevity of the breed. The total population of animals was divided by their state of birth and these subpopulations were analysed for type (Table 5) and structure (Table 6). Only Commercial and Multiplier herds were seen, with no nuclei or disconnected herds. Most multiplier states (14/ 15) use their own breeding males, buy in animals and sell sires to other states. As for commercial herds, 63.64 Table 5. Number of herds by type and characteristics for the Brazilian Crioula horse. Type Buy-in sires Use own sires Sell sires No. of states % Sires bought Nucleus No Yes Yes 0 0.00 Multiplier Yes Yes Yes 14 24.22 Multiplier Yes No Yes 1 100.00 Commercial Yes Yes No 4 93.69 Commercial Yes No No 7 100.00 Isolated No Yes No 0 0.00 Table 6. Structure of the Brazilian Crioula Horse herds by state. State 1 Region Type % parents born in state/ number of births % males born in state which are sires/number of sires Acre N Commercial 0.00 0.00 Bahia NE Multiplier 20.19 77.78 Ceará NE Commercial 0.00 0.00 Federal CO Multiplier 0.48 3.51 District Espirito SE Commercial 0.83 100.00 Santo Goiás CO Multiplier 5.25 46.81 Minas Gerais SE Commercial 5.47 100.00 Mato Grosso CO Multiplier 8.25 39.76 do Sul Mato Grosso CO Multiplier 0.74 15.79 Pará N Commercial 0.00 0.00 Pernumbuco NE Commercial 0.00 0.00 Piauí NE Multiplier 32.79 9.13 Paraná S Multiplier 36.01 44.95 Rio de SE Commercial 6.82 100.00 Janeiro Rondônia N Commercial 8.06 100.00 Roraima N Commercial 0.00 0.00 Rio Grande S Multiplier 81.74 92.70 do Sul Santa S Multiplier 12.83 70.37 Catarina Sergipe NE Commercial 0.00 0.00 São Paulo SE Multiplier 12.29 37.97 Tocantins N Multiplier 16.36 60.27 1 Brazilian States; N, North; NE, Northeast; S, South; SE, southeast; CO, Midwest. percent (7/11) did not use their own sires for reproduction, while 36.36 percent (4/11) using both purchased and own males for reproduction. Rio Grande do Sul state (RS) had the highest percentage of sires born in the State, which is not surprising as this is the home state for this breed. Even when subpopulations are studied, genetic parameters (F, AR, AF) vary little from the mean (Table 7), showing Table 7. Genetic variation by subpopulation (sires and grandsires with largest number of offspring, farms and municipalities) for the Brazilian Crioula horse. Grandsire Sire Farms Municipalities Number 196 803 106 49 Offspring 53 632 163 817 935.68 4242.22 F 0.010 0.011 0.013 0.011 AR 0.018 0.018 0.016 0.015 Gen Max 6.043 6.952 6.878 7.308 Gen Com 1.552 1.852 1.915 2.085 Gen Eq 2.737 3.196 3.236 3.509 AF 0.005 0.005 0.005 0.004 GCI 4.722 6.014 2.818 1.793 Genetic conservation index (GCI), average relatedness (AR), mean individual inbreeding (F) and average individual inbreeding (AF), GenMax, maximum number of generations; Gen_Com, common number of generations; Gen_Eq, equivalent number of generations.

Population structure of the Brazilian Crioula Horse 7 Table 8. Population data depending on recently born (from 1998) or family of sires with largest offspring numbers (family) in Crioula horses from Brazil. All Recent Family No. of reference animals 333 715 184 646 184 005 No. of ancestors 5086 4230 3334 Effective number of founders 91 62 49 Effective number of ancestors 75 (1.21) 45 (1.38) 35 (1.40) No. of ancestors explaining 50% 56 41 31 Base population (one or more unknown parents) 7901 Actual Base Population (one unknown parent = half founder) 6891.5 Effective Population Size of Founders 95.19 Expected F by unbalancing of founders contribution 0.53% Computed mean F 1.21% Mean Average Relatedness 1.58% Age first becoming parent 10.21 6.92 10.19 Regression coefficient of age becoming parent on inbreeding coefficient 13.18 ± 8.89 1.11 ± 0.51 21.35 ± 1.00 Regression coefficient of age becoming parent on increase in inbreeding coefficient 15.75 ± 16.71 0.32 ± 1.22 34.77 ± 2.25 Ne from regression birth date 274.48 215.59 Ne from log regression birth date 270.80 230.19 Ne via individual increase in inbreeding 98.94 ± 20.01 97.23 ± 19.83 Ne from Log regression on equivalent generations 68.88 58.06 Ne via regression on equivalent generations 70.14 58.90 Ne, effective population size. Figure 8. Percentage of animals registered per year depending on inbreeding level on all (ALL) animals registered, those from the major families (FAMILY) and the present generation (RECENT)Levels of inbreeding (0.00 (0.0), 0.001 0.05 (0.05), 0.051 0.10 (0.10), 0.101 0.15 (0.15), 0.151 0.20 (0.20), 0.201 0.25 (0.25) and greater than 0.25 (0.30). Figure 9. Mean inbreeding (F) of Crioula Horse depending on the date the farm started to register animals, the period of time the farm registered animals and the number of animals registered per farm. that sires are imported into farms as a general rule. GCI was significantly higher than herd average for those stallions that left more offspring, as expected. More recently animals are becoming parents at a relatively younger age. This may be a reflection of the rapid expansion of the breed and the need for breeding females to reproduce (Table 8). Effective population sizes were similar for different subgroups. This is a reflection of the fact that the major sire families produce a large proportion of the recently born animals, reflected by the large number of animals in common in both groups. It can be seen that the more recently farms started to register animals the higher mean F (Figure 8, Figure 9). This is also reflected in the longer the farm has been registering animals and the higher the number of animals registered the lower the mean F.

8 F.C. Maciel et al. Discussion Due to their high commercial value, the horses, especially those used for sports, need have their pedigree known, both for registration and purchase. Confirmation of genealogy is extremely important, not only for ensuring the ascendancy of the animals, but also because a pedigree can reliably enable the buyer to identify the origin of genetic problems in the herd and reduce or eliminate them (Coelho and Oliveira, 2008). The study of genetic variability and population structure in some horse breeds using pedigree analysis, alone or in combination with information such as genetic markers, has grown in recent years (Chiofalo et al., 2003; Valera et al., 2005). The results on the distribution of this breed in Brazil were in accordance with expected, with a high concentration of animals in the southern region, since the breed was developed in Southern Latin America, a region known as South American Pampa (Argentina, Southern Brazil and Uruguay). The increase in the use of the Crioula Horse in southeastern states such as Sao Paulo, as well as in Northern and Northeastern Brazil is also seen. This reflects the expansion and distribution of the breed, where the largest breeder (Estância Nazareth) has only 1.2 percent of the animals, unlike other local breeds, such as the Pantaneiro Horse, where the largest breeder has 7.25 percent of animals recorded. The amount of information on the Crioula has increased with the number of generations. These data are consistent with those found in studies with other breeds where the most distant generations have less genealogical information, as this may have been lost over the years or were not recorded (Valera et al., 2005). Vicente, Carolino and Gama (2009), as here, found an increase in complete Lusitano pedigrees over ten generations. The increase in the number of records per year is evident (Figure 2), mainly due to two factors: (a) partnership between the Breeder s Association and the Rural TV Channel, where the most important competition involving the Crioula, Freio de Ouro (which involves several tests, such as breed standard, gait, barrel racing, separating cattle, lassoing, horse control and obedience), is televised across the country and worldwide over the internet. This channel also televises most of the auctions with these horses, so farmers from distant regions can acquire animals without leaving home; (b) projects developed by the ABCCC, such as Vaquejada Project (A northeastern tradition where horses are used to drive cattle in a confined space.), created to insert Crioula horse in this activity in the northeast. The number of registers in the last two years (2010 and 2011) should be disregarded as the animals are subjected to technical inspection until they are nine months old and most of the animals born in this period had not passed inspection at the time of data collection. The greater number of females compared to males can be explained by fact that some breeders do not register males which do not meet their specifications. In domestic animals some individuals assume importance in the origin of the breed and its development (Fletcher, 1945, 1946; Rhoad and Kleberg, 1946; Gazder, 1954) as evidenced in this population. This is due to several factors, such as propaganda, success in auction rings and competitions as well as functional and morphological traits. The concentration of births in the months from September to January is mainly due to the occurrence of marked photoperiod in southern Brazil (Winter, 2007). Most mares show anestrus and transitional phases in summer and autumn and, under natural conditions, the transition ends near or after the equinox, in September and October. The mean coefficient of inbreeding found in Crioula was lower compared to other local breeds. Schurink, Arts and Ducro (2012) found average inbreeding coefficient of 0.053 while the Crioula horse was found 0.88 percent for the total population (including imported) and 1.21 percent for those born in Brazil. Wolc and Balińska (2010) also found much higher values (5.9, 5.1 and 5.6 percent) on three farms breed Polish Konik. Abrahão et al. (2002) found similar levels of inbreeding (0.8 percent) to those found in Crioula when studying inbreeding in Thoroughbred mares reared in Brazil. Low levels of inbreeding found in Crioula in Brazil can be explained by the large population base, large number of founders and ancestors, and the introduction of animals from other countries (Argentina, Uruguay and Chile), introducing different racial strains. The number of inbred animals found in the full third generation (80.99 percent of the population) was similar to that found by Mota et al. (2006) found that 88.0 percent of inbred animals in Mangalarga horses in the period 1936 2003. The results in Crioula show a clear increase in the number of inbred animals, reaching 100 percent at the 15th generation, although there are only two animals in this generation. Inbreeding of 10 percent was estimated for ethnic breeds such as Spanish Purebred (Valera et al., 2005) and Lipizzano (Zechner et al., 2002), and after the closing of the studbook this parameter has aggravated (Vicente et al., 2009). For Mangalarga inbred animals, the mean coefficient of inbreeding was 5.7 percent, with a maximum of 46.9 percent (Costa et al., 2005), similar to those found by Procópio et al. (2003) in the Campolina. The increase of inbreeding (Boichard, Maignel and Verrier, 1997) can be used to derive the effective size of a population. However, this method reflects mainly the long-term effects of selection and, furthermore, is very sensitive to incomplete information from a pedigree. An alternative would be additional parameters based on the probability of gene origin, and the effective number of founders and effective number of remaining genomes are commonly used in wild populations, but less common in studies of domestic animals. However, there are several studies in domestic species and different breeds, such as in cattle (Faria et al., 2002; Vercesi Filho et al., 2002; Pereira et al., 2005; Vozzi et al., 2006, 2007; Hammami

Population structure of the Brazilian Crioula Horse 9 et al., 2007; Martínez, García, and Gallego, 2008), in horses (Costa et al., 2005), sheep (Goyache et al., 2003), pigs (Toro et al., 2000) and even donkeys (Gutiérrez et al., 2005), based on some sort of pedigree analysis. Ideal levels of inbreeding range from 0.05 (Nicholas, 1989) to 0.01 (FAO, 2000) per generation, and the Crioula are below these levels. Moureaux et al. (1996) reported that when studying inbreeding in horse breeds, two groups can be distinguished: a group of international breeds, with values ranging from 0.81 to 2.89 percent, and another group with small population sizes with values ranging from 2.25 to 14.7 percent. The high variation in mean individual levels of inbreeding and inbreeding rate seen here between 1940 and 1970 may have been a reflection of the lack of breeding animals, the difficulty in finding stallions of the breed as well as in moving between farms, since the increase in effective herd size came after 1970 (Figure 3). The mean coefficient of relatedness (kinship) increased from the start of breed registration until about the year 2003. These results can be explained by selection of animals which showed excellent results in morphological and functional competitions, and therefore many breeders sought to acquire its offspring. Mean coefficient of relatedness provides additional information to explain relations between relatives (Gutiérrez, Goyache and Cervantes, 2009). In conservation genetics, knowledge of relatedness is required to optimize conservation strategies. Coancestry relationship is expressed relative to the base population in which all alleles are not defined as being identical by descent, so that the population coancestry base is zero by definition (Falconer and Mackay, 1996; and Lynch Walsh, 1998). Kinship average here was low compared to other breeds of horse (Zechner et al., 2002; Valera et al., 2005). Valera et al. (2005) found average values of inbreeding and AR of the Andalusian horse population were, respectively 8.48 and 12.25 percent, while Cervantes et al. (2008) found inbreeding average of 7 percent to Spanish Arabian. The average individual inbreeding and variation in inbreeding increased from the mid-1930s to mid-1970s, and remained relatively constant thereafter. This can be explained by an intense exchange of animals between Brazil, Argentina, Uruguay, and especially Chile, where different strains were incorporated into the local population after the 1970s. When comparing inbreeding in the three populations studied (ALL, FAMILY and RECENT), a clear increase in the percentage of animals with relatively low levels of inbreeding (0.05). The percentage of animals with F > 0.05 has not grown since the late 1990s and although an increase in inbreeding (F > 0.1) was seen in the ALL population in the 1940s to 1980, this has largely disappeared. Higher levels of inbreeding are not seen in RECENT or FAMILY. The percentage of animals with no inbreeding has steadily decreased in all populations and should increase with the closure of the herd book. In Andalusian horses, Valera et al. (2005) found 39.6 effective founders and 27 ancestors, with only six animals responsible for 50 percent of genetic variability in the breed. The situation of Crioula is better, with nearly 95 founders, but this is lower than for Pantaneira (McManus et al. 2013), perhaps due to rapid expansion of the breed in recent years. The small number of founder animals highlights the need for monitoring of inbreeding of the herd studied, control of breeding and introduction of animals that have no direct relationship with the principal ancestors identified in this study. Alderson (1991) used the GCI to calculate an effective number of founders in the pedigree of an animal. The higher the value GCI better it is for use in conservation of a breed. The GCI can be used both by individual breeders as an aid in selecting breeding stock, or within a breed to formulate a comprehensive improvement program. However, the index has limitations, as it is not representative of any concentration of breeding animals for non-founders in subsequent generations and cannot be used without pedigree records (Alderson, 1991). The interval between generations found in Crioula was higher when compared with other breeds such as Costa (2002) and Gonçalves (2010), who evaluated animals Mangalarga Marchador found generation interval of 8.9 and 8.98 years, respectively. This may be due to the prohibition of use of reproductive biotechnologies until 2011, as well as the large number of morphological and functional competitions in which these horses compete, means that these horses begin reproductive life at a later age. As natural breeding is still used, many mares exposed to the stallion do not impregnate. Vicente, Carolino and Gama (2009) also found a high generation interval for the Lusitano breed (10.4 years: males 11.2 and females 9.6 years), also possibly due to use after the end of their sporting career. McManus (2013) found lower intervals between generations in the Pantaneiro horse and considered this was because this breed is used on-farm animals instead of sports and therefore enter reproduction earlier. The use of artificial insemination and embryo transfer are beginning to become more widely used in horse populations (2011 in Crioula) but less frequently than in other farm animals (Laat, 2001). The population aspects of horses, especially those related to inbreeding and effective size therefore have an individual character (Costa et al., 2005) and continued monitoring is necessary to avoid future problems. Vinocur et al. (2003) determined the allele frequencies of seven blood group systems and eight protein systems in six herds of Crioula horses raised in Rio Grande do Sul of State. The herds presented a significant component due to isolation (F ST =0.0866, p < 0.01). They found high values for average heterozygosity (0.4631). When all herds were considered in the analysis, the inbreeding level (F IS ) was zero, in line with seen here where values near zero indicate that levels of genetic variability of the flock

10 F.C. Maciel et al. are high, since these values indicate an excess of heterozygous animals, both within and paternal lineages (subpopulations) and for the population as a whole (metapopulation). For the F ST index, the value calculated was 0.000, demonstrating that there is no differentiation between the parental strains with the formation of subpopulations. The Crioula is not at risk status (FAO 1998; Database http://dad.fao.org DAD-IS) as opposed to other breeds of naturalized horses (Luís et al., 2005). The integration of herdbooks from other countries with Crioula horses used for crossing with the Brazilian Crioula (such as Uruguay, Argentina and Chile) could help to identify some genetic bottlenecks that may exist but were not seen here. The fact that the smaller, younger farms have higher levels of inbreeding may reflect a policy of the larger farmers to import stallions or sell off inbred animals to those starting up in the business. It may also reflect a lack of breeding policy by the smaller farmers who may have a single stallion. This is also from the case seen with some pig breeds in the USA (Welsh et al., 2010) but to a lesser extent. Conclusion The indices obtained from the genetic data of genealogical record of the Crioula Breed show that in general inbreeding is under control in the total population and the effective population size is not at a critical level. Recent decrease in generation intervals may be a reflection of the rapid expansion of the breed and the need for mares for reproduction. Acknowledgements Thanks are due to the Breeder s Association for the Crioula Horse, INCT Pecuária (CNPq, FAPEMIG, MCT) and CAPES for scholarships. Statement of interest No conflicting interests were identified. References Abrahão, A.R., Mota, M.D.S., Oliveira, H.N. & Madureira, A.P. 2002. Endogamia em éguas da raça Puro-Sangue Inglês. In Proceedings IV Simpósio da Sociedade Brasileira de Melhoramento Animal, Campo Grande, Brazil, pp. 224 225. Alderson, G.L.H. 1991. A system to maximize the maintenance of genetic variability in small populations. In L. Alderson, ed. Genetic conservation of domestic livestock, pp. 18 29. Wallingford, UK, CAB International. Ballou, J.D. & Lacy, R.C. 1995. Identifying genetically important individuals for management of genetic diversity in pedigreed populations. In: Ballou, J.D., Gilpin, M. & Foose, T.J. (Eds) Population management for survival & recovery. Analytical methods and strategies in small population conservation. New York: Columbia University Press, p.76 111. Boichard, D., Maignel, L. & Verrier, É. 1997. The value of using probabilities of gene origin to measure genetic variability in a population. Genet. Sel. Evol., 29: 5 23. Caballero, A. & Toro, M.A. 2000. Interrelations between effective population size and other pedigree tools for the management of conserved populations. Genet. Res., 75: 331 343. Caballero, A. & Toro, M.A. 2002. Analysis of genetic diversity for the management of conserved subdivided populations. Conserv. Genet., 3: 289 292. Cervantes, I., Molina, A., Goyache, F., Gutiérrez, J. P. & Valera, M. 2008. Population history and genetic variability in the Spanish Arab Horse assessed via pedigree analysis. Livestock Sci., 113: 24 33. Chiofalo, L., Portolano, B., Liottal, L., Rundo Soteral, A. & Finocchiaro, R. 2003. Demographic characterization, inbreeding and genetic variability within Sanfratellano population horse from genealogical data. Ital. J. Anim. Sci., 2(Suppl. 1): 592 594. Coelho, E.G.A. & Oliveira, D.A.A. 2008. Testes genéticos na eqüideocultura. Revista Brasileira de Zootecnia, 37: 202 205. Special Number. Costa, M.D. 2002. Caracterização demográfica e estrutura genética da raça Mangalarga Marchador. Universidade Federal de Minas Gerais, Belo Horizonte, Brazil (Thesis). Costa, M.D., Bergmann, J.A.G., Resende, A.S.C. & Fonseca, C.G. 2005. Análise temporal da endogamia e do tamanho efetivo da população de equinos da raça Mangalarga Marchador. Arquivo Brasileiro de Medicina Veterinária e Zootecnia, 57: 112 119. Cothran, E.G., Santos, S.A., Mazza, M.C.M., Lear, T.L. & Sereno, J. R.B. 1998. Genetics of the Pantaneiro horse of the Pantanal region of Brazil. Genet. Mol. Biol. (available at http://dx.doi.org/10.1590/ S1415-47571998000300009). Escola Superior de Agricultura Luiz de Queiroz ESALQ. Centro de Estudos Avançados em Economia Aplicada da ESALQ. 2006. Estudo do Complexo do Agronegócio do Cavalo no Brasil. Brasília, CAN, MAPA. Falconer, D.S. & Mackay, T.F.C. 1996. Introduction to Quantitative Genetics. Harlow, Longman. Faria, F.J.C., Vercesi Filho, A.E., Madalena, F.E. & Josahkian, L.A. 2002. Estrutura populacional da raça Nelore Mocho. Arquivos Brasileiro de Medicina Veterinária e Zootecnia, 54: 501 509. Fletcher, J.L. 1945. A genetic analysis of the American Quarter Horse. J. Hered., 36: 346 352. Fletcher, J.L. 1946. A study of the first fifty years of Tennessee Walking horse breeding. J. Hered., 37: 369 373. Gazder, P.J. 1954. The genetic history of the Arabian Horse. J. Hered., 45: 95 98. Gonçalves, R.W. 2010. Efeito da endogamia nas características morfométricas e reprodutivas dos eqüinos da raça Mangalarga Marchador. Universidade Estadual de Montes Claros, Montes Claros, Brazil. (Dissertation) (available at http://www.dominiopublico.gov.br/pesquisa/detalheobraform.do?select_action=&co_obra= 193645). Goyache, F., Gutiérrez, J.P., Fernández, I., Gomez, E., Alvarez, I., Díez, J. & Royo, L.J. 2003. Using pedigree information to monitor genetic variability of endangered populations: the Xalda sheep breed of Asturias as an example. J. Anim. Breed. Genet., 120: 95 105. Gutiérrez, J.P., Marmi, J., Goyache, F. & Jordana, J. 2005. Pedigree information reveals moderate to high levels of inbreeding and a

Population structure of the Brazilian Crioula Horse 11 weak population structure in the endangered Catalonian donkey breed. J. Anim. Breed. Genet., 122: 378 386. Gutiérrez, J.P., Goyache, F. & Cervantes, I. 2009. Useŕs Guide, ENDOG v4.6, A Computer Program for Monitoring Genetic Variability of Populations Using Pedigree Information, (available at http://www. ucm.es/info/prodanim/html/jp_web_archivos/en_us_g_.pdf). Hammami, H., Croquet, C., Stoll, J., Rekik, B. & Gengler, N. 2007. Genetic diversity and joint-pedigree analysis of two importing Holstein populations. J. Dairy Sci., 90: 3530 3541. James, J.W. 1977. A note on selection differentials and generation length when generations overlap. Anim. Prod., 24: 109 112. Kelly, L., Postiglioni, A., De Andrés, D.F., Vega-Plá, J.L., Gagliardi, R., Biagetti, R. & Franco, J. 2002. Genetic characterisation of the Uruguayan Crioula horse and analysis of relationships among horse breeds. Res. Vet. Sci., 72(1): 69 73. Laat, D.M. 2001. Contribuição genética de fundadores e ancestrais na raça Campolina. Universidade Federal de Minas Gerais, Belo Horizonte, Brazil. (Dissertation) (available at http://www.ebookcult. com.br/produto/contribuicao_genetica_de_fundadores_e_ancestrais_ na_raca_campolina-10151). Lacy, R.C. 1989. Analysis of founder representation in pedigrees: founder equivalents and founder genome equivalents. Zoo Biol., 8: 111 123. Luís, C., Cothran, E.G., Oom, M.M. & Bailey, E. 2005. Major histocompatibility complex locus DRA polymorphism in the endangered Sorraia horse and related breeds. J. Anim. Breed. Genet., 122: 69 72. Lynch, M. & Walsh, J.B. 1998. Genetics and analysis of quantitative traits. Sunderland: Sinauer Assocs. Inc., 980p. Maccluer, J.W., Boyce, A.J., Dyke, B., Weitkamp, L.R., Pfenning, D.W. & Parsons, C.J. 1983. Inbreeding and pedigree structure in Standardbred. J. Hered., 74: 394 399. Malécot, G. 1948. Les Mathématiques de l Hérédité. Masson et Cie, Paris Martínez, R.A., García, D., Gallego, J.L. 2008. Genetic variability in Colombian Crioula cattle populations estimated by pedigree information. J. Anim. Sci., 86: 545 552. McManus, C. 2013. Evaluation of Conservation Program for the Pantaneiro Horse in Brazil, Revista Brasileira de Zootecnia, in press. McManus, C., Santos, S.A., Dallago, B.S.L., Paiva, S.R., Saraiva, R. F., Braccini Neto, J., Marques, P.R. & Abreu, U.G. 2013. Evaluation of conservation program for the Pantaneiro horse in Brazil. Rev. Bras. Zoot. 42:404 413. Mirol, P.M., Peral García, P., Vega-Pla, J.L. & Dulout, F.N. 2002. Phylogenetic relationships of Argentinean Crioula horses and other South American and Spanish breeds inferred from mitochondrial DNA sequences. Anim. Genet., 33: 356 363. Mota, M.D.S., Prado, R.S.A. & Sobreiro, J. 2006. Characterization of the Mangalarga horse population in Brazil. Arch. Zoot., 55:31 37. Moureaux, S., Verrier, É., Ricard, A. & Mériaux, J.C. 1996. Genetic variability within French race and riding horse breeds from genealogical data and blood marker polymorphisms. Genet. Sel. Evol. 28:83 102. Nicholas, F.W. 1989. Incorporation of new reproductive technology in genetic improvement programs. In: Hill, W.G., Mackay, T.F.C. (Eds) Evolution and animal breeding. Wallingford: CAB International, p.201 209. Pereira, M.C., Mercadante, M.E.Z., Albuquerque, L.G. & Razook, A. G. 2005. Estimativa de ganho genético a partir de diferenciais de seleção e parâmetros populacionais em um rebanho Caracu. Revista Brasileira de Zootecnia, 34: 2245 2252. Procópio, A.M., Bergmann, J.A.G., Costa, M.D. 2003. Formação e demografia da raça Campolina. Arq. Bras. Med.Vet. Zoot. 55:361 365. Rhoad, A.O. & Kleberg, J.R. 1946. The developments of a superior family in the modern Quarter Horse. J. Hered., 37: 227 238. Rodero, A., Delgado, J.V. & Rodero, E. 1992. Primitive Andalusian livestock and their implication in the Discovery of America. Arch. Zootec., 41: 383 400. Schurink, A., Arts, D.J.G. & Ducro, B.J. 2012. Genetic diversity in the Dutch harness horse population using pedigree analysis. Livestock Sci., 143: 270 277. Toro, M.A., Rodrigañez, J., Silio, L. & Rodriguez, C. 2000. Genealogical analysis of a closed herd of Black Hairless Iberian pigs. Conserv. Biol., 14: 1843 1851. Valera, M., Molina, A., Gutiérrez, J.P., Gómez, J. & Goyache, F. 2005. Pedigree analysis in the Andalusian horse: population structure, genetic variability and influence of the Carthusian strain. Livestock Prod. Sci., 95: 57 66. Vercesi Filho, A.E., Faria, F.J.C., Madalena, F.E. & Josahkian, L.A. 2002. Estrutura populacional do rebanho Tabapuã registrado no Brasil. Arquivos Brasileiro de Medicina Veterinária e Zootecnia, 54: 609 617. Vicente, A., Carolino, N. & Gama, L.T. 2009. Indicadores Demográficos no Cavalo Lusitano. Arch. Zootec., 58(Supl. 1): 501 504. Vinocur, M.E., Brass, K.E., Rubin, M.I.B. & Silva, C.A.M. 2003. Genetic variability in the Brazilian Crioula horse breed. Cienc. Rural, 33: 137 142. Vozzi, P.A., Marcondes, C.R., Magnabosco, C.U., Bezerra, L.A.F. & Lôbo, R.B. 2006. Structure and genetic variability in Nellore (Bos indicus) cattle by pedigree analysis. Genet. Mol. Biol., 29: 482 485. Vozzi, P.A., Marcondes, C.R., Bezerra, L.A.F. & Lôbo, R.B. 2007. Pedigree analyses in the Breeding Program for Nellore Cattle. Genet. Mol. Res., 6: 1044 1050. Welsh, C.S., Stewart, T.S. & Schwab, C. 2010. Pedigree analysis for 5 swine breeds in the United States and the implications for genetic conservation. J. Anim. Sci., 88:1610 1618. Winter, G.H.Z. 2007. Características reprodutivas sazonais da égua Crioula em um propriedade latitude 29 38 S no Rio Grande do Sul. Universidade Federal de Santa Maria, Brazil. (Dissertation) (available at http://www.openthesis.org/documents/reprodutivas-sazonais-dacrioula-em-473051.html). Wolc, A., Balińska, K. 2010. Inbreeding effects on exterior traits in Polish konik horses. Arch. Anim. Breed., 53: 1 8. Zechner, P., Sölkner, 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. Livestock Prod. Sci., 77: 137 146.