Classifying different Iberian pig genetic lines by applying chemical–instrumental parameters of dry-cured Iberian shoulders

• Published in: Journal of food science and technology 2018-11, Vol.55 (11), p.4589-4599
• Main Authors: Caballero, D., Asensio, M., Fernández, C., Martín, N., Silva, A.
• Format: Article
• Language: English
• Subjects: Animal models; Chemistry; Chemistry and Materials Science; Chemistry/Food Science; Classification; computer software; Data mining; decision support systems; Decision trees; fatty acid composition; Fatty acids; Food Science; fruits; genetic lines; herds; Iberian (swine breed); Iberian Peninsula; livestock and meat industry; Mathematical models; Meat; meat quality; Morphology; Nutrition; Organic chemistry; Original; Original Article; Parameters; Pasture; pastures; Shoulder; shoulders; Studies; Suidae; Swine; Training; Volatile compounds
• ISSN: 0022-1155; 0975-8402
• DOI: 10.1007/s13197-018-3396-3

Iberian pigs are an autochthonous porcine breed mainly bred in southwestern of Iberian Peninsula, this herd traditionally feeds on acorn and pasture, in an extensive production system. In this study, the main objective was for classifying Iberian shoulders from different genetic lines of pigs and their crossed genetic lines. For that, morphology parameters, fatty acid profile and volatile compounds were determined on two batches of dry-cured Iberian shoulders. The samples were acquired from the same place of the dry-cured Iberian shoulders. From these data, two databases were created, for training and validation purpose. Results on this study, firstly demonstrate the capability of data mining techniques in order to classify shoulder as function as Iberian pig genetic lines by using of different quality characteristics. Four classification models were developed and tested in the training database and all models achieved a percentage of correct classification higher than 75%. Then, the same four classification models were performed in the validation database and the model of J48 decision tree and fatty acid profile reached the highest correct classification percentage (91.67%), and from Retinto genetic line (95.65%). From this classification model, a software application was developed in order to determine in a semi-automatic way the genetic line of the dry-cured Iberian shoulder. This application could be used in the meat industries for determining relationships between genetic line and meat quality or performance of noble parts.