Identification of genomic regions related to lipid and cholesterol content in beef

Identificación de las regiones genómicas relacionadas con el contenido de lípidos y colesterol en bovinos

F.M. Rezende,^I G.A. Oliveira Junior,^II M.E. Carvalho,^II R.V. Ventura,^II-III J.B.S. Ferraz,^II J.P. Eler,^II

^IFederal University of Uberlandia, Uberlandia, MG, Brazil.
^IIResearch Center of Animal Breeding, Biotechnology and Transgenesis of the University of Sao Paulo/FZEA Brazil. ]]> ^IIIBeef Improvement Opportunities, Guelph, Ontario, Canada.

ABSTRACT

The aim of this study was to identify genomic regions that potentially have association with lipid and cholesterol content in Nellore cattle meat. Phenotypes of 615 Nellore bulls were obtained according to the methods described by Bligh and Dyer (1959) and Saldanha et al. (2004). Data of 658 genotyped Nellore bulls were used for association studies through the ssGWAS method. Those animals were genotyped with the Illumina Bovine HD® Beadchip (777 962 SNP) or Gene Seek GGPi (74,153 SNP). Based on another Nellore population genotyped for Illumina Bovine HD® Beadchip, genotypes were determined using the FImpute software. After quality control (MAF <2% and call rate <90%), 535 824 SNP in autosomal chromosomes were used in the association analyses. Single step analyses were performed using a pedigree composed by 4065 animals by the BLUPf90 program considering windows of 10 markers to estimate their effects. This procedure enables the identification of regions associated with lipid and cholesterol content by chromosome. Results of this research showed regions on chromosomes 5, 10, 12, 23 and 29 related to lipid content and on chromosomes 3, 10, 11, 12 13, 17 and 18 associated with cholesterol deposition.

Key words: GWAS, Nellore, meat quality, SNP.

RESUMEN

El objetivo de este estudio fue identificar las regiones genómicas que tienen asociación potencial con el contenido de lípidos y colesterol de la carne de ganado vacuno Nellore. Se obtuvieron fenotipos de 615 toros Nellore, de acuerdo con los métodos descritos por Bligh y Dyer (1959) y Saldanhaet al. (2004). Los datos de 658 toros Nellore genotipados se utilizaron para estudios de asociación a través del método ssGWAS. Estos animales se genotiparon con Illumina Bovine beadchip HD® (777 962 SNP) o GeneSeek GGPi (74 153 SNP).Sobre la base de otra población de Nellore genotipada por Illumina beadchip BovineHD®, los genotipos se determinaron con el software FImpute. Después del control de calidad (MAF <2% y call rate <90%), se utilizaron 535 824 SNP en cromosomas autosomales para el análisis de asociación. Los análisis de paso simple se realizaron utilizando un pedigrí compuesto por 4065 animales con el programa BLUPf90, considerando ventanas de 10 indicadores para estimar sus efectos. Este procedimiento posibilita la identificación de regiones asociadas al contenido de colesterol y lípidos por cromosoma. Los resultados de esta investigación demostraron regiones en los cromosomas 5, 10, 12, 23 y 29 relacionadas con el contenido de lípidos, y regiones en los cromosomas 3, 10, 11, 12 13, 17 y 18 relacionadas a la deposición de colesterol.

Palabras clave: GWAS, Nellore, calidad de la carne, SNP.

INTRODUCTION

The localization of genomic regions related to the expression of quantitative traits allows identifying genes and their pathways that provides a better comprehension of their genetic control. Lipid and cholesterol contents are becoming major concerns to consumers because of excessive consumption of high density calorie food has harmful effects on human health mainly on increasing cardiovascular diseases. Therefore, the aim of this study was to research on genomic association methods that allow the use of pedigree and genomic information for the identification of chromosomal regions that potentially have association with lipid and cholesterol content in Nellore beef.

MATERIALS AND METHODS

Animals. Nellore young bulls raised under pasture conditions until 18 months of age and, after, fed in feedlots, were slaughtered between 21 and 27 months of age in six different dates, what makes them standardized. All animals are progenies of bulls selected for production and reproduction traits.

Phenotypic traits. Longissimus dorsi sample was collected, individually identified and, after 7 days of ageing, kept frozen at -18˚C till the analyses. The determination of total lipids was based on methodology described by Bligh and Dyer (1959). Cholesterol extraction and quantification was made according to method described by Saldanha et al. (2004).

Animal genotyping. Animals were genotyped with Illumina Bovine Beadchip HD® (777962 SNP, N = 449 animals) or Illumina Gene Seek SNP Beadchips Bovine GGP-HDi (74153 SNP, N = 209 animals) technologies, according to the manufacturer protocols. After animal genotyping, the data from the lower density panel were imputed using the software FImpute, for panel Illumina Bovine HD®. The accuracy of imputation was determined by cross validation for each animal, and the concordance rate between the imputed and the real genotype was higher than 97.51%.The quality control of markers SNP was made, excluding those with unknown genomic position, placed in sexual chromosomes, with minor allele frequency (MAF) of 2%, markers that presented call rate lower than 90% and markers with heterozygous genotype excess. After quality control 658 animals samples and 535824 SNP remained to be analyzed.

Association analysis. Analyses were performed using a pedigree composed by 4,065 animals. Single step analyses were realized by BLUPf90 program considering windows of 10 SNP to estimate their effects, this procedure enables the identification of regions associated with lipid and cholesterol content along the chromosomes. The animal model considered as fixed effects slaughter group and analysis date, besides, as covariates, slaughter age and backfat and, as random effects, animal and residual effects. The variance components and genetic parameters were estimated using the Bayesian inference (Gianola and Fernando 1986), considering a linear animal model, and the GIBBS2F90 and ssGWAS computer programs were used (Misztal et al. 2002 and Aguilar et al. 2011). Chains with size of 100 000 interactions, with a burn in criteria of the first 1000 interactions were generated. For the estimation of parameters, the samples were stored every 100 cycles, forming samples with 990 data. The data convergence was verified with the graphic evaluation of the values sampled versus interaction the according to Heidelberger and Welch (1983) and Raftery and Lewis (1992) by package of analysis Bayesian Output Analysis (BOA) of software R 2.9.0 (The R Development Core Team, 2009).

Genomic region prospecting. The gene identification was realized with BioMart from Ensembl Genome Browser tool. It was possible to make the identification of gene association analyses for lipid and cholesterol content.

RESULTS AND DISCUSSION

The residual and the additive variance and heritability were 0.06, 0.34, 0.17 and 5.54, 31.98, 0.17 for lipid and cholesterol content, respectively. In table 1 were presented statistics description of analyzed data.

Figures 1 and 2 contain genomic regions and the percentage of variance explained by chromosome by windows of 10 SNP adjacent for lipid and cholesterol content, respectively.

A total of 7 and 10 associated genomic regions, distributed in 5 and 7 different chromosomes were obtained for lipid and cholesterol content, respectively. In these regions, 7 and 9 genes were identified for each trait in this order (tables 2 and 3).

With ssGWAS method, using a high density panel, it was possible to identify regions related to lipid and cholesterol content in Nellore beef. Posteriorly, those genes and their pathways will be researched to evaluate their then importance for meat quality traits.

REFERENCES

Aguilar, I., Misztal, I., Legarra, A. & Tsuruta, S. 2011. “Efficient computation of the genomic relationship matrix and other matrices used in single-step evaluation”. Journal of Animal Breeding and Genetics, 128 (6): 422–428.

Bligh, E. G. & Dyer, W. J. 1959. “A rapid method of total lipid extraction and purification”. Canadian journal of biochemistry and physiology, 37 (8): 911–917.

Heidelberger, P. & Welch, P. D. 1983. “Simulation run length control in the presence of an initial transient”. Operations Research, 31 (6): 1109–1144.

Misztal, I., Tsuruta, S., Strabebl, T., Auvray, B., Drueta, T. & Lee, D. H. 2002. “Proceedings of the 7th World Congress on Genetics Applied to Livestock Production”. In: vol. 28, Montpellier: INRA, p. 21.

Raftery, A. E. & Lewis, S. M. 1992. “Comment on ‘The Gibbs sampler and Markov chain Monte Carlo’”. Statistical Science, 7: 499–501.

R Development Core Teamn.d. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, ISBN: 3-900051-07-0, Available: <http://www.R-project.org> .

Saldanha, T., Mazalli, M. R. & Bragagnolo, N. 2004. “Avaliação comparativa entre dois métodos para determinação do colesterol em carnes e leite”. Ciência e Tecnologia de Alimentos, 24 (1): 109–113.

WTSI & EMBL-EBI. 2016. Ensembl genome browser 84. , Available: <http://www.ensembl.org/index.html>, [Consulted: May 5, 2016].

Received: November 26, 2015
Accepted: January 6, 2016

F.M. Rezende, Federal University of Uberlandia, Uberlandia, MG, Brazil. Email: frezende@ingeb.ufu.br