Genetic Variability within the Murgese Horse Breed Inferred from Genealogical Data and Morphometric Measurements

Abstract

The Murgese horse (MH) is a native breed from Apulia (Italy). This study aimed to evaluate the population status with regard to the available pedigree information (6923 animals born between 1900 and 2020), as well as its demographic and morphological evolution. The mean equivalent generations were 5.88. The average relatedness, inbreeding coefficient and increase in inbreeding by equivalent generation (ΔF) were 9.88%, 5.22%, 1.05%, respectively. The effective population size based on ΔF was 47.46. The effective number of founders (fe) was 36, and that of ancestors (fa) was 19. The ratio fe/fa was 1.89 witnessing a bottleneck effect. The ratio fa/fe indicates a 52% reduction of the genetic diversity as expected, given the fact that, for the current population (now recovered to 5000 breeding animals), the fe is 34 and the fa only 17, with 50% of diversity being explained by only six ancestors. Basically, the results reflect a substantial loss of genetic diversity in the MH breed over generations since its official founding, and unbalanced use of sires in the population, highlighting the importance of continuous monitoring and implementation of more effective conservation measures, especially in view of the growing request for boosting genetic improvement for MH morpho-functional traits.

1. Introduction

The Murgese belongs to the group of baroque-type horse breeds and is additionally characterized by a black or blue roan coat (Figure 1). Originally being bred as a working horse, suitable for both military and civilian use, a shift in breeding objectives in the 1990s resulted in an emphasis on performing traits [1,2]. The Murgese is appreciated today in equestrian tourism, with increasing interest also in horse riding sports, mainly dressage. The history of the Murgese horse (MH) is strictly connected with that of the arid and rocky hills of Apulia, Le Murge, from which it takes its name. The breed has a rustic nature [3], necessary for survival in such a difficult environment characterized by harsh climate (cold in the winter, hot and dry in the summer), poor pastures and the presence of enzootic pathogens [4,5].

At the dawn of the twentieth century, the horse in Italy was reduced to a mere source of food and, even worse, almost all the national equestrian resources were lost in World War I. Due to the geographical isolation and the production of very valuable mules derived by crossing Murgese mares with Martina Franca jacks, the Apulian horse was not involved in cross-breeding programs which affected most of the native Italian breeds.

During the Fascist regime, the substantial failure of the hippo-technical measures taken by previous governments came to light and a package of legislative actions to encourage equine breeding was set out. In 1925, the Murgese, a mesomorphic horse with wide transverse diameters, was officially distinct from the rest of the equine population present in Apulia. However, in 1926, when the official denomination of the breed, as well as the first individuals’ registration in the Murgese Stud Book, were established, the MH population counted only 46 founder brood-mares and 9 stallions [3]. Since the 1970s, the revival of equestrian leisure culture has led to a long-lasting interest in the MH, whose population has progressively increased. In 1990, the Italian Ministry of Agriculture and Forestry established the “Anagraphic Register of the Equine Populations identifiable as Ethnic Groups” [1], among which the MH was included. In 2008, the Murgese Stud-book was established, and it was entrusted to the Italian breeders association (Associazione Italiana Allevatori, A.I.A.). In 2019, it was transferred from A.I.A. to the Italian national association of Haflinger breeders (Associazione Nazionale Allevatori Cavallo Razza Haflinger Italia, A.N.A.C.R.HA.I.) and, in 2022, from A.N.A.C.R.HA.I. to the “National association of the Murgese horse and Martina Franca ass breeders” (A.N.A.M.F.). Over time a number of additional bodies have been involved in the MH conservation policy, among which the regional “Service for the Valorization of Natural Resources and Biodiversity” (evolved from the Apulia regional “Institute for the Improvement of Horse Populations). The MH population has now recovered to 6437 breeding animals [6].

A precondition for decisions in livestock selection and breeding programs is the robust knowledge of breed attributes, including demography and genetic variability [7]. Particularly pedigree analysis enables the estimation of genetic variability and its evolution over time using information from genealogical records [8]. Moreover, demographic analyses may help to understand the factors that have affected the genetic history of a population [9]. In the last years, a number of studies have been carried out on the population structure and genetic variability of horse breeds by analyzing pedigree data [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]. Classically, monitoring of genetic diversity via the use of genealogical information has been carried out by assessing inbreeding (F) and relationships (R) in the population of interest, converted into effective population size (N_e), which is regarded as a good indicator to estimate genetic diversity for conservation purposes.

The main objective of this study was, therefore, a complete evaluation of the demographic and genetic status of the MH breed, combining pedigree information with data about the temporal evolution of basic descriptive morphometric traits. The knowledge generated by our research will be beneficial for orienting future genetic management and breeding strategies of the MH.

2. Material and Methods

2.1. Genealogical Data

Genealogical data were obtained from the “National Association of Murgese horse and Martina Franca ass breeders” (A.N.A.M.F.). The pedigree dataset includes animals born between 1900 and 2020. The above set will be hereinafter referred to as “total population”. In addition, we defined a reduced sub-set, hereinafter referred to as “reference population”, including only those animals born from 2005 onward, assuming that they may represent the current population. For each animal in the dataset, individual ID, given name, year of birth and sex were available.

2.2. Pedigree Analysis

All demographic and genetic parameters were evaluated by using the ENDOG (v4.6) software [27]. The pedigree completeness level was characterized by computing: (i) the percentage of ancestors known at various generations, (ii) the number of traced generations (number of generations separating the offspring from its furthest known ancestor in each path; ancestors with no known parent were considered as founders, i.e., generation 0) and (iii) the number of equivalent generations (the sum over all known ancestors of (1/2)ⁿ where n is the number of generations separating the individual to each known ancestor [8]. Generation intervals (the average age of parents at the birth of offspring kept for reproduction) and ages at first foaling (the average age of the sire or dam at birth of first registered offspring) were calculated, from birth dates of single animals together with those of their fathers and mothers, for each of the 4 gametic pathways: sire to son, sire to daughter, dam to son and dam to daughter. To characterise the genetic variability of the population we analysed the following parameters: (i) the individual inbreeding coefficient (F), a measure of the probability that, in an individual, the two alleles at any given locus are identical by descent (i.e., descendants from a single ancestor) [28]; (ii) the average relatedness coefficient (AR) of each individual, defined as the probability of an allele, randomly chosen from the whole population in the pedigree, to belong to a given animal [29,30]; (iii) the effective number of founders (fe), defined as the number of equally contributing founders that would be expected to produce the same genetic diversity as in the population under study [31]; the effective number of ancestors (fa), i.e., the minimum number of ancestors (founders or not), explaining the complete genetic diversity of a population [32]; (iv) the ratio between the effective number of founders and the effective number of ancestors (fe/fa) as an indirect indicator of pedigree consistency (fe, but not at the same extent, fa may be overestimated when pedigree information is missing) and unbalanced use of animals in reproduction; (v) the ratio between the effective number of ancestors and the effective number of founders (fa/fe). Finally, the effective population size (N_e) was computed using the following three approaches: via an individual increase in inbreeding [33], via regression and via Log regression, on the birth day. The resulting N_e have been compared to the value obtained following the predictive formulae for the effective size of X-linked genes originally given by Wright [34]:

$$N_e=4NmNf/(Nm+Nf)$$

While the increase in inbreeding by maximum, complete and equivalent generations was obtained using the ENDOG package for the total population, the increase in inbreeding by equivalent generation in the reference population was inferred using the equation ΔF = 1/(2N_e) [35].

2.3. Morphometric Data

Three morphometric traits were available in the dataset obtained from the “National Association of Murgese horse and Martina Franca ass breeders” (A.N.A.M.F.), namely withers height (WH), chest girth (CG) and cannon bone circumference (CBC). In the frame of the yearly exhibition held in autumn in Martina Franca (Apulia, Italy), linear measurements were taken from about 30-month-old horses. Following a standard protocol, measurements were taken as follows: the smallest CBC and CG were measured by using plastic measuring tape; the WH was taken from the ground to the highest point of the withers. The dataset used in this study contains a sub-set of animals, out of those in the pedigree dataset, including 3282 animals (885 stallions and 2397 mares) registered in the official stud-book and approved as breeding animals in the period from 1987 to 2017 for which morphometric records were available on the three parameters. Summary statistics (average, standard deviation) were estimated from the data in order to assess the phenotypic variability of the Murgese horse population for the considered morphometric parameters. We also used the morphometric data to estimate the evolutionary change in each of the three considered traits by computing the change in the mean (Z) of WH, CG and CBC in both sexes over five subsequent periods from 1987 to 2017.

3. Results

3.1. Demographic and Genetic Breed Status and Trends Inferred from Pedigree Data

The trends in the population size of the stud-book registered Murgese horses over the last century are reported in Figure 2.

A progressive increase in the number of registered animals can be observed for the total population. Males and females show similar trend lines even though the female slope was steeper than the male one. Pedigree data were analyzed in order to derive an overview of the demographic and genetic breed status and trends. A description of the pedigree structure of the Murgese horse is provided in

Table 1. Pedigree structure of the Murgese horse.

Parameter	Total Population	Reference Population
Total number of animals	6923	4465
Males	2124	1714
Females	4799	2751
Sires	551	208
Dams	2840	977
Individuals with both known parents	6574	4459
Individuals with both unknown parents	261	0
Individuals with one unknown parent	21	6 *
Individuals with unknown sire	282	6 *
Individuals with unknown dam	261	0
Females with both unknown parents	208	0
Males with both unknown parents	53	0
Effective Number of Founders	36	34
Number of Ancestors	313	297
Effective Number of Ancestors	19	17
fa/fe	0.52	0.50
fe/fa	1.89	2.00
Nº of ancestors explaining 50%:	7	6
Individuals with no progeny	3524	3280
Matings between full sibs	2	0
Matings between half sibs	120	13
Matings parent-offspring	28	5

* These 6 individuals with one unknown parent in the reference population are females whose genealogical records were lost during the transition between the “Anagraphic Register of the Equine Populations identifiable as Ethnic Groups” and the establishment of the Stud-Book.

. The total registered population included 6923 horses, out of which 2124 were males and 4799 were females. In the reference population (animals born in the last 15 years), these figures were 4465, 1714 and 2751, respectively (with a theoretical male:female sex ratio roughly corresponding to 1:1.6). However, given that some animals older than 15 years are still alive, the number of animals in the alive population, namely around 5000 [6], may be better considered as representative of the total census size (N_c) of the MH breed. Out of the number of animals in the total and the reference population, 551 and 208 were the reproducing males and 2840 and 977 the reproducing females, respectively.

As shown in Table 1, in the total population, 95% of animals (6574) had both parents known, while about 3.7% (261) had both parents unknown (founders). These figures were, respectively, 99.8% and 0.0% in the reference population, pointing, as expected, to the improvement of the pedigree quality for the most recent animals. Indeed, as presented in

Table 2. Percentages of pedigree completeness in the total and the reference populations at various generations.

Generation *	Total Population	Reference Population
1	0.956	0.999
2	0.925	0.994
3	0.894	0.989
4	0.853	0.983
5	0.778	0.957
6	0.638	0.845
7	0.442	0.622

* For compactness, percentages of pedigree completeness are shown here only for the first seven generations, with generation 1 referring to the most recent and generation 7 referring to the most remote one.

, at the seventh generation, the pedigree completeness had already fallen below 50% for the total population and was close to 60% for the reference population.

More in detail, the level of pedigree completeness for the Murgese horse breed is described in

Table 3. Pedigree completeness for the total and the reference populations.

Parameter	Total Population (SD)	Reference Population (SD)
Mean Maximum Generations	10.09 ± 3.20	11.82 ± 1.19
Mean Complete Generations	3.79 ± 1.50	4.52 ± 0.95
Mean Equivalent Generations	5.88 ± 1.98	6.98 ± 0.75

. In the total population, the Mean Maximum Generations were equal to 10.09 ± 3.20, while the Mean Complete Generations and Mean Equivalent Generations were equal to 3.79 ± 1.50 and 5.88 ± 1.98, respectively. In the reference population, the Mean Maximum Generations were equal to 11.82 ± 1.19, while the Mean Complete Generations and Mean Equivalent Generations were equal to 4.52 ± 0.95 and 6.98 ± 0.75, respectively.

The number of founders (both parents unknown) in the total population was 261, while the effective number of founders dropped to 36 and 34 for the total population and reference population, respectively (Table 1). The number of ancestors was 313 for the total population (and 297 for the reference population), while the effective number of ancestors was remarkably lower (19 in the total population and 17 in the reference population, respectively), and only 7 and 6 ancestors were able to explain 50% of the relationships in the total and the reference populations, respectively. The f_e/f_a ratio was 1.89 and 2.00 in the total and the reference population, respectively (Table 1).

Out of the 6923 animals registered in the overall dataset, 3532 (51.0%) had no progeny. These were 1573 males (22.7%) and 1959 females (28.2%). The scenario was even worse in the reference population, where 1506 males (representing about 88% of the males in the reference population) had no progeny. A plot of the frequency of sires falling in different classes of offspring numbers is shown, for both the total and reference population, in Figure 3. As can be observed, in the total population (Figure 3A), a large number of sires (254, 45.8%) had a number of offspring ranging between 1 and 5. On the other side, 17 sires had more than 50 offspring (and among them, two sires had more than 100 offspring, data not shown). In the reference population, the sires with a number of offspring ranging between 1 and 5 were 15 (0.9%), only one sire had a number of offspring included between 41 and 50 and no sire had a number of offspring higher than 50, suggesting a slight improvement, in the last decades, in the balanced usage of the available reproducing animals.

While mating between known close relatives was limited (2 between full sibs, 120 between half sibs and 28 between parents and offspring; Table 1), the mean inbreeding was 5.22% for the total population (

Table 4. Inbreeding, average relatedness and effective population size estimated for the total and reference populations.

Item	Total Population		Reference Population
	%	Ne	%	Ne
Mean Average Relatedness (ALL generations)	5.22	-	11.00	-
Maximum Average Relatedness (ALL generations)	9.88	-	14.90	-
Average Relatedness SD (ALL generations)	40.4	-	1.60	-
Mean Inbreeding (ALL generations):	3.83	-	6.30	-
Maximum Inbreeding (ALL generations)	16.40	-	40.00	-
Inbreeding SD (ALL generations)	3.01	-	3.50	-
Mean Inbreeding (10 generations):	5.19	-	6.21	-
Maximum Inbreeding (10 generations)	40.45	-	40.44	-
Inbreeding SD (10 generations)	3.82	-	3.45	-
Mean Inbreeding (5 generations):	2.72	-	2.92	-
Maximum Inbreeding (5 generations)	38.28	-	38.28	-
Inbreeding SD (5 generations)	3.31	-	3.27	-
Increase in Inbreeding by Maximum Generation	0.59	84.15	-	-
Increase in Inbreeding by Complete Generation	1.39	35.94	-	-
Increase in Inbreeding by Equivalent Generation	1.05	47.46	-	-
Ne obtained from regression on the birth date	-	49.73	-	47.53
Ne obtained from Log regression on the birth date	-	49.64	-	47.48
Ne obtained from individual increase in inbreeding	-	-	-	53.67
Ne obtained from regression on equivalent generations	-	-	-	36.54
Ne obtained from Log regression on equivalent generations	-	-	-	36.26
Increase in Inbreeding by Equivalent Generation (ΔF = 1/(2Ne)).	-	-	1.36	-

SD, Standard Deviation.

). In the reference population, mating between half sibs and between parents and offspring were, respectively, 13 and 5, while no mating between full sibs was observed. The mean inbreeding in the reference population was higher than in the total population (6.31 vs. 5.22). Since the average relatedness was 9.88% for the total population and 11.10 for the reference population, this possibly suggests a narrow/unbalanced gene pool as a likely explanation of the observed inbreeding values.

Average inbreeding levels per birth year (Figure 4) showed a steady increase, and more precisely, the estimated increase in inbreeding per generation, no matter whether maximum or complete or equivalent generation, was roughly around 1% for the total and the reference populations (Table 4).

The effective population size, either obtained considering the increase in inbreeding by equivalent generation or obtained from regression (or Log regression) on the birth date (Table 4), had values close to 50. Thus, the estimate of N_e/N_c was 0.01, suggesting that the loss of genetic variation per generation is relatively high.

Generation intervals (GI) for each of the four pathways are presented in

Table 5. Average generation intervals (GI) and standard deviations (SD) expressed in years for the four parent–offspring pathways in the whole pedigree of the MH.

Pathway	N	GI	SD
Dam_Son	488	10.7	7.1
Dam Daughter	2588	12.1	10.8
Sire_Son	502	9.5	5.7
Sire Daughter	2627	10.2	8.7

. The sire-to-son pathway was the lowest (9.5 years ± 5.7), and the dam-to-daughter pathway was the highest (12.1 ± 10.8 years). These estimates held high standard deviation values, suggesting heterogeneous management practices within each of the four gamete pathways in the total population. Similarly, for the age at first foaling (AFF), the sire-to-son pathway was the lowest (9.2 years ± 4.8), and the dam-to-daughter pathway was the highest (11.2 ± 9.1 years) (

Table 6. Age at first foaling (AFF) and standard deviations (SD) expressed in years.

Pathway	N	AFF	SD
Dam_Son	2061	10.2	5.4
Dam Daughter	4550	11.2	9.1
Sire_Son	2075	9.2	4.8
Sire_Daughter	4582	9.8	7.4

). For all four gamete pathways, GI values were higher than AFF values.

3.2. Morphometric Diversity

Descriptive statistics for the three considered biometric traits are shown in

Table 7. Summary statistics for the three considered body measurements (cm).

Variable	Sex	N	Mean	SD	CV(%)
Wither height (WH)	Females	2397	156.14	3.40	2.17
	Males	885	159.52	2.97	1.86
Chest girth (CG)	Females	2397	184.33	7.22	3.94
	Males	885	190.25	6.30	3.33
Cannon bone circumference (CBC)	Females	2397	20.21	0.73	3.66
	Males	885	21.27	0.91	4.30

Standard deviation, SD; coefficient of variation, CV.

, disaggregated for the male and female populations. In general, females had slightly lower values, suggestive of a possible, although narrow, sexual dimorphism, as previously observed, as an example, also in Mangalarga Marchador horse [36], Hucul horse [37], and Minorcan horse [38].

When analyzing the variation trend for the three biometric traits in the MH population (

Table 8. Trend of variation, assessed from 1987 to 2017, for wither height (WH), chest girth (CG) and cannon bone circumference (CBC), expressed as mean and standard deviation (SD), both for males and females, in five subsequent periods of about five years each.

Trait	Sex	N	Average ± SD for Each Year’s Group
			1987–1995	1996–2000	2001–2005	2006–2010	2011–2017
WH	Males	885	160.5 ± 2.3	160.2 ± 3.2	160.9 ± 3.5	159.1 ± 2.9	159.2 ± 3.0
	Females	2397	156.2 ± 3.4	156.2 ± 3.8	156.8 ± 4.0	156.2 ± 2.9	156.1 ± 3.0
CG	Males	885	193.9 ± 5.5	189.2 ± 6.2	190.2 ± 5.9	191.7 ± 7.4	189.1 ± 6.3
	Females	2397	186.7 ± 8.7	184.1 ± 7.7	184.3 ± 7.2	185.2 ± 8.0	183.4 ± 7.2
CBC	Males	885	21.5 ± 0.6	21.3 ± 0.7	21.1 ± 0.9	21.3 ± 0.9	21.3 ± 0.9
	Females	2397	20.9 ± 0.8	20.4 ± 0.8	20.4 ± 0.8	20.1 ± 0.9	20.1 ± 0.7

), only very slight changes could be observed over time for both males and females.

4. Discussion

Depth and completeness of pedigrees are of fundamental importance for accurate genetic diversity inferences, as incomplete pedigree information can bias, generally underestimating, inbreeding and relationship coefficients. In the MH breed, the analysis of the pedigree completeness parameters suggests that, while a relatively high number of maximum generations are traced on average (10.09), thus ensuring a good pedigree depth, each animal can count, on average, on a number of equivalent generations that is almost half the average number of maximum generations (5.88); this implies the presence of a certain number of unknown ancestries in the MH pedigree. As predicted, the lack of information was mainly concentrated in the earlier generations, with pedigree completeness falling below 50% in the seventh generation, while being higher than 95% in the first generation. As such, the MH pedigree, while being poorly able to capture ancient inbreeding phenomena, can still accurately detect recent inbreeding. The latter is expected to exhibit larger unfavorable effects than ancient inbreeding [39]; thus, the MH pedigree can be considered a valuable tool for orienting genetic management practices and optimizing conservation strategies [21]. Not surprisingly, percentages of pedigree completeness were higher, at any given generation, in the reference population compared to the total population.

In other horse populations, such as the Slovak Sport Pony [15], the two endangered Spanish horse breeds, Asturcón and Mallorquí [11], the recently revived, yet still “vulnerable” Croatian Posavina Horse [22], the Brazilian Creole horse [19] and the Trakehner [40], pedigrees were not as deep-rooted/complete as in the Murgese horse. Notably, the latter represent an ancient breed, with a partially closed population since the year 1732 and a post-Second World War recovery using the few rescued mares and stallions. Similar pedigree depth/completeness was reported for the Lusitano horse [16]; despite the long tradition of this breed in Portugal for several centuries, when the stud-book was established in 1967, the number of remaining animals was rather small, so, much alike the MH in Italy, the current population is considered to have derived from a reduced number of founder animals. Furthermore, the Maremmano horse breed from Italy [21] was reported to have values for complete and equivalent complete generations very close to those observed in the Murgese. On the other side, for several horse breeds, most of them having a long breeding tradition, such as the Arab [20,41], the Pura Raza Español [10], the Lipizzan [9], the Slovack Hucul [15], the American Shire [17], the Noriker [14] and the Hanoverian warmblood horses [13], deeper pedigrees were reported.

The trends in demography shown in this study help to understand important circumstances affecting the genetic history of MH. In fact, due to the low initial numbers of registered animals, the considered time frame has been subdivided into five periods, each covering two decades, with the exception of the period 2011–2020 (Figure 2). Particularly, the first period, including animals born from 1930 to 1950, gives an account of the bottleneck experienced by the MH population during the first half of the 20th century. As a consequence, more than 60% of the available genealogical records belong to animals born in the last 15 years, supporting the fact that the current population experienced a population expansion in the last decades, as clearly depicted by the time series of registered males and females per year. The similarity between the two sexes in the observed demographic trends may suggest that these were driven by general contingent socio-economical phenomena affecting the entire population (renovated international interest toward the horse species, as well as a revived interest in this specific breed for its good attitude in equestrian sports, mainly dressage), rather than by sex-focused selection practices. This scenario of recent demographic expansion is reported for other horse breeds that had formerly experienced, mainly in Europe, severe population size contractions [41] and followed the present position of the European horse sector, which is growing again thanks to the green assets of equines in the European context of the ecological transition of agriculture [42].

A close-to-unity sex ratio is theoretically possible in the reference population, although the numbers of dams and sires used to generate this population seem to suggest a scenario where, on average, one stallion would serve about five females. This figure is not dissimilar to the total population, likely suggesting that no significant changes have occurred over time for this parameter. However, the effective number of founders in the total Murgese population was about 1/7th of the total number of founders, thus supporting their unbalanced use, confirmed by the distribution of male animals in the total population per class of offspring numbers (Figure 3). When comparing the distribution of male animals per class of offspring numbers in the total population and in the reference population, a higher number of males with no progeny was observed in the latter. This could possibly account for the increased use of the MH in horse riding sports.

Our generation interval (GI) values were longer than the 8.20 years found by Ivanković et al. [22] in the Croatian Posavina, but similar to the 10.1 years reported for the Lusitano horse by Cervantes et al. in 2009 [43] and confirmed by da Silva Faria et al. in 2018 [44]. GIs are higher for dams than sires. This reflects the use of mares for longer periods in the herds than the sires. Murgese horses can be expected to live a long natural lifespan of 25 to 30 years, similar to the general horse population, particularly those partly bred in the woodland. However, in general, there is a large variation in the age of horses because some are killed at a young age owing to illness, breeding problems or injury, whereas some survive until senescence. The age class suffering most from premature mortality is 0 to 1 year of age [45] for both sexes, but because of breeding policies in domestic horses, few stallions are graded for the breeding pool and most of the stallions born are castrated before the age of 3. The fact that 74% of the MH male population (88% in the MH male reference population) have no progeny is a confirmation of the above statement.

The effective number of founders in the MH was 36, while the effective number of ancestors was 19. The rate of the effective number of founders compared to the effective number of ancestors can be used to determine the bottleneck in the population [29], and if the ratio is 1, the population is stable. Higher or lower values of the fe/fa ratio always indicate an imbalanced use of sires, which poses a risk to the original genetic diversity. When a population suffers a demographic decrease, such as that experienced by the MH at the beginning of the last century, fe is overestimated by ignoring some genetic bottleneck effects, but the ratio (f_e/f_a) allows us to consider the loss of genetic variation promoted by the overuse of few sires. The bottleneck ratio (f_e/f_a) in the MH population was 1.89, which is similar to that observed in the Czech-Moravian Belgian horse (1.86), more favorable compared to the Austrian Noriker (4.00) [14] and American Shire (3.65) [17], but lower than that found in the Slovak and Hungarian Hucul horse and the Croatian Posavina, that were 1.6, 1.41 and 1.29, respectively [15,22,37]. The value of the fa/fe ratio in Table 1 indicates a 50% reduction of the genetic diversity as expected by the fact that for the MH current population, the effective founders are 34 and the effective ancestors only 17. In addition, as far as the ancestor contributions to the current population, 50% of the genetic diversity could be accounted for by only six individuals. Basically, these statistics reflect the substantial loss of genetic diversity within the breed over the generations since its official founding.

The effective population size (N_e) is a key parameter in conservation and population genetics because it has a direct relationship with the level of inbreeding, fitness and the amount of genetic variation loss due to random genetic drift [24]. In the MH population, the N_e obtained from the Wright equation was 1118 (data not shown), while that from the increase in inbreeding was 47.48. The estimates of N_e based on the individual increase in inbreeding more accurately reflect the genetic history of the populations, namely the size of their founder population, their mating policy or bottlenecks due to abusive use of reproductive individuals for the period in which the genealogies are known. Thus, figures close to 50 appear more acceptable as estimates of the effective population size for the MH, suggesting that the breed has experienced a significant loss of genetic diversity.

Likely, as a consequence of the narrow breed gene pool, the average relatedness was not negligible; this possibly points to the complexity in mating among unrelated animals (as attested by the mean inbreeding value in the MH, possibly underestimated due to the presence in the reference population of some individuals with unknown parents), although the limited mating between known close relatives seems to suggest certain care in mating planning. Notwithstanding, the estimated increase in inbreeding per generation was roughly around 1%, a value corresponding to the critical threshold generally accepted as the safe limit to assure population survival in the long run.

Effective and resilient population genetic management should be regarded as a balancing act between conservation and genetic improvement. The analysis of the basic morphometric descriptors did not highlight major conformational changes during the last 30 years. This would indirectly suggest that no high selection pressure has been applied in the considered temporal interval to conformational traits in the MH. In view of the growing request for boosting genetic improvement of the MH morpho-functional traits, the implementation of precision genetic breeding practices looks essential to assure a long-term potential for genetic gain and survival.

5. Conclusions

In summary, the overall results from the analysis of genealogical data seem to highlight that the MH population has lost a part of its original genetic diversity mainly due to the bottleneck at the early official breed foundation event and to the unbalanced contribution of founders during the following decades. Efforts in making a more balanced use of the current sires, as well as in exploring the possibility of recovering as sires the not-neutered males currently still without progeny, look of vital importance.