Next generation sequencing as a genomic used tool to detect species substitution and mislabelling in meat products sold in South Africa

Abstract

Meat species adulteration is a subject that has brought a lot of controversy to the meat industry and has led to meat adulteration studies being conducted worldwide. Correct labelling of meat products promotes fair trade and ensure that consumers make well informed decisions when choosing the meat they want to consume. Unfortunately, the unabated increase in the cost of resources has resulted in an increase in meat species substitution and/or mislabeling in the global meat industry. Deoxyribonucleic acid-based methods have previously been used in the past for species identification, however, they are limited to a few targeted species. Next Generation Sequencing is a universal technology that can be used to identify meat species. The main objective of this study was to evaluate genomics and bioinformatics pipelines that will facilitate utilization of genetic markers and Next Generation Sequencing (NGS) technology to identify and discriminate meat species in the South African meat industry. The first study investigated the discriminatory potential of 16S rRNA, 12S rRNA, COX3 and ATP6 mitochondrial genes in mammalian species identification and differentiation. To achieve this, a phylogenetic analysis was conducted using the entire database of 262 mammalian species for each of the above genes downloaded from Genbank (http://www.ncbi.nlm.nih.gov/nuccore). All four genes managed to separate distantly related species and group closely related species with a common ancestry, however, the ATP6 and COX3 genes grouped some species that were not closely related together. Overall, the 16S rRNA gene performed the best with bootstrap values of 97 – 100% in all clades observed whilst the ATP6 gene performed the least. There was no improvement in performance when the 12S rRNA, COX3 and ATP6 genes were individually combined with the 16S rRNA gene. There were high bootstrap values of 100% observed in most clade groupings when all four genes were combined. Consequently, the study recommends the use of the 16S rRNA gene on its own for species identification, as it performed well in comparison to the 12S rRNA, COX3 and ATP6 genes. The use of four different genes in a species identification experiment will be more expensive and time consuming albeit yielding relatively higher bootstrap values. The second experiment developed a molecular and bioinformatics diagnostic pipeline that utilizes the mitochondrial 16S ribosomal RNA (rRNA) barcoding gene, to determine processed meat product mislabelling through Next Generation Sequencing. Pure meat samples were artificially mixed at different ratios, to verify the sensitivity and specificity of the pipeline. Processed meat samples (n = 155) namely, minced meat (n = 49), biltong (n = 28), burger patties (n = 35), and sausages (n = 43) were collected from across South Africa and sequenced using the Illumina MiSeq sequencing platform. All the species used in the artificially mixed pure samples were identified, confirming the specificity and sensitivity of the pipeline. Processed meat samples had reads that mostly mapped to the Bos (90% and above) genus, with traces of the Ovis and Sus (2 – 5%) genus. This confirmed that the majority of the processed meat samples were from beef. Amongst all processed meat samples, sausages had the highest level of contamination, with 46% of the samples having mixtures of beef, pork or mutton in one sample, which was in contrast to the labelling, as the only labelling provided was of samples labelled as beef sausages. The pipeline further demonstrated its specificity by identifying species with percentages as low as 0,1% in both the artificially mixed pure samples and processed meat samples. Overall, the developed pipeline can be used with confidence to authenticate meat products and furthermore, investigate and manage any form of mislabelling in the meat industry. The third and last experiment investigated the presence of breed-specific Single Nucleotide Polymorphisms (SNPs) using the entire mitochondrial genome of 13 European and Indigenous cattle breeds reared in South Africa for use in breed assignment and traceability. Whole genome sequencing was performed on 13 European and Indigenous cattle breeds reared in South Africa. A total of 42 animals were used from Afrikaner (n = 4), Beefmaster (n = 4), Boran (n = 4) Charolais (n = 2), Hereford (n = 2), Nguni (n = 2), Simbra (n = 3), Bonsmara (n = 4), Brahman (n = 4), Drakensberger (n = 4), Limousin (n = 2), Santa (n = 3) and Simmentaler (n = 4) breeds. Whole genome sequencing was performed using the Illumina HiSeq 2500 (Illumina, San Diego, CA, United States) at 10X coverage. A total of 12 996 variants were identified and of these 12 633 were SNPs and 363 were Indels. The highest number of variants were identified in the European Brahman breed (n = 2 066) and the lowest in the Indigenous Nguni breed (n = 340). The SNPs were also divided into homozygous and heterozygous SNPs. The highest number of homozygous SNPs were found in the Limousin breed (n = 534) and the highest number of heterozygous SNPs were found in the Brahman breed (n = 1 872). To identify breed-specific SNPs we used all homozygous SNPs identified that have the same alleles. A total of 125 breed-specific SNPs were identified in all breeds except for the Charolais breed that did not contain any breed-specific SNPs. The Limousin breed had the highest number of breed specific SNPs (n = 59) and the lowest were found in the Nguni breed (n = 1). The COX3 mitochondrial gene had the highest number of breed specific SNPs (n = 22), followed by the 16S mitochondrial gene (n = 19). Nineteen of the breed-specific SNPs were shared amongst breeds and the ND5 gene contained the highest number of shared SNPs. This study provides an insight on the presence of SNPs within the mitochondrial genome of cattle breeds reared in South Africa. The breed specific SNPs identified provided an understanding of the regions within mitochondrial genes that are unique in each breed and can be used in the authentication of beef meat in the meat industry.

In conclusion the study illustrated that NGS can be used as a genomic tool to detect meat species mislabelling in meat products in South Africa and worldwide. Mitochondrial sequences were used to determine the discriminatory potential of four genes (16S rRNA, 12S rRNA, COX3 and ATP6) in mammalian species. The 16S rRNA gene demonstrated the highest discriminatory potential amongst the four genes. Using the 16S rRNA gene, NGS technology was further used to identify meat species in both artificially mixed pure and processed meat samples. The NGS technology proved that it can be used as a universal tool in meat species identification. Finally, using NGS cattle breeds reared in South Africa were sequenced and the sequences were used to identify breed specific SNPs in the mitochondrial genome of the cattle breeds. The mitochondrial genes that contained the most breed-specific SNPs were within the ND5, COX3 and 16S rRNA genes.