Factors Affecting Milk Yield and Composition of Indigenous Balkan Goat Breed Reared in Semi Extensive Conditions

Goats are reared in different geographical and climatic conditions and play very important role in less favored areas (Memisi et al., 2004, Bolacali et al., 2017). The abundance of goat populations in the harsh, arid, semi-arid areas most likely reflects their good adaptation to such environments, especially indigenous breeds (Kumar et al., 2010). Lack of information on production potential of local breeds is one of the main reasons why farmers replace local breeds  by highly productive breeds (Curro et al., 2019). Many authors have studied effect of different environmental factors on milk production traits of goats, like age or lactation, stage of lactation, season and others (Antunac et al., 2001, Ciappesoni et al., 2004; Mia et al., 2014). Some of the previous studies have shown that yield and composition of goat milk were affected by breed (Kume et al., 2012), by genetic groups in the same breed (Silva et al., 2013) and also by genotype of alpha S1 casein (Vazquez-Flores et al., (2012), Soares et al., 2009).
        
Balkan goat is the most important indigenous breed in Montenegro, where it constitutes about 50% of total goat population. Many farmers are still interested in rearing Balkan goat breed, primarily due to its advantages for rearing in semi extensive production systems. They also intend to improve its production potential, especially milk production. One of the prerequisite for that improvement is to develop breeding program based on comprehensive knowledge on the production traits. Thus, the objective of this study was to analyze effect of environmental and genetic factors on milk traits of Balkan goat breed reared in semi extensive production systems.

Investigation was carried out on 338 goats of Balkan breed, reared in four family farms from south part of Montenegro that is characterized by Mediterranean climate with hot and dry summers and rainy winters. System of rearing was semi extensive, based on grazing over the year. Hay and small amount of concentrate feeds were used in nutrition only at the beginning of lactation.  
        
For the purpose of this study, milk recording was organized in accordance to ICAR rules for AT control. The first recording was done 5 to 15 days after kidding, the next ones were consequently until the end of lactation, with 28-33 days between subsequent milk test days. Animals were milked by hand twice per day (morning and evening milking). Milk yield was measured and milk sample (50 ml) was taken from each animal in the morning milking. Samples delivered to the laboratory refrigerated at 4°C temperature. The total 2159 milk samples were collected in seven consecutive test days. Milk samples were analyzed on fat, protein and solid non-fat content in the Dairy lab at Biotechnical faculty using infrared spectrophotometry on MilkoScan FT120. 
        
The following production traits were studied: milk yield in lactation (MY, kg), days in lactation (DIL, days) or lactation duration, daily milk yield (DMY, kg), fat content (F, %), protein content (P, %), solid non-fat content (SNF, %). All recorded animals were grouped into four lactation groups (first, second and third, while doers in fourth and subsequent lactations were treated together). Clearly distinct strains in coat color or genetic groups of Balkan goats were treated as three separate groups: red – R, black – B and multicolored strain – MC. 
        
In order to analyze relationship of allelic variants of alpha S1 casein with milk yield and milk composition, genotyping was done for 94 animals randomly chosen from all flocks. Genomic DNA was extracted by phenol-chloroform method, while the 750 bp sequence of alpha S1 casein gene from exon 12 was amplified by the Polymerase Chain Reaction (PCR). Sequencing of alleles A and F was carried out by ABI PRISM®BigDye™ Terminator Cycle Sequencing Ready Reaction Kit and the automatic DNA-sequencer Model 377.
        
In order to study variability of milk traits, three different statistical models were applied: Model 1 for total number days in lactation (DIL) and milk yield in total lactation (MY), Model 2 for daily milk yield and milk composition and Model 3 to estimate the effect of genotype alpha S1 casein on milk yield and composition. The models were as follow:
 
Model 1:  Yijkl = µ + Fi + Lj + Vk + (FL) ij + eijkl
Model 2: Yijklm = µ + Fi + Lj + Vk + R(ij) + Cl + (FC)il + +  eijklm
Model 3: Yijkl = µ + Fi+ +  Gj + R(ij) + Cl + (FC)il + eijkl
 
Where
Yijkl, Yijklm = the phenotypic value of DIM (days), MY (kg),fat, or protein or SNF (%),
μ - mean of the analyzed traits,
Fi - fixed effect of flock, (i = 1, … 4),
Lj - fixed effect of lactation (j = 1, … 4),
Vk - fixed effect of strain, (k = 1, 2, 3) 
Cl - fixed effect of stage of lactation or test day (l = 1, 2, …7),
Gj - fixed effect of the genotype of alpha S1 casein (j = 1, 2, 3)
R(ij) - random effect of animal - doer
(FL)ij - effect of interaction between flock and lactation 
(FC)il - effect of interaction between flock and  stage of lactation,eijkl, eijklm  - residuals.

Statistical analysis was performed by using the generalized linear model (PROC GLM) procedure of the Statistical Analysis System program (SAS, 2009, SAS Institute Inc., Cary, NC, USA). 

Montenegrin population of Balkan goat breed reared under semi-extensive conditions is relatively low productive breed with average milk yield of 150.43 kg in lactation lasted 211.09 days (Table 1). However, very wide range of milk yield variation (from 67.4 kg to 260.2 kg) suggests that this breed could achieve significantly better production results, by appropriate selection even in the existing population. Much higher milk yield was determined by Bogdanović et al., (2010) for Balkan goat breed reared in Serbia, where the rearing conditions were generally better, as well as Kominakis et al., (2000) for Greek dairy Skopelos breed reared in semi-intensive conditions. Similar results were reported by Memisi et al., (2004) and Kume et al., (2012) for local goat breeds reared in the region of South East Europe (Kosovo and Albania) under extensive condition. Much lower daily milk yield and milk yield in total lactation were obtained for indigenous goat breeds reared in arid or semi-arid condition of Asia and Africa (Alkass and Merkhan, 2011; Mia et al., 2014. and Rout et al., 2017).
 

        
The mean values of fat and protein content (3.35% and 3.29%) were lower than the results reported for Italian local breeds Girgentina, Maltese, Jonica and Mediteranean Red (Curro et al., 2019 and Carnicella et al., 2008). Regarding milk composition, our results for fat, protein and SNF content were in agreement with the results of Mioè et al., (2007) and Zazharska et al., (2018) for Alpine, Saanen or Anglo Nubian breeds. Our results confirm that Montenegrin population of Balkan goat breed possesses very important biological variability, typical for autochthonous breeds, which can be used for improvement of milk production traits.
 
Effect of flock, lactation and strain on MY and DIL
 
High statistical significance of Model 1 showed that the observed factors explained significant part of the phenotypic variability of MY and DIL (Table 2). These two traits were significantly affected (P<0.001) by flock and consecutive lactation, while effect of strain or genetic group was not significant (P>0.05). Obtained determination coefficient (R²) for DIL and total MY is very low (0.295 and 0.391) that corresponds to generally low heritability for these traits. Wide range of variation in MY among flocks (from 129.3 kg to 171.7 kg) indicates different level of applied selection and different rearing conditions in the flocks (Table 3). As expected, MY and DIL were increased from first to third lactation and decreased in subsequent lactations. Very low milk yield in first lactation (only 110.3 kg) was a consequence of too early first mating of females, at age of 9 to 10 months. Lower milk production in the first lactation is in agreement with results obtained in Mediterranean, North-European and Mexican breeds (Fernandez et al., 2002; Macciotta et al., 2005), as well as with results of  Mavrogenis et al., (2006)  for Damascus goats in Cyprus. Similar to our results, significant effect of flock and consecutive lactation on MY and DIL reported also Goetsch et al., (2011) and Mia et al., (2014).
 

 

 
Effect of the observed factors on DMY and milk composition
 
The applied model showed that flock, stage of lactation, interactions of these two factors and random effect of animals significantly affected DMY and milk composition (Table 2). Consecutive lactation significantly affected only DMY. The coefficient of determination (R²) for milk composition parameters was ranged from 0.478 for fat to 0.506 for SNF content and for daily milk yield (0.675). It means that larger part of variations explained by the applied model and that is in accordance with the results of Ciapresoni et al., (2004), Bogdanović et al., (2010).
        
Daily milk yield and milk composition significantly varied across the observed flocks (P<0.05). These differences caused by different rearing conditions, nutrition and level of applied selection in the flocks. DMY in the first lactation (0.54 kg) was significantly lower than in the subsequent lactations (P≤0.05) (Table 3). Regarding milk composition, there were no significant differences between lactations (P>0.05). These results are in accordance with results of Memisi et al., (2004) for Balkan goat at Kosovo and Upadhyay et al., (2014) for Indian local goat breeds.
        
Stage of lactation had the strongest effect on variation of DMY and milk composition. DMY was increased from the first to the second test day and after that it has slowly been decreased to the end of lactation (Fig 1). Fat content has been increased from the beginning to the end of lactation, with significant differences between each consecutive test days (P<0.05), while the protein and SNF content were relatively stable in the first half of lactation (approximately up to 120 days) and after that these components have been constantly increased to the end of lactation (Fig 1). These trends of protein and SNF content are in agreement with the results of Antunac et al., (2001). Similar to our results, significant effect of flock on DMY, fat and protein content was obtained in the research conducted by Ciapresoni (2004) for Czech shorthair breed.
 

 
Effect of genotype of alpha S1 casein on DMY and milk composition
 
The results of genotyping of alpha S1 casein gene showed presence of three genotypes (AA, AF and FF) and domination of “strong” allele A (0.633) which related to the higher protein content in milk over “weak” F allele (0.367). The model 3 showed that genotype alpha S1 casein had no significant effect on DMY, while significantly affected milk components. Relatively high level of determination (R2), ranged from 0.522 for fat content to 0.655 for DMY (Table 2), indicates that applied model described more than 50% of total variation. Goats with AA genotype of alpha S1 casein had significantly lower DMY, than FF genotype (Table 4). On the other side, AA genotype had higher fat, protein and SNF content (3.34 %, 3.46 % and 8.49%, respectively) than other two genotypes (AF and FF) and differences among them were significant (P£0.05).
 

        
Slightly higher milk yield in AF and FF genotypes than in AA genotype is in agreement with the results of Szatankoova et al., (2013) obtained for Check dairy goat. Close relationship of AA genotype with higher content of protein and fat obtained for Balkan goat breed is in agreement with studies of Vazquez-Flores et al., (2012). Higher protein content in goat milk, influenced by AA genotype of alpha S1 casein gene, resulting in higher cheese yield, even up to 20% (Moioli et al., 2007; Maga et al., 2009). Though, the genotype of alpha S1 casein can be used as selection criteria to improve these traits of goats.

The obtained results show that Montenegrin population of Balkan goat possesses high phenotypic variability of analyzed milk production traits. These results could be good base for designing breeding program for improvement of its production traits. Since this breed has not been exposed to selective pressure so far, the existing high variability in production potential would enable relatively fast progress and significant achievements. For successful improvement  it is also necessary to apply adequate rearing system,  primarily by improving  nutrition and rearing conditions. After improvement production capacity, the next selection step could be focused on increasing of presence of AA genotype of alpha S1 casein.