Factor and Discriminant Analyses in the Morphostructure of Batur and Wonosobo Sheep Breeds

Batur sheep and Wonosobo sheep are two Indonesian indigenous sheep breeds from Central Java Province, Indonesia. Batur sheep resulted from a crossbreed between Merino sheep and Thin-Tailed sheep that originated and spread widely in Batur District, Banjarnegara Regency. Wonosobo sheep is a new breed developed from crossing the local sheep and Texel sheep and spread widely in the Wonosobo Regency. Batur and Wonosobo sheep have been cultivated, maintained and raised by smallholder farmers for generations and are suitable for meat production (Ibrahim et al., 2021; Noviani et al., 2013; Hakim et al., 2019).
       
The characterization of livestock is the first approach to sustainable use of its animal genetic resources. Body dimensions have been used to indicate the breed, origin and relationship or shape and size of an individual as they give an idea of body conformation. Breed characterization in goats and sheep can be carried out based on body measurements and body indices (Markovic et al., 2019; Putra and Ilham, 2019). The morphometric characterization of livestock is essential for planning improvement, sustainable utilization, conservation strategies and breeding programs for a breed (FAO, 2012).   
       
The Batur and Wonosobo sheep were still considered non-descript sheep in the country and many things of morphostructures have not been studied. While some studies reported that the Batur and Wonosobo have similar ancestors, namely Thin-Tailed sheep (Hakim et al., 2019; Ibrahim et al., 2020b; Prayitno et al., 2008). Genetically, the Batur, Wonosobo and Thin-Tailed sheep have a close genetic relationship (Ibrahim et al., 2020b). Certainly, with similar ancestors and relatively close genetic distance, it is expected that the sheep provide performance and productivity so close different. Therefore, the present study was undertaken to describe and discriminate the morphostructure of Batur and Wonosobo sheep with body measurements and body indices.
 
 

Research site and animals
 
The samples were collected from 2019 to 2020 in Wonosobo and Banjarnegara Regencies, Central Java Province, Indonesia. However, this study was conducted in July-October 2021 in the Laboratory of Meat, Drought and Companion Animals, Universitas Gadjah Mada. The body measurements of Wonosobo sheep (22 rams and 61 ewes) and Batur sheep (13 rams and 24 ewes) were collected from smallholder farmers in Wonosobo and Banjarnegara Regencies, respectively. All animals in this study have two permanent pairs of incisors conditions with an average of 1.5 years of age (non-pregnancy stage). The phenotypic characteristic of Wonosobo sheep and Batur sheep and their breeding site were illustrated in Fig 1.
 
Animal’s measurements
 
The body measurements were taken from animals in a standing position with a raised head. Body measurements of animals were performed using a measuring stick and flexible measuring tape and taken based on FAO (2012) Guidelines. Fourteen body measurements of body length (BL), withers height (WH), rump height (RH), rump width (RW), chest circumference (CC), chest depth (CD), shoulder-width (SW), head length (HL), head width (HW), ear length (EL), ear width (EW), tail length (TL), tail width (TW) and tail circumference (TC) were measured in each animal as described in Fig 2. Moreover, nine body indices of length index (LI), thoracic index (TI), depth index (DI), height index (HI), thoracic development (TD), conformation index (CI), body index (BI), proportionality (Pr) and area index (AI) were computed in this study with mathematical formula (Boujenane, 2015; Khargharia et al., 2015) as follows:
 
Data analysis
 
The statistics parameter of mean, standard deviation and Pearson’s coefficient of correlation (r) for body measurements and body indices were calculated with SPSS 25.0 computer program (IBM, USA). Therefore, factor analysis [Principal component analysis (PCA)] and discriminant analysis [Canonical discriminant analysis (CDA)] were computed using a similar computer program.
 

Animal morphostructure and Pearson’s coefficient of correlation
 
Nine body measurements of RW, CC, HL, HE, EL, EW, TL, TW and TC in Batur sheep were higher than in Wonosobo sheep (P<0.05) as presented in Table 1. Two body indices of HI and BI in Wonosobo sheep were higher than in Batur sheep (P<0.05), as presented in Table 2. Therefore, body indices of CI in Wonosobo sheep were lower than in Batur sheep. The Pearson’s coefficient of correlation (r) value between WH and RH in both sheep breeds was highest than the other correlations (0.97 in Wonosobo sheep and 0.93 in Batur sheep), as presented in Table 3. In addition, the highest r value among body indices showed in Pr-AI (-0.99) for both sheep breeds, as presented in Table 4.








       
The body measurements in the present study were higher than the previous study reported by Haryanti et al. (2015) and Haren et al. (2018) in Wonosobo and Batur sheep, respectively. According to the BI value, Wonosobo and Batur sheep included brevigline (BI<85) animals type similar to Ripollesa sheep (83.90±6.10) (Esquivelzeta et al., 2011). Handiwirawan et al. (2011) obtained LI value of 109.00±2.00 in Barbados Black Belly Cross (BC) sheep and lower than Wonosobo and Batur sheep. Therefore, the DI value in BC sheep was 52.0±0.9 and close to Batur sheep. The r-value of WH-RH in Uda (0.90) sheep (Yakubu and Akinyemi, 2010) was close to the present study.
 
Factor analysis
 
The factor analysis (PCA) of body measurements reveals three principal components (PCs) that explain 76.58% and 73.51% of the total variance in Wonosobo and Batur sheep morphostructures, respectively (Table 5). The PCA of body indices in sheep study reveals three PCs that explain 89.57% (Wonosobo) and 79.16% (Batur) of the total variance in sheep morphostructure (Table 6). Previous studies were revealed PCA of body measurements with a total variance of sheep morphostructure of about 59.00% (2PCs) in Ripollesa (Esquivelzeta et al., 2011); 61.53% (4PCs) in Rampur-Bushair (Sankhyan et al., 2018) and 96.65% (3PCs) in Pramenka (Markovic et al., 2019) sheep. Unfortunately, a study of the PCA of sheep’s body indices is limited. Previous studies reported that the PCA of body indices accounted for 86.84% (4 PCs) in Katjang does and 89.38% (4 PCs) in Pasundan cows (Putra et al., 2020; Putra and Ilham, 2019), showing higher than in sheep study.




 
Discriminant analysis
 
The discriminant analysis (CDA) reveals six body measurements (RW, SW, EL, EW, RH and TL) and two body indices (TI and CI) that were selected as the discriminant variables (Table 7). According to the CDA, about 95.2% (Wonosobo) and 97.3% (Batur) of sheep can be classified into their original breeds group based on body measurements (Table 8). Meanwhile, about 80.7% (Wonosobo) and 75.7% (Batur) of sheep’s study can be classified into their original breeds based on body indices. The discriminant plots of body measurements and body indices were illustrated in Fig 3 and 4, respectively.








       
The CDA showed that body measurements were more accurate for discriminating Wonosobo and Batur sheep than body indices. Ofori et al. (2021) obtained the rc value of 0.62 (high) in the first canonical factor of body measurement in West African Dwarf goats and lower than in the present study. Thus, (Asamoah-Boaheng and Sam, 2016) obtained the rc value of 0.89 (very high) in the first canonical factor of body measurements in sheep and similar to the present study. More than 90% of the morphostructure of sheep study were successfully classified into their original breeds group with six selected discriminant variables of body measurements. A previous study reported that the body measurements could classify many sheep breeds into their original breeds group (Asamoah-Boaheng and Sam, 2016). Moreover, 58% of Uda rams and 70.8% of Uda ewes can be classified into their original sex group based on two discriminant variables of HW and HL measurements (Yakubu and Akinyemi, 2010). The difference in body measurements, genetics (breed), location (geographical area), agro-climatic conditions and management system of animals can be caused by different results compared to previous studies.
 
 

In conclusion, about 95.2% of Wonosobo sheep and 92.7% of Batur sheep can be classified with body measurements (RW, SW, EL, EW, RH and TL). Meanwhile, about 80.7% of Wonosobo sheep and 75.7% of Batur sheep can be classified with body indices (TI and CI). Wonosobo sheep and Batur sheep can be accurately characterized by their body measurements.
 

The authors are thankful to the Direktorat Penelitian dan Tim Peningkatan Reputasi UGM menuju World Class University-Kantor Jaminan Mutu UGM, for funding this study with the Post-Doctoral Program (grant no. 6144/UN1.P.III/DIT-LIT/PT/2021). The authors also thank the Faculty of Animal Science Universitas Gadjah Mada, the National Research and Innovation Agency, the Department of Agriculture, Fisheries and Animal Husbandry of Banjarnegara Regency and the Department of Food, Agriculture and Fisheries of Wonosobo Regency for the support and thank all farmers.
 

None.