Selection Criteria for Meat Production Based on the Morphometry of Morada Nova Lambs: A Systematic Approach


 The objective of this study was to characterize the meat production potential of Morada Nova lambs from different genetic groups using morphological and zoometric data, performance and carcass characteristics. A total of 24 non-castrated male lambs from two genetic groups were used: Morada Nova red (MNR) and Morada Nova F1 red x Morada Nova white (MNF1) breed, with an initial mean body weight of 16.9 ± 3.43 kg and an average age of four months were distributed in a completely randomized design. Evaluations of morphometric data and zoom indices, performance and carcass characteristics were measured and calculated. Univariate and multivariate techniques were used to differentiate and reduce the number of variables, to group them and then develop and validate a sheep meat production index. Only 12 differences in the 29 morphometric measurements and zoometric indices analyzed were observed (P < 0.05). Furthermore, only 7 of the 28 quantitative and qualitative characteristics of the carcass were greater (P < 0.05) in MNF1 lambs. The yields of commercial cuts and the physical-chemical characteristics and qualitative measurements of the carcass were similar (P > 0.05) between the genetic groups. All groups of variables, except the cut yield, have discriminatory power (P < 0.05) with the highest classification percentages for body morphometry (91.3%) and zoometric indices (95.7%). The four identified factors, which together explain more than 85.0% of the total data variation, were renamed in the following order of importance: morpho-economic (38.5%), morpho-productive (18.8%), morpho-commercial (17.5%), and equilibrium (11.0%). The heat map showed the suitability in the morpho-commercial and morpho-economic factors for MNF1 lambs, while the MNR lambs showed emphasis for the equilibrium factor. Thus, the Sheep Meat Morphometric Index (SMMI) resulted in the following equation SMMI =\(\frac{\text{A}\text{I} \times 0.771 + \text{C}\text{H} \times 0.759 + \text{R}\text{H} \times 0.878 + \text{T}\text{P}\text{R}\text{I} \times 0.876 + \text{B}\text{I} \times 0.550 + \text{C}\text{W} \times 0.905}{100}\) In which: AI is the anamorphosis index; CH is chest height; RH is rump height; TPRI is the thoracic perimeter relation index; BI is body index; and CW is chest width. The application and validation of the developed index showed superiority (P < 0.001) for meat production in the crossed lambs. MNR and MNF1 lambs have high thoracic development, compact character and aptitude for meat production. Lambs crossed between genetic groups of the MN breed are another sustainable option for sheep production in drylands. It is recommended that the SMMI be used as a tool for morpho-productive selection criteria in breeding programs in the MN breed, as well as other small native breeds.


Introduction
The Brazilian semi-arid region has traditionally raised sheep with a total of 68.5% of the national effective herd [Brazilian Institute of Geography and Statistics (IBGE, 2018)], with the potential for semi-intensive exploitation in sheep meat farming mainly using native breeds and their crosses to increase meat production (Castro et al., 2020;Silveira et al., 2021a). Among these breeds, Morada Nova (M N ) is one of the main genetic resources of hair sheep in the Brazilian semi-arid region, being recognized for its Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js resistance to gastrointestinal endoparasites (Ferreira et al., 2019), adaptability to the thermal environment of the semi-arid region (Leite et al., 2018;, productive potential as a maternal race in terminal crossbreeding , fertility without seasonal in uence (Moura et al., 2019), carcasses with fatty acid pro les with potential bene t to human health (Lage et al., 2020) and excellent productive potential and skin quality (Vasconcelos-Filho et al., 2021; Alencar et al., 2021).
The Brazilian Association of Sheep Breeders (Associação Brasileira de Criadores de Ovinos -ARCO) recognizes the genetic record of two genetic groups of the M N breed: red and white hair. These present genetic differentiation between themselves, being considered distinct genetic groups . However, there is population fragility with the native breeds due to extinction risk, low effective number of the herd and increased inbreeding rate caused by the indiscriminate introduction of exotic breeds . Speci cally, the creation of the MN breed is still undervalued in production systems, with the white genetic group being empirically recognized by farmers as "not adapted" to the thermal environment of the Brazilian semi-arid region, and the red genetic group for its small size. Thus, conducting a survey on the breeding pattern and meat production potential of this breed is essential to encourage the sustainable insertion of these animals in sheep production systems (Lobo, 2019).
A tool used to determine the breeding pattern is by means of body biometrics, which is performed from metric measurements on the animal's body in vivo. This simple tool is applied in commercial herds and/or in nuclei to preserve genetic resources in order to assist in zootechnical characteristic studies.
With the association of these measures from different body regions, it is possible to generate zoom indices which identify the aptitude and productive potential of the animals, de ning the breeding pattern and ability for commercial exploitation (Guimarães, 2018), in addition to relating these indicators to the performance and carcass characteristics. The diversity of this information requires adopting a systematic approach to explore the relationship and identify the main predictor variables for studies of this nature.
Multivariate analysis techniques are useful tools for relating sets of indicators (Silveira et al., 2021b) and to develop indexes which assist in decision-making for animal selection and adaptability (de Vasconcelos et al., 2020).
To the best of our knowledge, there is no information on characterizing the potential for meat production using zoomometric measures and indices, nor are there any studies evaluating the performance using the white M N breed genetic group crossbred with the red (M NR ). Therefore, this study aimed to characterize the meat production potential of M NR and M N F 1 lambs using morphological and zoometric data, performance and carcass characteristics. and 56 m asl). The region's climate is Bsh (B -Dry, S -Semi-arid, low h-latitude) according to the Köppen Climate Classi cation (Alvares et al., 2013).

Animals, experimental design and diet
A total of 24 non-castrated male lambs from two genetic groups were used: Morada Nova red (M NR ) and F 1 Morada Nova red x Morada Nova white (M N F 1 ) with an average initial body weight of 16.9 ± 3.43 kg and an average age of four months distributed in a completely randomized design. The lambs were fed a standard isoproteic and isoenergetic diet during the experimental period according to the nutritional requirements recommended by the National Research Council (NRC; 2007) for lambs of late maturity in the nishing phase with a predicted weight gain of 200 g day -1 . Water and mineral salt were supplied ad libitum.

Measurements and morphometric indices in vivo
Morphometry was measured in the last week prior to slaughtering the animals with a tape measure and a biometric rod adapted according to (Bianchini et  index provides an understanding of the general shape of the animal's body, meaning if it is long and leggy or short, also indicating production capacity; and 9) Thoracic body index (T BI ): ratio given by dividing body length and withers height by withers height and chest depth, multiplied by 100.

Slaughter and carcass characteristics
The lambs were slaughtered at approximately seven months of age. The animals were fasted on a solid diet for 16 hours and then weighed prior to slaughter. The slaughter procedure was carried out following the rules of the Industrial and Sanitary Inspection of Products of Animal Origin (RIISPOA -Regulamento Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js da Inspeção Industrial e Sanitária de Produtos de Origem Animal) (Brazil, 2017). Then, the non-carcass components were eviscerated and weighed. After obtaining the carcasses, the pH 0 hour after slaughter was measured using a portable digital pH meter (Testo 205, São Paulo, Brazil) provided with a glass electrode, calibrated with pH 7.0 and pH buffer solution 4.0. The carcasses were identi ed and weighed to determine the hot carcass weight (H CW ). The empty body weight (E BW ) was determined according to Silva Sobrinho, (2001). Then, the carcasses were packed in plastic bags and sent to a cold chamber (4°C for 24 hours). After this period, the cold carcass was weighed and the pH 24 hours after slaughter was determined. The cooling weight loss [C WL ,%) = (H CW -C CW ) × 100/H CW ], in which: H CW = hot carcass weight (kg) and C CW = cold carcass weight (kg). The cuts and carcass composition were longitudinally sectioned using a cold saw (G Paniz SF248, Rio Grande do Sul, Brazil), and the left half carcasses were divided into six commercial cuts: leg, shoulder, rib, breast, neck, and loin. These cuts were individually weighed to determine their respective yields (%). In addition, subcutaneous fat thickness (S FT ) was measured using a 200 mm digital caliper (ZAAS-01.0013, São Paulo, Brazil) (Silva Sobrinho, 1999).
Finally, the loin eye area of the Longissimus lumborum (L L ) muscle exposed on the 12 th thoracic rib was calculated using the geometric method.
Qualitative analysis of the carcass Conformation, nishing and evaluation of perireneal fat were determined by visually evaluating the carcass according to the methodology described by Cezar and Souza (2007). For conformation, a score of 1 was given for poor conformation to 5 as very good according to the musculature and adipose tissue deposition in relation to the skeleton. Meanwhile, nishing was classi ed into ve categories (score 1 = very thin and 5 = very fat) according to the subcutaneous fat amount and distribution in the chilled carcass. A score of 1 to 3 was assigned for perirenal fat assessment according to the amount of fat in the abdominal cavity around the kidneys (1 = left kidney without fat cover; 2 = fully coated left kidney and the right partially; and 3 = both are covered by a thick layer of fat).

Morphometric measurements and carcass indexes
The morphometric measurements were obtained from the carcass and were carried out as recommended

nd
Step(Simple effect): The data were submitted to ANOVA followed by Tukey's means test (P < 0.05) for all continuous variables. The effect of genetic groups at two levels (M NR and M N F 1 ) was considered in the linear mode of ANOVA. The Kruskal-Wallis non-parametric analysis (P < 0.05) was used for qualitative carcass variables which did not present a normal distribution using the median as a measure of central tendency.

rd
Step(Differentiation): Canonical discriminant analysis (C DA ) was used to differentiate genetic groups into 9 groups: I. body morphology, II. zoometric indices; III. carcass and muscle; IV. carcass morphometry; V. carcass indices; VI. carcass performance; VII. commercial cuts; VIII. cut yield; and IX. mix of variables (all variables simultaneously). The homogeneity of multivariate covariance matrices was assessed using Box's M statistical test (P > 0.05). The discriminant function signi cance was determined using the Wilks' Lambda statistic (P < 0.05). The main discriminating characteristics between genetic groups was identi ed by the highest value of the standardized coe cient of the canonical discriminating function (r > 0.50).

th
Step (Identi cation of variables): A factor analysis (F A ) was performed to identify the most relevant characteristics in meat production for the evaluated genetic groups. The analysis was initially performed for each group of characteristics in a similar way to that described in the C DA . After identifying the main variables within each group of characteristics, an F A was performed to extract the factors. The number of extracted factors followed the Kaiser criterion (eigenvalues > 1; (Kasier, 1960) with orthogonal rotation (varimax method). The factors were renamed according to the importance of the factor weights of the variables within each factor (> ± 0.50). The model adequacy was assessed using Kaiser-Meyer-Olkin statistical test (KMO = 0.59) and the signi cance by Bartlett's sphericity test (P < 0.05).

th
Step(Clusters): The de nition of the potential for each genetic group according to the renamed factors was performed through the hierarchical cluster analysis (HCA). Ward's method (Ward, 1963) and Euclidean distance were used to measure similarity. Next, individual eigenvalues of the M NR and M N F 1 lambs were used in the regression method to generate the matrix. From these, the average factor score by genetic group was generated and for each renamed factor in Step 4. As a complementary analysis, the result of the grouping was plotted using a heatmap generated using the HeatMapper software program (Babicki et al., 2016).
to facilitate identifying lamb selection to provide greater potential for meat production. The morphometric characteristics and zoom indices which integrated the index were selected according to the commonality of each characteristic in the F A . The S MMI was built from the factor weights of the variables as performed by de Vasconcelos et al., (2020) using morphometric measurements, zoom indices and productive characteristics. An ANOVA between genetic groups was performed followed by a means comparison by the Tukey test (P < 0.001) to validate and apply the S MMI .

Results
Differences (P< 0.05) were observed in only 12 of the 29 variables for the body morphometric characterization and zoometric indices ( Table 1). The zoometric indices (B I , T PRI , A I , B SI , T BI ) classify both genetic groups as longitudinal, compact biotype, prone to meat production ability and respiratory capacity. Differences were observed (P < 0.05) for the anamorphosis index (A I ) and thoracic body index (T BI ). Regarding the carcass measurements and indexes: carcass chest girth, arm length, carcass inner width, and carcass compactness index were heavier (P < 0.05) in the crossbred lambs.
Regarding the productive and physico-chemical characteristics of the carcass, commercial cuts and cut yields ( Table 2), only 7 of the 28 variables differed between genetic groups, observing a higher value (P < 0.05) for M N F 1 lambs. Regarding these variables, the fasting live weight, hot carcass weight, rib eye area and rib, loin and ank weights stood out. Cut yields and carcass physicochemical characteristics were similar (P > 0.05) between genetic groups. Qualitative characteristics (conformation, nishing and evaluation of perirenal fat) did not show an effect on genetic groups (P < 0.05).
The C DA showed that it is possible to differentiate M NR from M N F 1 lambs from the groups of the 9 clusters (Table 3). The zoometric indices were the group of variables which had the highest percentage of correctly classi ed cases (C CC ) in their group of origin, followed by morphometry. The mix of variables in the simultaneous assessment of all variables belonging to the discriminating groups pointed out that T BI , leg yield, C G and rib weight are the main predictor variables to differentiate purebred lambs from crossbred. The yield of the cuts in isolation did not show any power to differentiate (P = 0.092) the genetic groups.
The eigenvalues and commonality of the variables distributed in the factors are shown in Table 4. High commonality values (> 0.90) were obtained in the ranked variables of C WL , T PRI , A I and C EI , while the other variables had values > 0.70. The rst four factors explained 85.90% of the total variation in the data, in which the rst factor retained 38.5% of the total variance, and constituted the variables with the highest factor load, being renamed morpho-economic. There was a positive association between C CI , L EA , R H , C H , T D , C W , A I and C EI , but negative with C WL . Higher coe cients in the morpho-productive factor were obtained by C Y , C WL and T PRI , with 18.8% of the explained variance. The factor associated with the commercial characterization was composed of two variables with greater load (Leg Y. and B I ). Finally, the Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js equilibrium factor was given by the list of variables which met the analysis assumptions (> 0.5), and that it was not possible to allocate these variables in the rst three factors due to their compared low factor load.
The heat map (Figure 1) demonstrated the dynamics of the genetic groups based on the four factors extracted, thus revealing the productive potential of each group, in which the M N F 1 lambs showed aptitude in the morpho-commercial and morpho-economic aspects (positive value of Z -scores) and M NR showed a high association for the balance factor. There was also the formation of two clusters: (i) morpho -economic and morpho-commercial; and (ii) morpho-productive and equilibrium.
The S MMI resulted in the following equation: The application and validation of the S MMI is shown in Figure 2, in which crossbred lambs presented greater (P < 0.001) potential for meat production.

Discussion
The morpho-functional and productive characterization of the M N breed is non-existent, since the morphological evaluation is limited to considerations about breeding pattern and is only used in promotional events, fairs and exhibitions organized by breeders' associations. This study assesses the productive potential of the M N breed through morphometric, zoometric, productive data and carcass characteristics using a systematic approach with successive multivariate techniques for different objectives.
Our main ndings were: (i) zoometric analyzes show the capacity for morpho-productive classi cation of lambs from the two genetic groups of the M N breed; (ii) the zoometric indices and body morphometry are the main factors responsible for differentiation between the genetic groups evaluated in the study; and (iii) the S MMI has potential in the area of selection criteria for conservation programs and genetic improvement of small native sheep breeds.
The morphological indices used in this study were able to characterize the phenotypic and production potential of M N lambs using the T PRI , A I , B CI "1" and B CI "2" variables, enabling the classi cation of breeding groups in high thoracic development, ability to produce meat, weight gain, feeding e ciency and breathing capacity, infusing greater breed rusticity and adaptability for both genetic groups. These characteristics are in agreement with other studies which evaluated adaptability and productive potential of the M N breed, especially the red variety which has greater potential for meat production (Melo et  The superiority observed in some morphometric and productive characteristics in crossbred lambs are justi ed by the fact that this crossbreed brought together intrinsic characteristics of the white group, meaning that we still observe the result of the heterosis process and complementarity of characteristics between genetic groups, even within the same breed. This is mainly because the M NW breed is larger when compared to the M NR . Therefore, it is inferred that the crossing between the genetic groups of the M N breed is another alternative to increase the carcass weight resulting in greater cuts, especially noble ones (rib, loin and breast), but with similar cut yields indicating proportionality between carcass and non-carcass components within the two genetic groups. This result may be a peculiar characteristic of using the M N breed in crossbreedings, since Landim et al. The direct relationship between S T and C WL ( Table 2) is associated with the carcass quality during the nal post-slaughter process, since the cover fat has the main objective to act as a thermal insulator in cooling the carcass, thus providing less loss of weight during the maturation process Landim et al., 2021). However, S T was not su cient to maintain thermal insulation for the evaluated genetic groups, exceeding the extremity considered adequate (C WL < 7 %; Oliveira et al., 2017).
The discrimination in only one canonical discriminating function (Table 3) for the nine selected groups is explained because the genetic groups (white and red) belong to the same breed, which reduced the discriminatory variance shared between the genetic groups. Even the cut yields do not have an isolated capacity to discriminate genetic groups, which is justi ed by the similarity between yields ( Table 2). The shank yield showed discriminatory power between the genetic groups when all variables were simultaneously analyzed. However, this characteristic acted in a complementary way to the others (T BI , C G and rib weight) to classify 87.0% of the lambs in their group of origin. It is emphasized that adopting the mix of variables is important to identify the main predictor variables for inclusion in genetic breeding programs for native breeds, especially small breeds due to the lack of research. Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js The higher classi cation rates observed when the zoometric indices and body morphometry were partially analyzed are justi ed by the higher values in the characteristics related to the thoracic part and the height of the crossbred lambs (Tables 1 and 2), which infer in the e ciency response from the productive point of view, i.e. the ability to gain weight and intake dry matter, which justi es the higher performance of these animals. In contrast, the lowest rates were observed in carcass performance and indexes, as the effect of crossbreeding between genetic groups did not in uence most carcass variables (Melo et al., 2018;Ramos et al., 2019). The discrimination of the thoracic perimeter when evaluating the carcass morphometry is explained because this characteristic is a predictor of live weight at slaughter (Castro et al., 2012), for which both variables were higher in M N F 1 lambs.
In this study, the potential for meat production was assessed by factor analysis, in which the characteristics linearly associated with each other were distributed through their respective loads, making it possible to compose factors with biological and economic interpretation. The rst factor grouped a larger number of variables including those which indirectly affect the nal carcass characteristics, and therefore acted in the economic scope of meat production. This factor also pointed out that C WL is the single factor which most compromises the economic index, since losses due to cooling interfere with the carcass weight and quality, and consequently reduce its price. For (Landim et al., 2017;, an alternative to increase C WL is the use of the Rabo Largo breed in terminal crossbreeding. The C EI inferred greater explanatory power due to the four factors extracted, in addition to having a high correlation with the A I , which is considered an important indicator of meat production (Sabbioni et al., 2016). Therefore, if there is selection in the M N breed for this morphological index, there will indirectly be a positive in uence on the economic aspect. Considering the variables of greatest importance and association within the rst factor, it is observed that it is possible to carry out sheep selection through measurements and zoometric indices, indirectly resulting in the carcass traits, which in principle would only be analyzed after slaughter.
The second factor associated characteristics which exert great in uence on the productive potential of the M N breed. It is possible to predetermine the digestive capacity and weight gain of a given animal using T PRI , while C Y can be used in sheep farming in response to the relative percentage of meat to be marketed, as lambs generally have high yields and small variations between biotypes (Oliveira et al., 2018).
The factor associated with commercial characterization, which in turn is directly related to the animal's production capacity, and consequently greater body and carcass weight, has a direct in uence with LegY, as this variable has the greatest quantitative representation of a sheep carcass (McManus et al.,2008;Trindade et al., 2018). The chest height indicates the stature of the animal, and it is recommended to raise small sheep for the Caatinga biome since the availability of food during the dry period of the year is mostly located in herbaceous extract . In addition, animals with a compact pro le and small to medium body structure should be chosen if the breeder intends to raise animals for meat production with a high capacity for weight gain (Ramos et al., 2019).
The grouping plotted on the heat map clari ed the speci cities for meat production of the genetic groups. M N F 1 lambs presented superiority in the productive, commercial and economic aspects, while the M NR lambs were balanced. Moreover, Issakowicz et al. (2016) report that the lower meat production potential for the M NR breed when compared to other native sheep breeds is related to the small size of this breed group; however, morpho-productive characterizations are useful approaches to circumvent this limitation, as they identify the main characteristics related to the productive potential and use them as a selection criterion. The cluster formed by the morpho-economic and morpho-commercial factors is justi ed by their direct relationship between factors, while the balance was related to the morpho-productive factor, especially with emphasis on the crossbred lambs due to their greater production when compared to the purebred lambs.
The S MMI suggests that sheep obtained by crossbreeding the white and red genetic groups of the M N breed have a greater potential for meat production when compared with M NR lambs, demonstrating the importance of the white genetic group participating in terminal crossbreeding. Finally, this work identi ed predictive morphological parameters for meat production with the creation of an aggregate index obtained from the variables with greater participation in the multivariate analysis. This index presents itself as an early and practical selection criterion for meat production in native lambs, and can be used in breeding programs for sheep and by farmers, since biometrics is a simple and easy tool to perform. It is also expected to contribute to the recovery process of the MN white breed variety, as its use has been shown to be favorable in sheep meat production.

Conclusions
M N V and M N F 1 lambs present high thoracic development, compact character and aptitude for meat production. Traits such as T BI , Leg Y., C G and rib weight were generally the most important for differentiating the genetic groups of the M N breed. The successive adoption of multivariate techniques demonstrated greater potential for meat production for M N F 1 lambs, with this crossbreeding being another alternative to increase sustainable sheep production in the Brazilian semi-arid region. It is recommended that M MI be used as a tool for morpho-productive selection criteria in M N breeding programs, as well as for other small native breeds. Resultados apresentados para esse grupo de variável é mediana -intervalo interquartil (IQR)