# Protein requirements of woolly sheep reared in hot climates: a multifaceted approach

### Ethical considerations

Animal ethics committee approval was not necessary in this study because data were collected from previously published sources.

### Model proposal

Experiments conducted only with woolly sheep or crosses raised in tropical regions of Brazil provided individual data of the following quantitative data: BW, EBW, average daily gain (ADG), EBW gain (EBWG), total digestible food intake (TDNI), crude protein intake ( CPI) and body protein (BPC) and fat (BFC) content. Studies included information on individual animals fed at least two levels above and at maintenance levels based on comparative slaughter methodology.

The database consisted of 11 experimental studies (Nascimento Junior29; Silva et al.30; Pereira31; Costa and others.32; Regadas Filho and others.9; Oliveira et al.33; Rodrigues et al.34; Pereira and others.35; Pereira and others.7; Pereira and others.23and Mendes et al.36), consists of 382 animals in total. Of these, 74 animals belonged to the reference group, 308 to the experimental group, two gender classes: intact (n = 269) and aborted (n = 113) males. Dietary crude protein (CP) and metabolizable energy (ME) ranged from 47–236 g/kg dry matter (DM) and 0.9–3.4 Mcal/kg DM, respectively; the main feeding system was forage (Table 3). Nascimento Junior29 The study was not included to estimate protein requirements for maintenance due to lack of intake data (TDNI and CPI).

### Dissection, chemical analysis and body composition

All studies used comparative slaughter methodology. After slaughter, body components were analyzed for DM content (AOAC37; method 930.15); oil content was determined by ether extract (EE) using a Soxhlet apparatus for 12 hours (AOAC).37; method 920.39) and CP (AOAC37; method 984.13). In general, intake, digestibility measurements and estimates of ME intake, stored energy (RE) and stored protein (RP) were similar across studies, and details can be obtained directly from the original publications. EBW, BPC, and BFC of reference animals slaughtered at the beginning of the experiments were used to estimate baseline EBW, BPC, and BFC of experimental animals individually. The body energy (BEC) of the animals of each study was calculated using the equation recommended by the ARC12:

$$mathrm{BEC}=left(mathrm{BPC}times 5.6405right)+(mathrm{BFC}times 9.3929)$$

(5)

where BEC is body energy content (Mcal/day), BPC is body protein content (kg), BFC is body fat content (kg). RP and RE were estimated by the difference between the final BPC and BFC and the baseline BPC and BFC of each study, respectively. Descriptive statistics of the variables used to fit the models are shown in Table 4.

### BW and body gain adjustments

Fasted body weight, lean body weight, and lean body weight gain were calculated according to the equations recommended by Herbster et al.38:

$$mathrm{FBW}=-0.5470+0.9313times mathrm{BW}$$

(6)

$$mathrm{EBW}=-1.4944+ 0.8816times mathrm{FBW}$$

(7)

$$mathrm{EBWG}=0.906times mathrm{ADG}$$

(8)

where BW is body weight (kg), FBW is estimated fasting body weight (kg), EBW is estimated lean body weight (kg), EBWG is estimated lean body weight gain (kg/day), ADG is average daily gain (kg/day). Factors of 1.23 (BW/EBW) and 1.21 (FBW/EBW) were used to convert requirements expressed in g/kg EBW to g/kg BW and g/kg FBW, respectively.

### Metabolic protein intake

MCP synthesis was estimated using the equation recommended by Santos et al.39:

$$mathrm{MCP}=12.7311+59.2956times mathrm{TDNI}$$

(9)

where MCP is estimated microbial crude protein synthesis (g/day) and TDNI is total digestible food intake (kg/day) calculated for each study. From the rear, rumen degradable protein (RDP) was considered equivalent to MCP. The following equation was used to estimate the actual digestible microbial crude protein:

$$mathrm{tdMCP}=mathrm{RDP}times 0.64$$

(10)

where tdMCP is the actual digestible microbial crude protein (g/day), RDP is the estimated rumen degradable protein (g/day), a value of 0.64 given that MCP is composed of 80% amino acids with an intestinal digestibility of 80. %21. RUP intake was calculated as the difference between CP intake and RDP. Thus, digestible rumen indigestible protein was obtained from the following equation:

$$mathrm{dRUP }=mathrm{ RUP }times 0.80$$

(11)

where 0.80 refers to 80% digestibility of RUP in the small intestine21. Thus, metabolizable protein intake (MPI) was calculated as the sum of tdMCP and dRUP.

### Metabolizable protein requirements for maintenance

Metabolizable protein requirement for maintenance (MPm, g/kg0.75 EBW/day) was estimated based on an adaptation of the equations given by Wilkerson et al.18 and NRC21. First, a linear regression of MPI against EBWG of animals was constructed:

$$mathrm{MPI}={upbeta }_{0}+{upbeta }_{1}times mathrm{EBWG}$$

(12)

where MPI is metabolizable protein intake (g/day), EBWG is lean body weight gain (kg/day), and β0 and β1 are linear regression coefficients. Cross section at the back (β0) value of the adjusted model was divided by the total mean metabolic EBW of the animals, and this result was taken as MPm (g/kg).0.75 EBW/day):

$$mathrm{MPm}= frac{{upbeta }_{0}}{{mathrm{EBW}}^{0.75}}$$

(13)

### Net protein requirements for maintenance

To estimate the net protein requirement for maintenance (NPm, g/kg0.75 EBW/day), a linear regression of RP against MPI was fitted according to the following equation:

$$mathrm{RP}={upbeta }_{0}+{upbeta }_{1}times mathrm{MPI}$$

(14)

where RP is stored protein (g/kg0.75 EBW/day), MPI is metabolizable protein intake (g/kg0.75 EBW/day), β0 NPm and β were taken as1 was the effectiveness of using metabolizable protein for weight gain (kp). The efficiency of metabolizable protein utilization for maintenance (kevening) was obtained as NPm/MPm.

### Net protein requirements for weight gain

A regression was fitted between weight gain requirement (NPg), EBWG and RE versus RP to estimate net protein. This method assumes that animal performance and body gain composition are related to the proportion of energy stored in the gain.21:

$$mathrm{NPg}={upbeta }_{0}+{upbeta }_{1}times mathrm{EBWG}+{upbeta }_{2}times mathrm{RE}$$

(15)

where NPg is net protein requirement for weight gain (g/day), EBWG is lean body weight gain (kg/day), RE is stored energy (Mcal/day), β0β1 and β2 are linear regression coefficients.

### Statistical analysis

In this study, a linear mixed model was used to estimate and test the parameters and effects. Because the dataset consisted of different individual studies, we used a meta-analysis approach that included the study effect as a random effect.40. The inclusion of the study effect was also tested for each model slope and intercept. The fixed effect of gender classes on model parameters was tested, and when differences were significant (P< 0.05), a unique equation was used for all gender classes. Normality and dispersion of residuals were checked, and we considered records with Studentized residuals greater than 2.5 and/or Cook's distance greater than 1 as influential points.41,42,43,44. In this study, we first tested three covariance structures using unstructured covariance and no convergence and/or covariance significance (P< 0.05), variance components (VC) and complex symmetry (CS) structures were tested and selected based on the corrected AIC value. Statistical analysis was performed using the MIXED and NLMIXED procedures of SAS (SAS Institute Inc) for linear mixed and nonlinear mixed models, respectively.