Use of the canonical discriminant analysis to select a reduced pool of single nucleotide polymorphism markers for sheep breed assignment and traceability purposes

Several market research studies have shown that consumers are concerned with the provenance of the food they eat. Then, the development of methodologies able to trace animal derived foodstuffs is become of primary interest. DNA-based technologies seem to be the most useful tools to completely prevent consumers from frauds. Moreover, to satisfy the increasing number of consumers demanding for edible products of certified origin, the development of a breed DNA traceability and authentication methods might represent an important marketing strategy. In this study, a new statistical method to discriminate among 21 different Italian sheep breeds and assign new individuals to groups was applied. For each breed, 22 animals were genotyped using the Illumina OvineSNP50 BeadChip: 20 animals were designed as training population (TP), 2 as validation population (VP). For each breed, monomorphic markers and SNP with MAF<0.15 were discharged and, at the end of the data editing, the retained markers were 26,450. The canonical discriminant analysis (CDA) was used to discriminate among breeds. CDA was first applied at chromosome level and markers whose canonical coefficients were higher than 0.27 were retained. A genome-wide CDA was then developed with only the selected markers to test the separation among breeds. Finally, the discriminant analysis (DA) technique was used to assign the VP to the proper breed. The by-chromosome CDA selected a total of 265 top discriminant SNPs and the genome-wide CDA gave a significant separation among all groups. The DA was able to assign 42 over 46 independent individuals to the proper breed. Only for four breeds, the assignment percentage was 50%. These results suggested that an assay with the selected 265 SNP could be used to routinely track monobreed products.