Berganês, a new sheep breed from the Brazilian semiarid: performance and carcass traits in different sex-classes by univariate and multivariate approaches

. Twenty-four lambs between four and five months of age were divided into sex classes: uncastrated males, castrated males and females, eight for each treatment, with an average body weight of 27.00 ± 3.13 kg. dry matter and water intakes were higher for uncastrated males (p < 0.001). final body weight, body weight at slaughter, average daily gain and cold carcass weight were greater for uncastrated, intermediate for castrated and lower for females (p < 0.05). hot and cold carcass yields were higher for castrated males and females, respectively. (p < 0.05). based on principal component analysis, twenty-five variables were selected out of thirty-two, in some of the seven principal components generated, summarizing 21.87% data dimension. Discriminant analysis identified greater discrimination power for body weight at slaughter, thorax width, thorax depth, fatness, dry matter intake and hot carcass weight. All lambs were classified into their respective sex classes. Therefore, uncastrated males have performed better than castrated ones and both showed higher performance than females. Females and castrated males showed higher carcass yield. Discriminant analysis indicated heterogeneity between sex classes.


Introduction
Berganês sheep is an ecotype emerged in the Brazilian semiarid around 1980 and is derived from a crossbreed between Bergamácia and Santa Inês (Silva Filho et al., 2021a;Soares et al., 2019).These animals have large body size and are highly adapted to semiarid conditions, thus becoming an alternative for meat production (Nogueira Filho & Yamamoto 2017).Male castration and female slaughter are common practices in this region.However, no studies have evaluated sex class effects on performance and carcass traits of this sheep breed.
Sex class can affect productive performance and carcass traits since metabolic and hormonal changes influence growth rate and fattening (Everitt & Jury 1966;Rodrigues et al., 2016).In general, uncastrated males have greater testosterone synthesis, which has an anabolic effect, increasing muscle growth (Schanbacher, Crouse, & Ferrell, 1980;Li et al., 2020).Araújo et al. (2017) assessed sex class effect on the performance of Morada Nova sheep and reported that dry matter intake, average daily gain, and body weight at slaughter were greater for uncastrated males, intermediate for castrated, and lower for females.Conversely, females and males may have higher carcass yield, which can be explained by early puberty, resulting in lower carcass weight and higher fat deposition than uncastrated males (Facciolongo et al., 2018;Okeudo & Moss, 2008).This result is important since fat serves as a barrier against extravasation of carcass fluids (Costa et al., 2011;Warner, 2017).
Based on the above, this study aimed to evaluate the effect of sex class on performance and carcass traits of Berganês sheep, using univariate and multivariate statistical methods.

Experiment location and date
The experiment was carried out at the Federal Institute of Sertão Pernambucano (IF Sertão-PE), municipality of Petrolina, state of Pernambuco -Brazil (9° 23' 39'' S latitude, 40° 30' 35'' W longitude, and 373 Acta Scientiarum.Animal Sciences, v. 46, e64120, 2024 m altitude) between July and September 2017.This area has a semiarid climate and total rainfall in 2017 was 140 mm, with an average air temperature of 25°C and relative air humidity of 60%.

Ethical aspects, animals, treatments and experimental design
The study was approved by the Animal Ethics Committee of the IF Sertão, Pernambuco State (under code 029/2017).
We used 24 Berganês lambs, eight of each belonging to one of the following sex classes: uncastrated males, castrated males, and females (eight for each).The animals had an average body weight of 27.00 ± 3.13 kg and aged between four and five months.For evaluations, lambs were distributed in a completely randomized design and housed in 2-m 2 individual pens with a dirt floor and equipped with feeders and drinking fountains.

Diets and chemical analysis
The experiment lasted 70 days, with the first 14 days for adaptation to facilities and handling.Lambs were fed a single diet with a roughage: concentrate ratio of 15:85, containing Buffel grass hay (Cenchrus ciliaris L.), ground corn, soybean meal, calcitic limestone, and mineral mixture.Table 1 lists the ingredient proportions and chemical composition of diets.The diet was based on National Research Council (NRC, 2007) to provide 250 g day -1 gains.Diets were supplied twice daily (8:00 and 15:00h) and adjusted to allow 10% leftovers.Contents of dry matter (DM -method 934.01), crude protein (CP -942.05), and ether extract (EE -920.39) were determined according to Association of Official Analytical Chemists (AOAC, 2000).Neutral (NDF) and acid (ADF) detergent fiber were obtained according to the method described by Van Soest, Robertson, and Lewis (1991).Total digestible nutrients (TDN) were calculated using the equation described by Weiss (1999).

Dry matter, nutrient and water intakes, and animal productive performance
Dry matter intake was determined as follows: DMI = (diet offered daily -leftovers in the feeder the next day) x dry matter.Nutrient intake [CP (CPI), NDF (NDFI), and EE (EEI)] consisted of the difference between their amounts in diets and leftovers.TDN intake was obtained as follows: TDNI = (CPI -CP in feces) + 2.25 (EEI -EE in feces) + (total carbohydrate intake -total carbohydrate in feces).Water intake (WI) was calculated as WI = water offered daily -leftovers in drinking fountains the next day -evapotranspiration in control drinking fountains.Initial (IBW) and final (FBW) body weights were obtained on days 1 and 56, respectively, with animals being weighed after 16 hours fasting.We also estimated average daily gain [ADG = (FBW-IBW) /56] and feed conversion (FC = dry matter intake/average total gain).

Carcass traits and morphometric evaluation
After the end of the experiment, lambs were transported to a slaughterhouse in the municipality of Petrolina, where they fasted for 16 hours and were weighed for body weight at slaughter (BWS) and then slaughtered.Slaughter process started with brain concussion, followed by bleeding, skinning, evisceration, and head and feet removal.Carcasses were weighed to obtain hot weight (HCW) and yield (HCY = HCW / BWS * 100).After 24-hour cooling, carcasses were weighed again for cold weight (CCW), yield (CCY = CCW / BWS * 100), and cooling loss [CL = (HCW -CCW) / HCW * 100] (Cezar & Sousa, 2007).Carcass conformation and fatness were also evaluated using a 0.25-point scale, from 1 to 5, as described by Cezar and Sousa (2007).

Statistical Analysis
Data were evaluated using the Statistical Analysis System (SAS, 2002).The Shapiro-Wilk normality test was carried out using the PROC UNIVARIATE at 5% significance (p < 0.05).Analysis of variance was run using the PROC GLM, with means subjected to Tukey's test at 5% significance (p < 0.05).The variables IBW and HCW were included as covariates for analysis of animal performance and carcass traits, respectively.A principal component analysis (PCA) was performed using PROC PRINCOMP, with principal components retained for analysis by the Kaiser (1960) criterion.A discriminant analysis (DA) was also performed to classify and discriminate individuals for performance and carcass traits.The procedure PROC STEPDISC was used to determine the variables with the greatest discrimination power.The classifier PROC DISCRIM was used to determine the percentages of individual classifications in each sex class.Finally, canonical roots were generated by the PROC CANDISC.
The FBW, BWS, and ADG were higher for uncastrated males, intermediate for castrated males, and lower for females (p < 0.05).Uncastrated males showed lower FC, followed by castrated, and both were lower than FC of females (Table 2).The HCW and CCW were higher for uncastrated males, intermediate for castrated males, and lower for females (p < 0.001) (Table 3).Females and castrated males had higher HCY and CCY (P < 0.05) On average, HCY and CCY were 52.88 and 50.95%, respectively.Lower CL were found for females, intermediate for castrated males, and higher for uncastrated males (P=0.002)(Table 3).

Discussion
The highest values for DMI, CPI, NDFI, TDNI, and WI observed for uncastrated males increased FBW, BWS, and ADG but decreased FC.Increases in nutrient intake promote greater ADG and FBW, with water being involved in muscle growth metabolic reactions (Albuquerque et al., 2020).Other authors (Mahgoub & Lodge, 1994;Pereira et al., 2018) also reported that uncastrated males have better productive performance than other sex classes.
Testosterone blood levels in uncastrated male lambs are higher than in castrated males and emales.Many studies have shown that higher testosterone levels are associated with higher growth rates in sheep (Arnold & Meyer 1988;Adam & Findlay 1997;Araújo et al., 2017).When bound to myofibril receptors, testosterone triggers reactions in the cell nucleus through signaling pathways to activate mammalian target protein of rapamycin (mTOR), which plays a role in structural protein synthesis, such as actin and myosin, in muscle fibers (Vingren et al., 2010).
However, castrated males and females had similar DMI and WI, but castrated males had better productive performance.Therefore, muscle growth decreased in females to start fat deposition, which is related to reproductive development (Rodrigues et al., 2016), as females were approaching the reproductive stage by the time of this study.
The highest HCW and CCW for uncastrated males, intermediate for castrated males, and lower for females also reflect testosterone anabolic effect on male muscles and female pregnancy preparation.These findings are in agreement with Everitt and Jury (1966), who evaluated Southdown × Romney crossbred lambs and with Pereira et al. (2018), who evaluated Morada Nova lambs.
The lowest carcass yield for uncastrated males was already expected.However, unexpectedly, it was not associated with carcass fattening since sex classes did not differ for fatness (P=0.099).Similar results were verified by (Araújo et al., 2017).Complementing our study, Silva Filho et al. (2021b) evaluated non-carcass components and found higher total weights of by-products (head, blood, skin, and feet) for uncastrated males than in castrated males and females.Thus, the greater representation of these non-carcass components in BWS of uncastrated males decreased carcass yield.
The CL results suggest that females and non-castrated males may have higher and lower intramuscular fat contents, respectively.Higher intramuscular fat contents indicate greater availability of molecules or free radicals to make bonds with proteins or water, inhibiting fluid extravasation between muscle fibers (Warner, 2017).
Better carcass conformation in uncastrated males corroborates the results of performance and carcass weight.This confirms a greater muscle hyperplasia in this sex class (Butterfield, 1988).
Uncastrated males also had a greater TD.By studying allometric growth in ruminants, Owens, Dubeski, and Hansont (1993) verified that thorax matures in the penultimate stage of life.Therefore, we may suggest that uncastrated males prolonged thorax growth for longer, thus castration anticipated the maturation of this region.Females are also known to mature earlier.
The PCA showed that 25 out of the 32 variables analyzed had high explanatory power in some of the seven principal components generated, summarizing data dimension by 21.87%.Dataset size reduction and explaining variation in a few PCs is a PCA premise (Härdle & Simar, 2015;Macena et al., 2022) and indicates the technique suitability.
In PC1, most of the total variance (49.73%) indicated FBW, BWS, and ADG for performance results and HCW, CCW, HP and CCI for carcass responses.These variables were also associated with nutrient intake (DM, CP, TDN, and EE).Body growth and carcass weight increases result from the supply of ingested nutrients for fat deposition in body tissues (Santos et al., 2019).
The contrast in PC2 indicates that an increase in FC was related to higher HCY and rib yield, and lower shoulder yield in castrated males and females (Tables 2 and 3).This suggests that earliness in castrated males and females increased FC, HCY, and rib yield, reducing shoulder yield.However, an increased FC is not a desirable characteristic.Moreover, HCY in uncastrated males is compatible with those reported by Vargas Junior et al. ( 2014) for Pantaneiro lambs, and superior to those observed by Araújo et al. (2017) for Morada Nova and by Souza, Selaive-Villarroel, Pereira, Silva, and Oliveira (2016) , for Santa Inês lambs.Ribs and shoulder are cuts of lower commercial value (Cezar & Sousa, 2010), therefore, reductions in the yield of the shoulder and rib increases may not lead to losses in carcass economic value.
The results verified in PC3 indicated that increases in EEI improved HCY and CCY.A high bioavailability of lipids in metabolism facilitates their deposition in muscle tissues (Kowalczyk, Orskov, Robinson, & Stewart, 1977;Fan et al., 2019).Consequently, higher lipid contents in muscles increase availability of molecules or free radicals to bind with water and proteins, preventing further cooling losses and improving carcass yield (Warner, 2017;Corazzin, Bianco, Bovolenta, & Piasentier, 2019) The results verified in PC4 suggest that carcass fatness increase was associated with higher LL.These variables also showed a negative correlation with LCI and shoulder yield.Carcass fatness and LL positive correlation indicates that the ecotype Berganês has long legs and adequate carcass fat cover.Moreover, the high LL was decisive for LCI reduction.
The contrast found in PC5 indicates that the ELC increase was correlated with increases in loin and leg yields, as well as reductions in CL and saw.These are desirable characteristics, since CL reduction increases carcass yield.Similarly, saw reductions, associated increases in with loin, leg, and ELC, improve carcass commercial value (Cezar & Sousa, 2010).Also, the contrast in PC6 is also desired, as indicates that reductions in neck yield increase HCY.
The PC7 indicated that EEI increases enhanced carcass fatness and leg yield, but reduced ELC.As mentioned before, high lipid bioavailability in animal metabolism increases their deposition in body tissues (Fan et al., 2019), which could justify the increase in carcass fattening.As for leg yield, there was an increase in subcutaneous and intramuscular fat contents, reducing its CL and increasing its yield in carcass (Corazzin et al., 2019).Furthermore, an increase in EEI may anticipate fat deposition and tissue growth maturation (Kowalczyk et al., 1977), which may be related to ELC reduction in this study.
The greater discrimination power of BWS, TD, DMI, and HCW can be explained by the statistical difference detected by analysis of variance (Tables 2 and 3).However, selecting thorax width and fatness show the importance of these variables in sex class discrimination in our study.The Wilks's lambda represents the proportion of total variance in discriminant scores not explained by differences between groups (Hair, Black, Anderson, & Tatham, 2009).Therefore, reductions in these values indicate that the model was properly fit to treatment discrimination and classification (Dossa, Wollny, & Gauly, 2007), suggesting heterogeneity between sex classes in our study.

Conclusion
In general, productive performance, carcass weight, and carcass conformation are higher for uncastrated Berganês males, intermediate for castrated males, and lower for females.Uncastrated males have lower carcass yields and higher cooling loss.Except for thoracic length, carcass and meat cut morphometric characteristics are not affected by sex classes in Berganês lambs.Berganês lamb sex classes were heterogeneous, and body weight at slaughter, thorax width, thorax depth, fatness, dry matter intake, and hot carcass weight showed greater discriminatory power.

Figure 1 .
Figure 1.Cumulative variance and eigenvalues (y-axis) generated by principal component analysis for nutrient intake, animal performance, and carcass traits of Berganês lambs from different sex classes.

Figure 3
Figure3illustrates the interrelationship between the variables and the first two components (PC1 and PC2) through a two-dimensional plane.Variables DMI, CPI, EEI, TDNI, FBW, BWS, ADG, HCW, CCW, HP and CCI all approached each other and positioned themselves better along the PC1.In PC2, HCY, FC, and Rib were positively positioned, while shoulder, negatively.

Figure 4 .
Figure 4. Canonical representation of nutrient intake, animal performance, and carcass traits of Berganês lambs from different sex classes.

Table 1 .
Ingredient proportions and diet chemical composition for Berganês lambs from different sex classes.

Table 2 .
Dry matter, nutrient and water intake, and performance of Berganês lambs from different sex classes.

Table 3 .
Carcass traits and commercial meat cuts of Berganês lambs in different sex classes.
CCI = Carcass compactness index; LCI = Leg compactness index; SEM = Standard error of the mean; P = Probability value.Means followed by different letters, in the same row, are significantly different by Tukey's test at 5%.

Table 4 .
Discriminant linear function of the variables selected by the stepwise method and percentages of classification of Berganês lambs of the sex-classes uncastrated males, castrated males and females.