1. Introduction
Animal genetic resources play an important role in local economies and in maintenance of territories and landscapes [
1]. Among the large number of autochthonous cattle populations in Italy, the Aosta breeds play an important role for the Aosta valley, located in the northwest Alpine territories of Italy [
2]. In addition to their milk and meat production, their economic value is also related to the farming activity itself, closely linked to the use of local territories: farming activity, in fact, allows to maintain the territory and the environment and to keep alive the strong cultural and societal value that these breeds and their farming represent.
Aosta breeds are the Aosta Red Pied (Valdostana Pezzata Rossa (VRP)), the Aosta Black Pied (Valdostana Pezzata Nera (VBP)) and the Aosta Chestnut (Valdostana Castana (CAS)). All three are dual-purpose cattle breeds and possess a considerable milk production (2019 average production per lactation: 3000 kg of milk for CAS and VBP and 4000 for VRP) in proportion to their body size (average adult live weight of 550 kg for males and 400 kg for females). Even if their production selection goal includes milk and meat, their capacity to adapt to the harsh alpine environment, i.e., their functionality, has been strongly pursued by farmers for decades [
3]. Additionally, the active mating scheme in these populations, done for decades, was implemented in order to avoid loss of genetic variability while applying some directional selection for milk and meat [
4]. The summer pastures occurring at high altitudes, up to 2500 m above sea level, challenges the cow’s functionality and capability to cope with severe environmental conditions. These include feeding with fresh grass, ability to walk and climb steep territories and resisting harsh climates. The alpine summer pasture is of central economic value for the Aosta Valley region, as this practice maintain territories and landscapes, making them usable for tourism and, as such, closely linked to the local economy. CAS and VBP are appreciated by farmers, also for their particular vitality and impetuosity, which it takes the form of a dominant behavior within the herd. For decades in the Aosta valley, annual tournaments have been organized among the cows of the CAS and VBP cattle from different herds (Batailles des Reines) and weekly disputed from March to October [
5]. The innate instinct for the territoriality and hierarchical dominance of the cows translates first into a non-ferocious ritual of combat, where dominance of a cow over the other is expressed in knock-out battles, leading then to the awarding of the title of Queen of the Valley in a final contest held in Aosta at the “Arena Croix Noir”. Farmers give particular care to select Aosta Chestnut cattle for their temperament [
6], as the cultural value of the “
Battailes des Reines” has a central role for farmers and the Aosta valley’s cultural value and sociality [
1]. Nowadays, CAS and VBP can be considered as three purpose breeds because the fighting ability is a selection criteria included in the IRCMC (Indice Resa Casearia, Muscolosità, Combattività—Cheese yield, Muscularity and Combativeness Index), the official selection index of the two breeds [
7].
The availability of wide genome SNP genotypes on a large scale in cattle made it possible to obtain a detailed picture of the breeds’ genetic diversity across the genome, to investigate genomic variation among populations and to relate their genomic structure to the occurring selection that is undergoing in cattle populations. Runs of Homozygosity (ROH) are long tracts of homozygous DNA that were firstly identified in human populations by [
8]. McQuillan et al. [
9] used ROH to investigate genomic variation among European populations and used ROH to propose a new genomic inbreeding coefficient. They compared different coefficients of inbreeding, finding the largest correlation between genomic inbreeding (F
ROH) and the inbreeding coefficient calculated by genealogical information. Several authors used ROH to explore genomic variation in cattle [
10,
11,
12,
13]. ROH can be used to detect recently occurring autozygosity, generated when related individuals are mated and which discloses the genomic regions under selection, including putative candidate genes. Curik et al. [
14] discussed the use of ROH and F
ROH to identify hotspot regions (i.e., islands), where the frequency of ROH is high, and coldspot regions, where no ROH is found (i.e., deserts).
The nationally funded project DUALBREEDING (PSRN National Program of Rural Development 2017–2020 by Italian Ministry of Agriculture) had among its objectives the disclosure of genomic variation across Aosta breeds and its relationship with occurring selection. Among the activities of DUALBREEDING, all males and a part of the female population have been analyzed with SNP chips at a 150K density providing as such a solid reference genotype database.
The aim of this study was to map ROH in the three Aosta breeds using the 150K SNP chip genotypes available from DUALBREEDING. A further objective was to calculate the FROH based on the ROH and compare it with information derived from F (the inbreeding coefficient). Additionally, we aimed to investigate the genetic variability across breeds using Wright’s F statistics, a Principal Component Analysis and ADMIXTURE analysis. Finally, we aimed to annotate the genes mapped in the ROH in order to disclose the common and proprietary regions under selection in the three Aosta breeds.
4. Discussion
The three Aosta breeds have been sharing the same environment and farming practices for centuries. Nowadays, after the structuring of the breeding activities in herd books, even if they are part of the same herd book association, they are managed as three different populations with different selection indexes. A common denominator is that their milk is used for the Fontina cheese production, a DPO product that, in its manufacturing specifications, includes the rule that only the milk from Aosta cattle breeds can be used.
In term of selection, CAS and VBP are sharing a similar selection goal accounting for milk, meat and fighting ability. Differently, the dual-purpose selection of VRP is more oriented toward milk production [
3]. The differentiation in selection goal occurred for decades and this may reflect the findings in term of genomic regions under selection that may be identified as the ROH.
Even if the three breeds share the same environment and farming structure, they appear to be genetically different. The genetic relationship between CAS and VBP, previously reported using microsatellite markers [
25], was confirmed here with the use of SNP markers, as per the F
ST (both at population and at marker levels), PCA and ADMIXTURE results.
The FST statistic at the population level showed a value of genetic differentiation among the three breeds, which was around 0.05 when the comparison (both) involved VRP, but a lower one (FST = 0.019) between CAS and VBP. At the single-marker level, no differentiated SNPs (FST > 0.5) were identified in the CAS_VBP breed comparison, but several ones were found for the comparison of VRP with the other two breeds. These different values of FST can be affected by the origin of the three Aosta breeds: as known by historical evidences, the origin of VRP is independent from VBP and CAS. In fact, historical findings indicate that VRP was introduced in the area by the Burgundians in the 5th century AC.
It is interesting to mention that these genomic regions are harboring genes involved in the mammalian circadian rhythms regulation (
CLOCK) [
26], feed efficiency and growth traits (
CORIN) [
27], marbling (
LNX1) [
28] and immune response to mammary gland inflammation (
CBFA2T3) [
29]. It is worth noting that the VRP_CAS comparison identified the
KIT and
PDGFRA genes involved in melanogenesis and in coat color (spotting) (
KIT) [
30], as well as in intramuscular adipocyte development and marbling fat deposition (
PDGFRA) [
31], respectively. The coat color and shape (uniform and chestnut in CAS and red and white in VRP) is among the distinctive criteria of the breeds, whereas the differences in fat deposition can be ascribed to the peculiar constitution of the CAS breed. The vitality and attitude for dominance of this breed are a part indeed of a pretty “masculine” phenotype of CAS cows, showing curly hair and a large neck with developed anterior muscle masses, like the neighboring Hérens breed [
32,
33].
The genetic distinctness of the three populations is supported also by the PCA, showing clearly that the three breeds cluster separately: VBP and CAS are both placed on the same spatial position according to PC_1 (
Figure 1A), explaining 41% of the total variance, and both separated from VRP. PC_2 is on the other hand separating the two breeds in two differentiated groups. The ADMIXTURE results also clearly support the evidence that the three breeds have distinct genetic origins.
The ROH analysis is particularly interesting in these three populations as it allows to disclose recent inbreeding caused by the managing of the populations, for example, by performing artificial insemination and with structured breeding plans, even if the selection scheme was carefully evaluated and planned to minimize the increasing of inbreeding [
4].
The recent inbreeding’s loop produces small numbers of long ROH, influencing the sum of ROH much more than the total number of ROH itself. In the VRP, the regions linked to recent inbreeding are more frequent than the ones found in the other two breeds. In VRP, in fact, an ROH > 16 Mb has been found in 34% of samples (of which 28 animals had an ROH longer than 30 Mb), while in CAS and VBP, it was found in 26% and 30% of animals, respectively. Among the latter, 15 CAS and 12 VBP individuals had an ROH longer than 30 Mb.
Although a similar proportion of homozygous SNPs was identified among the three breeds (around 67%—
Table 6), a higher number of homozygous SNPs concentrated in the ROH was detected in VRP (10.12%—calculated as the number of SNPs defining ROH/observed homozygotes). In CAS and in VBP, the proportion of homozygous SNPs mapping within the ROH is lower: 8.8% and 8.04%, respectively.
In VRP, a higher number of ROH and a longer size with respect to CAS and VBP were identified, mainly as a consequence of no introduction of animals from other regions. This is not the case for CAS, who has been recently recognized as genetically similar to the Hérens breed. The possibility to enroll progeny of the two breeds in the studbook of any of them was recently approved by the two breeder’s associations. This recent advance in reciprocal recognition occurred after some generation of known exchange of reproducers. In fact, already in 1929 [
34], the closeness of CAS with Hérens was highlighted and the author already raised the hypothesis that the two breeds could be recognized genetically as one population. A previous study on microsatellites, carried out on the most important cattle breeds of the Alpine arc [
25], also support the strong relationship between these two breeds. Before the approval of using Hérens bulls for breeding, crossbred mating was, however, already occurring between the CAS and VBP breeds, contributing to maintain the average level of inbreeding lower than in VRP (both
F and F
ROH).
The ROH are not randomly distributed across the genome and there are regions with a high prevalence of ROH. In the CAS breed, on the BTA5 located around 79–80 Mb, the highest number of samples (
n = 94—31%) that shared the same TOP_ROH was identified. This TOP_ROH, defined by 18 homozygous SNPs (of which 5 are intronic), lies within the
TECRL gene encoding for an enzyme involved in chemical reactions and pathways involving lipids, also reported to possibly play a role in puberty and female fertility in cattle [
35]. In CAS, within the TOP_ROH on BTA4 11, the genes belong to the Homeobox family genes, of which
HOXA13, HOXA11, HOXA10, HOXA9, HOXA7, HOXA5, HOXA3 and
HOXA4 are involved in the reproductive tract and in development and fertility in males and females [
36,
37,
38]. It interesting to note that a negative genetic relationships occurs between fighting ability and fertility, as observed in CAS [
33], and a reduction in fertility has been found in Hérens cows, also empirically, for fighting ability across time [
39]. The breed has no problems with fertility, like all Aosta breeds, well known for their hardiness, but a strong selection towards fighting ability; disregarding fitness characteristics could, in the long term, have a detrimental effect on fertility.
The second highest peak (TOP_ROH) found in CAS is located on chromosome 5 (max number of samples = 78, 25.4%), where the
KITLG maps. This gene was included in regions under selection in cattle breeds [
40,
41] and is responsible for the coat color phenotype in different species [
42,
43]. In VBP, two higher peaks were identified on BTA6 and BTA23, where the maximum number of samples (
n = 59, 38.5%) and (
n = 45, 29.4%), respectively, shared intergenic and intragenic (
KHDRBS2 gene) homozygous regions. The
KHDRBS2 gene has been associated with fertility traits in goats [
44] and in Brahman cows [
45], and more recently with adaptability traits in Colombian cattle breeds [
46]. This finding may relate to the VBP as well as other Aosta breeds, which are well adapted to harsh environmental conditions, such as the alpine pastures farming. Lastly, the regions shared by a higher number of VRP samples (more than half) are those found on chromosomes 2 (
n = 150, 53.6%), 5 (
n = 157, 56%) and 6 (
n = 180, 64.3%). Except for the most representative regions on BTA2, for which an SNP maps in the intronic position of the
DARS1 gene, the other ones include SNPs lying closely to the
KITLG (BTA5) and
KIT (BTA6) genes, respectively.
We would like to underline the impact of selection on highly homozygous regions: several interesting genes in TOP_ROH include a large number of SNPs annotated within the gene itself. Although this information may be affected by the density of SNPs on the chip and by the length of the genes, it would be worth mentioning them. The
THSD7B gene of the VRP’s TOP_ROH has 48 intronic SNPs annotated (
Table S3), with an average distance among them of 17.7 kb, which is consistent with the average marker distance for this SNP chip spacing of approximately 19 kb (
https://genomics.neogen.com/en/ggp-hd150k-dairy). For this gene, classified as an integral component of membranes (GO:0016021), no association study, at the best of our knowledge, is available. Another example is represented by the
KHDRBS2 gene on BTA23 (described above in the text), in which 33 homozygous intronic SNPs are annotated for all the three Aosta breeds at an average distance of 20.5 kb. Lastly,
CTNNA3 (
n = 32 intronic SNPs spacing 20 kb), both in CAS and VBP, and
ADGRL3 (
n = 27 intronic SNPs spacing 20.2 kb),
CCSER1 (
n = 26 intronic SNPs spacing 17.9 kb) and
CACNA2D1 (
n = 25 intronic SNPs plus a synonymous variant spacing 18.8 kb) in CAS are genes mostly annotated with homozygous SNPs. These genes were previously shown to be under positive selection and associated with marbling score in Korean cattle (
CTNNA3) [
47], with protein yield and percentage (
ADGRL3) [
48], with aggressiveness during gestation (
CCSER1) [
49] and with carcass and meat quality traits in the cattle (
CACNA2D1) [
50]. This is in line with the selection performed in the three Aosta breeds.
In VRP and VBP, a homozygous cluster on BTA6, involving GABA-A receptor subunit genes (
GABRA2, GABRA4, GABRB1 and
GABRG1), was identified. These genes being the major inhibitory neurotransmitters in the mammalian brain, mediating anxiolytic activity and playing a key role in emotional and behavioral control in humans [
51,
52]. Even if speculative, we may comment that the fact that this region does not appear to be in autozygosity in CAS may lead to variability in expression regarding the behavior of cows and for the specific selection operated by the farmers on this breed for the “
Battailes des Reines”. We may assume that, during the non-ferocious match, the reaction to the visual and physical view of the opponent may be mediated by these groups of genes. The behaviors exhibited during the matches is the same that cows show in pastures when unfamiliar individuals from different herds meet at the beginning of the summer season.