Principal component analysis and cluster analysis for fatty acid assessment of backfat in three pig breeds

Irina Chernukha; Marina Nikitina; Elena Kotenkova; Liliya Fedulova

doi:10.18485/meattech.2023.64.2.60

Irina Chernukha V. M. Gorbatov Federal Research Center for Food Systems of RAS
Marina Nikitina V. M. Gorbatov Federal Research Center for Food Systems of RAS
Elena Kotenkova V. M. Gorbatov Federal Research Center for Food Systems of RAS
Liliya Fedulova V. M. Gorbatov Federal Research Center for Food Systems of RAS

DOI: https://doi.org/10.18485/meattech.2023.64.2.60

Keywords: pig breed, Duroc, Altai, Livni, cluster analysis, PCA

Abstract

Local pig breeds are bred by crossing pig breeds, and their carcass traits, technological, chemical and physical compositions differ. The fatty acid composition and content depend on both biological and handling specifity that can be assessed using statistical tools. Backfat composition of three pig breeds was studied: two were local (Livni and Altai), one was commercial (Duroc). The calculations were performed based on the results of fatty acids analysis using small datasets in the software R Studio 2022.07.2 Build 576. Principal component and cluster analysis revealed that the fatty acid composition of the two local pig breeds differed from each other and from the commercial breed, which resulted in clustering of the individual pigs studied into separate groups. Results of cluster analysis demonstrated similarity with PCA, but all breeds formed distinct groups; Altai pigs formed the most tight and distinct group, while Duroc and Livni pigs clustered in more expanded groups with overlapping regions. Non-parametric statistics (PCA and cluster analysis) were applicable to this small dataset.