Sheep and goat meat are important sources of dietary proteins with high biological value and nutrients (Webb et al., 2005; Ponnampalam et al., 2015). The global production of sheep and goat meat increased from 15 million tons to 16.2 million tons between 2015 and 2019 (FAOSTAT, 2021). Moreover, rearing these animals (largely composed of traditional breeds) has a socioeconomic role in less developed regions by providing occupation, income, and food security for local populations (Lobo, 2019; Morales-Jerrett et al., 2020). Among the many factors involved in meat production, animal feed composition is considered a crucial component to promote the proper animal development and finishing as well as the improvements in meat quality (Knapik et al., 2017). Specifically, meat quality comprises the nutritional value, physicochemical (such as color, water holding capacity, and shear force), and sensory properties that, collectively, govern the suitability of meat for consumption (Elmasry et al., 2012).

Animal diet has also gained another dimension due to the concerns about current practices to produce crops (Salami et al., 2019a). The use of edible crops (food-feed competition) that also have an environmental impact (land-use, soil erosion, and climate change, for instance) is a major component in meat production cost (Peacock & Sherman, 2010; Salami et al., 2019a). Consequently, the search for alternative sources of nutrients for meat animals has become an important aspect to improve the current commercial meat production systems (Salami et al., 2019a). In this sense, the concept of circular economy gained the spotlight. Circularly economy consist in the strategic utilization of resources, reduction of waste generation, and also reutilization of co-products of production chains, which fits in the current trend of sustainable development (Hamam et al., 2021).

Sunflower cake/biomass is the co-product obtained from cold pressing of sunflower seeds for biodiesel production and is considered a rich source of nutrients with high contents of protein, fiber, and minerals (29-61%, 4-45%, and 6-9%, respectively) (Brazil et al., 2019; Subaşı et al., 2021). In this context, the incorporation of sunflower cake in the diet of ruminant has been studied in the last decade. For instance, the use of increasing levels of sunflower cake (up to 30%) in Santa Ines lambs improved the fat content and the fatty acid profile of Longissimus lumborum muscle (Lima et al., 2018). Additionally, no significant effects were reported for chemical composition and sensory attributes. Conversely, a similar experiment with sunflower (8, 16 and 24%) in the diet of Boer goats revealed no significant effect in the pH or chemical composition of Longissimus lumborum muscle whereas tenderness was reduced (Oliveira et al., 2015).

Considering the utilization potential of sunflower cake in animal feed, this study aimed to evaluate the meat quality (proximate composition and physicochemical characteristics in Longissimus lumborum muscle; and pH (24 h) and sensory analysis in Semimembranosus muscle) of crossbred lambs (Santa Inês × Dorper) fed with increasing levels of sunflower cake.

The experimental diets had no effect on the chemical composition or the physicochemical parameters of the meat (Table 4). Moreover, the orthogonal contrast analysis (0% vs. 5, 10, and 15% addition of sunflower cake into sheep diet) for these data did not indicate significant differences, except for moisture content. Sensory evaluation of lamb meat revealed that selected descriptors were not affected by increasing levels of sunflower cake (Table 5). The average values found for hardness, juiciness, flavor, aroma, and color were 2.42, 3.90, 4.28, 4.11, and 3.96, respectively.

In order to evaluate the degree of association among the physicochemical composition and the sensorial attributes of the meat of lambs fed different levels of sunflower cake, a Principal Component Analysis (PCA) was carried out. As a result, it is observed that the first two components explained 95.59% of the total variability of meat quality characteristics, 69.71% explained by the first component and 25.88% by the second one (Fig. 1). Protein was the most relevant variable (0.331) in principal component one (PC1) and lipid was the most important factor (0.478) in principal component two (PC2).

Four main groups of variables can be observed in Fig. 1. The group 1 is composed of moisture content and CL; the group 2 is formed protein, hardness, and ash content; the group 3 composed of lipid content, SF, and WHC; the group 4 has juiciness, color, odor, and flavor. Fig. 1 shows that sensory analysis results were mainly distributed in PC1 wherein group 2 (protein, texture, and ash) and 4 (juiciness, color, odor and flavor) were positioned in opposite locations. A similar interpretation can be observed between moisture and lipid content.

Our results from chemical composition analysis of Longissimus lumborum are in agreement with other studies found in the literature for Santa Inês and their crossbreeds fed with licuri (Syagrus coronata) and crambe (Crambe abyssinica hochst) cakes, whose percentage values varied between 64 and 76% for moisture; from 14 to 27% for protein; between 1.5 and 7% for lipids; and from 1.0 to 1.2% for ashes (Issakowicz et al., 2017; Costa et al., 2018; Bagaldo et al., 2019). According to these studies, the incorporation of cakes from vegetable oil processing indicated a variable effect in terms of chemical composition of meat. The use of licuri cake (up to 260 g/kg feed) did not cause a significant effect in the composition of lamb meat. It is also relevant to mention that lipid and protein contents varied from 19.7 to 36.3 and from 200 to 189 g/kg feed, respectively, as the level of licuri cake increased in the diet (Bagaldo et al., 2019). Conversely, another study with licuri cake in lamb diet indicated significant increase in moisture content in meat whereas the content of lipids and proteins were reduced due to partial replacing of soybean meal and ground corn by licuri cake (8, 16 and 24% feed in DM) (Costa et al., 2018). In that experiment, increasing levels of licuri cake in the diet resulted in lipid content from 2.6 to 5.1% and protein content around 16%.

The difference in these studies seem to relay in the variations of lipid and protein content in the diet of lambs. High lipid level with reduced protein content has been suggested as the relevant factor that modify the progression of fermentation in the rumen and favor the deposition of fat in body and meat (Bagaldo et al., 2019). Considering this aspect, the results for the chemical composition obtained in our study may be explained by the relatively small variations in lipid and protein contents in diet (Table 2). Moreover, the orthogonal contrast analysis did not indicate significant differences for chemical composition (except for moisture) or physicochemical parameters (Table 4). Although moisture content varied between the meat of animals receiving control and supplemented diet, this effect was not observed in other parameters, indicating a marginal variation derived from diet composition.

A suitable acidification rate in muscles during the 24 hours post-slaughter period leads to pH values between 5.5 and 5.7 in lamb meat (Chauhan et al., 2019). In this experiment, the muscle presented average values higher than this range. This difference may be attributed to pre-slaughter factors such as stress, feeding, and environmental temperature that reduces the glycogen levels in the animal, which limits the pH decay in the post mortem period (Ferguson & Warner, 2008). It is also relevant to mention that our pH values are within the range reported in scientific literature (from 5.68 to 5.96) wherein lambs were fed with licuri cake (Costa et al., 2018), camelina (Camelina sativa) meal (Ramírez et al., 2018), and cardoon (Cynara cardunculus var. altilis) (Salami et al., 2019b).

The effect of agro-food residues in lamb meat varies across studies (L*, a* and b* in the ranges of 35.7-40.00, 14.85-24.4, and 3.12-11.4) (Bezerra et al., 2016; Costa et al., 2018; Bagaldo et al., 2019; Salami et al., 2019b). Similar to our study, Bezerra et al. (2016) did not observe significant differences in color of crossbreed (Dorper × Santa Inês) lambs fed with increasing levels of peanut cake (replacing up to 100% of soybean meal). Likewise, Salami et al. (2019a), explored the effect of cardoon meal (15% in diet) in Sarda × Comisana lambs meat and reported no significant changes in color. Conversely, the use of licuri cake in lamb feed was associated with significant changes in color (especially redness) of meat (Costa et al., 2018; Bagaldo et al., 2019). For consumers, lamb meats with brightness values below 34 and redness values below 9.5 for a* value are considered dark and unacceptable (Khliji et al., 2010).

Although non-significant differences were obtained for SF among treatments, it is important to note that SF values between 5.0 and 7.9 can be considered tender whereas meat with SF values in the ranges of 8.0–10.9 and above 11 can be classified as acceptable and thought, respectively (Bickerstaffe et al., 2001). Previous studies indicate different results about the effect of sunflower cake and seeds in SF of Longissimus muscle of goats and lambs. For instance, a significant reduction SF values (Longissimus muscle) was reported by Xazela et al. (2012), who compared the effect of two diets (Medicago sativa vs. sunflower cake) in the production of autochthonous goats (Boer and Xhosa × Boer cross) of South Africa. However, these authors also indicated similar SF for Xhosa lop eared and Nguni breeds fed with either Medicago sativa or Sunflower cake diet. Moreover, a related experiment showed that including 8% of sunflower seeds in diet of Ile de France lamb did not affect the SF in Longissimus lumborum muscle (Sales et al., 2013).

The WHC is directly linked to the final quality of the meat because the greater the capacity to retain water, the better the juiciness and the lower the weight loss and CL; which can also be associated to the yield in the preparation for consumption (Warner, 2017). The values of WHC in the present study were close to that reported by Bezerra et al. (2016), who reported non-significant differences (values in the range 63.41-59.86%) among the increasing levels of peanut cake as soybean meal replacer. Similarly, Sales et al. (2013) obtained a WHC value of 62.9% in Ile de France lambs fed diets containing sunflower seeds and vitamin E. These authors observed a significant increase in WHC due to sunflower seed supplementation. Additionally, high WHC values were also associated with the retention of water-soluble nutrients in the meat, since these components could be leaked in the exudate. Thus, high values of WHC are indicative of less shrinkage and more soluble myofibrilar proteins (Huff-Lonergan & Lonergan, 2005).

The CL values of lamb meat found in this study were in the range of values found for studies exploring the effect of agro-food residues in lamb feed. In agreement with our results, the inclusion of sunflower seeds (8%) in the diet of Ile de France lambs did not affect the CL of Longissimus lumborum muscle (Sales et al., 2013). Other similar studies with agro-food residues indicate a similar outcome. For example, CL mean values between 23.57 and 29.08% were reported for the meat of Dorper × Santa Inês lamb fed with peanut cake (Bezerra et al., 2016), from 25.5 and 29.0% for the incorporation of licuri cake into feed of Santa Inês lamb (Costa et al., 2018), and values in the range of 28.1-31.6% Boer × indigenous lambs with increasing levels of palm kernel cake (7, 14, and 21%) in diet (Ribeiro et al., 2018).

Considering the sensory evaluation data, it was possible to classify the meat as slight hard; with medium to reduced juiciness, flavor, and aroma (all the treatments received scores around the middle of the scale). It is relevant to remember that lamb meat stands out among other meats due to its characteristic and intense odor and flavor. In a similar way to observed in the present study, Ribeiro et al. (2018) reported non-significant differences in sensory analysis of lambs fed with palm kernel cake for aroma, flavor, tenderness, and juiciness. A related experiment with Barbarine lambs fed with olive cake did not indicated significant effects in flavor and tenderness whereas significant differences were reported for juiciness (Hamdi et al., 2016).

PCA revealed interesting outcomes regarding the characteristics of meat from animals fed with different levels of sunflower cake. The negative correlation between juiciness and hardness (evaluated by PCA) in lamb and hogget meat was reported in a previous study with animals from two breeds (Norwegian White Sheep and Norwegian Spel) (Bhatti et al., 2020). It is also relevant to mention that this negative correlation between the juiciness and tenderness of meats may occur across different species (Rødbotten et al., 2004). Regarding the negative association between lipid and moisture contents in meat samples from animals with increasing levels of sunflower cake, a recent experiment reported a similar outcome due to the addition of guava processing waste in the feed of lambs (Nobre et al., 2020). Increasing levels of guava (Psidium guajava L.) residue in animal feeding were associated with reduced moisture content and increased levels of fat in meat.

Some negative correlations between sensory and chemical composition were also observed in the PCA analysis. Considering the role of proteolysis in the formation of flavor and aroma in meat, a negative correlation between protein content and sensory attributes can be considered. This relation may be explained, at least in part, by a previous study proteomic of meat from Santa Inês × Dorper lambs feed with increasing levels of sunflower cake (5, 10, and 15%) (de Melo et al., 2020). That experiment revealed that the contents of most structural proteins in meat (such as myosin light chain 2, troponin C, tropomyosin beta chain, and actin) were gradually reduced (suggesting intensification of proteolysis) as the level of sunflower cake increased in the diet of animals.

In this context, the inverse relation between protein content and juiciness could be explained by the advance of proteolysis that could affect proteins structure and lead to a reduced capacity to retain water (Pearce et al., 2011). Another consideration related to the progression of proteolysis may explain the inverse relation with the flavor and odor of meat and protein. It is known that proteolysis is a key process for the formation of odor– and flavor–active compounds in meat (Khan et al., 2015).

In summary, sunflower cake when included in feed for lambs (up to 15%) does not interfere with the chemical composition or physicochemical characteristics of Longissimus lumborum muscle. Neither pH (24 h) nor sensory characteristics of Semimembranosus muscle were affected by the increasing levels of sunflower cake. Therefore, sunflower cake can be used as an alternative source of nutrients in diet of Santa Inês × Dorper sheep with minimal impact in meat quality. Further experiments with a large number of animals are necessary to expand our understanding about the effect of sunflower cake in the meat of Santa Inês × Dorper sheep.