Genome-wide Association Study for Rectal Temperature in Chinese Holstein population

H.Luo1, W. Xu1, A. Liu1, X. Li2, L. Liu3, Y. Wang1

1College of Animal Science and Technology, China Agricultural University, 100193, Beijing China 2Beijing Sunlon Livestock Development Company Limited, 100029, Beijing China 3Beijing dairy cattle center,100192, Beijing China [email protected] (Hanpeng Luo)

Summary

Heat stress causes a considerable economic loss due to unfavorable correlation with production, reproduction, and immunity in dairy cattle whilst rectal temperature (RT) is widely used as an indicator for response to heat stress. This study was designed with objective to identify SNPs and candidate for RT during heat stress in lactating Holstein cows using Genome-Wide Association Study (GWAS). The whole genomic single nucleotide polymorphisms (SNPs) were obtained from the IlluminaBovineSNP150 BeadChip and genotyped in a population of 1,114 Holstein cows. Genetic parameters and estimated breeding values (EBVs) were derived for morning RT (AMRT) and afternoon RT (PMRT), moreover the EBVs of AMRT was used to perform GWAS by using the Fixed and random model Circuitous Probability Unification (Farm CPU) method. In total, eight SNPs were identified to be significantly associated with AMRT and five candidate genes were discovered to be located near to them. The overall findings concluded the FAM107B and PHRF1 as strong candidate genes.

Keywords: Chinese Holstein, rectal temperature, genome-wide association study

Introduction

Heat stress in dairy cattle causes extensive economic losses due to its unfavorable correlation with production, reproduction, and immunity (Armstrong et al., 1994). Most of the negative effects of heat stress on animal performance are upshot of either physiological adjustment to maintain body temperature or unfavorable consequences of failure to regulate body temperature. Thus, selection for regulation of body temperature during heat stress could increase thermo-tolerance (Hansen, 2011), whereas rectal temperature (RT) rises significantly with higher environmental temperature and widely used as an indicator of heat stress in dairy cattle. Breeding animals with genetic resistance to heat stress is considered as one of the sustainable strategy, and revealing the genetic basis of RT become essential. During early time, the RT during heat stress was already found to be a heritable trait in dairy cattle, with an estimated range from 0.15 to 0.31 (Seath,1947). Recent research reported that the SLICK (Liu et al., 2011) and ATP1A1 (Olson et al., 2003) were coupled with both RT and milk yield. With the development of a high-density single nucleotide polymorphism (SNP) for the panel, genome-wide association study (GWAS) has become a useful tool for mapping markers and quantitative trait loci (QTL). In dairy cattle, many SNPs with high linkage disequilibrium with causative mutations have been detected for various traits (e.g. DGAT1 for milk fat). However, to our knowledge, GWAS have barely been performed for heat stress or traits related to heat stress in dairy cattle. Thus keeping in view the above cited studies, the current research was to carry out a GWAS for RT during heat stress in lactating dairy cows to identify SNPs and candidate genes for RT in Chinese Holstein.

Materials and methods

Data

Data of 6,446 lactating Holstein cows during 2013 to 2016 were collected from 9 herds in the Sanyuan Lvhe Dairy Cattle Center, Beijing, China and all herds were equipped with fans and sprinklers. Full pedigree was provided by the Dairy Association of China (Beijing, China), including 20,545 females and 2,194 males born from 1934 to 2014, moreover, the RT of milking cows was measured in the morning during 06:30-11:00 am (AMRT) and in the afternoon during 13:00-18:30 pm (PMRT) for two consecutive days. Environmental temperature (Tdb) was measured before and after the measurement of RT for each herd using a digital thermometer placed at a height of 1.5~3 m from the floor and single test-day records with the closest date to RT measurements were collected to match RT.

Estimated breeding values

Estimated breeding values (EBVs) were used as dependent variables in GWAS. EBVs of cows for AMRT and PMRT were estimated using a binary-trait animal model including herd by testing year, parity, stage of lactation and milking status as fixed effects, and Test-day milk yield and average environmental temperature during the measurement of each herd as covariates, and additive genetic effects, permanent environmental effects, and residual effects were included as random effects. Parities had 6 levels, namely: 1, 2, 3, 4, 5, and 6~10 parity. Stage of lactation had 5 levels based on days in milk (d): 1~50, 51~100,101~200,201~ 300 and 301~600. Milking status had 3 levels: RT measured before milking, after milking, and unknown. Average information restricted maximum likelihood method was applied to estimate (co)variance components using the DMU Package (Madsen et al., 2011). Summary statistics of EBVs for AMRT and PMRT for 1,114 genotyped cows are shown in Table 1. Since a high genetic correlation (0.97) was observed between AMRT and PMRT, thus, GWAS was performed using EBVs for AMRT only.

Table 1. Summary statistics of EBVs for AMRT and PMRT. Trait Mean Std Minimum Maximum AMRT1 -0.0078 0.00731 -0.2123 0.2494 PMRT2 -0.0065 0.0574 -0.1668 0.1927 1AMRT: morning rectal temperature 2PMRT: afternoon rectal temperature

Genotypic data

A total of 1,114 cows were genotyped with Illumina Bovine SNP150 BeadChip (Illumina, Inc., San Diego, CA, USA). Imputation was performed in Chinese Holsteins by BEAGLE software (Browning et al., 2009). Only SNPs met following criteria were used for GWAS: 1) minor allele frequency greater than 5%; 2). The P-value for Hardy-Weinberg equilibrium less than 1 × 10-6, and 3) chromosome and position were known. Ultimately, 116,579 SNPs of 1,114 cows were available for GWAS.

Association analysis

The Fixed and random model Circuitous Probability Unification (FarmCPU) model (Liu X et al., 2016) was used to perform the single-SNP analysis for the markers. Subsequently, Bonferroni-corrected threshold probability of 0.05/N was implemented to verify the significance levels for the GWAS results, where N was the number of trait-SNP tested. Quantile-quantile (Q-Q) plots and Manhattan plots were created by R 3.3.3 software.

Results and discussion

GWAS

Manhattan plot and Q-Q plots for AMRT are presented in Figure 1. Q-Q plots revealed that the population stratification was well controlled. A total of 8 genome-wide significant SNPs were detected for AMRT (P< 4.29×10-7). The 8 significant SNPs (Table 2) were located on Bos taurus autosome (BTA) 3, 8, 13, 14, 17 and 29, respectively. Those significant SNPs’ MAF ranged from 0.17 (ARS-BFGL-BAC-26939) to 0.46 (BovineHD1700005860) and P- value ranged from 2.18×10-8 (ARS-BFGL-NGS-86864) to 3.47×10-7(BovineHD1700005860). Candidate genes were searched by taking a 200 Kb window surrounding the 8 significant SNPs. Gene identification information were quoted from National Center for Biotechnology Information gene integrates information database (http://www.ncbi.nlm.nih.gov/gene/) or The Human Gene Database (http://www.genecards.org/).

Identification and annotation of candidate genes

The 8 significant SNPs found in the current study were entirely different from the results of relevant literature (Dikmen et al., 2013). The significant SNP (BovineHD0300007938) on BTA3 located in SPAG17 gene (sperm associated antigen 17), which encodes a central pair protein present in the axonemes of cells with a "9+2" organization of and the encoded protein is required for the proper function of the axoneme. A recent study showed that the SPAG17 gene was associated with growth in mice (Treves et al., 2015). The searching region of significant SNP (ARS-BFGL-NGS-86864) on BTA13 contained the FAM107B gene (family with sequence similarity 107 member B ). The FAM107 members contain an N-terminal domain which is conserved across species and regarded as heat shock- inducible tumor small protein which carried the promoter region providing heat shock (Nakajima et al., 2010). The significant SNPs (ARS-BFGL-NGS-70821) and (ARS-BFGL-BAC-26939) on BTA14 located in TSNARE1 (t-SNARE domain containing 1) gene and RALYL ( RALY RNA binding protein-like) gene respectively. The gene TSNARE1 may have a vertebrate-specific function in intracellular protein transport and synaptic vesicle exocytosis and a recent cited literature suggested its development within the vertebrate lineage from the harbinger transposon super family (Smithet al., 2012). The gene RALYL is related to nucleic acid binding and identical protein binding. An important paralog of this gene RALYL is RALY that encodes a member of the heterogeneous nuclear ribonucleoprotein gene family and responsible for pre-mRNA splicing and in embryonic development (Khrebtukova et al., 1999). The significant SNP (BovineHD2900014853) on BTA29 located in the PHRF1 gene (PHD and ring finger domains 1). The gene PHRF1 could promote the genome integrity by modulating non-homologous end-joining upon DNA damage insults (Chang et al., 2015), so it may repair cell damage caused by heat shock. Thus, basing on biological function, gene FAM107B and PHRF1 were more plausible related to RT.

Figure 1. Results of the genome-wide association studies. The strengths of genome-wide association studies (GWAS) are illustrated by the Manhattan plots on the right panel. The deviations of the signals from the null hypothesis are illustrated as the Quantile-Quantile (QQ) plots on the right panel. The negative logarithms of the observed (y-axis) and the expected (x-axis) P values are plotted for each SNP.

Table 2. Significant SNPs identified for AMRT in Chinese Holstein and related genes within 200 Kb of the significant SNPs.

Arguable candidate genes Start and SNP BTA Position (bp) Gene Gene name end position BovineHD0300007938 3 25184273 SPAG17 sperm associated 25111111 - antigen 17 25367379 BovineHD0800010566 8 35584460 BovineHD1300001265 13 4896494 ARS-BFGL-NGS-86864 13 29486605 FAM107B family with 29499895 - sequence similarity 29638719 107 member B ARS-BFGL-NGS-70821 14 3078843 TSNARE1 t-SNARE domain 3054703 - containing 1 3171547 ARS-BFGL-BAC-26939 14 80862874 RALYL RALY RNA 80029572 - binding protein like 80866821 BovineHD1700005860 17 20271465 BovineHD2900014853 29 50922990 PHRF1 PHD and ring 50907382 - finger domains 1 50930300 Note: nearest gene are symbols of gene full name in the NCBI database (http://www.ncbi.nlm.nih.gov/)

Conclusions

The present genome-wide association study identified eight significant SNPs as well as targeted five candidate genes associated with rectal temperature trait in Chinese Holstein cattle population. The two genes FAM107B and PHRF1 were strongly recommended for further functioned analysis, especially in relation to response to heat stress in dairy cattle.

List of References

Armstrong, D. V., 1994. Heat stress interaction with shade and cooling. Journal of Dairy Science, 77(7): 2044-2050. Browning, B. L., & Browning, S. R., 2009. A unified approach to genotype imputation and haplotype-phase inference for large data sets of trios and unrelated individuals. American Journal of Human Genetics, 84(2): 210. Chang, C. F., Chu, P. C., Wu, P. Y., Yu, M. Y., J-Y, L., & M-D, T., et al., 2015. PHRF1 promotes genome integrity by modulating non-homologous end-joining. Cell Death & Disease, 6(4): e1716. Dikmen, S., Cole, J. B., Null, D. J., & Hansen, P. J., 2013. Genome-wide association mapping for identification of quantitative trait loci for rectal temperature during heat stress in Holstein cattle. PloS One, 8(7): e69202. Hansen , P.J., 2011. Heat stress and climate change. In: Moo-Young M, (ed.), Comprehensive Biotechnology, Second Edition, vol. 4, 477–485. Amsterdam: Elsevier. Khrebtukova, I., Kuklin, A., Woychik, R. P., & Michaud, E. J., 1999. Alternative processing of the human and mouse raly genes. Biochimica Et Biophysica Acta, 1447(1): 107-112. Liu, X., Huang, M., Fan, B., Buckler, E. S., & Zhang, Z., 2016. Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PloS Genetics, 12(2): e1005767. Liu, Y., Li, D., Li, H., Zhou, X., & Wang, G., 2011. A novel SNP of the ATP1A1, gene is associated with heat tolerance traits in dairy cows. Molecular Biology Reports, 38(1): 83- 8. Madsen P, Sørensen P, Su G, et al. DMU - a package for analyzing multivariate mixed models.[C]// Proceedings of the 8th World Congress on Genetics Applied to Livestock Production, Belo Horizonte, Minas Gerais, Brazil, 13-18 August, 2006. 2014:27-11. Nakajima, H., Ishigaki, Y., Xia, Q. S., Ikeda, T., Yoshitake, Y., & Yonekura, H., et al., 2010. Induction of HITS, a newly identified family with sequence similarity 107 protein (FAM107B), in cancer cells by heat shock stimulation. International Journal of Oncology, 37(3): 583-93. Olson, T. A., Lucena, C., Jr, C. C., & Hammond, A. C., 2003. Evidence of a major gene influencing hair length and heat tolerance in Bos taurus cattle. Journal of Animal Science, 81(1): 80. Seath D. M., 1947. Heritability of heat tolerance in dairy cattle. Journal of Dairy Science, 30(3):137-144. Smith, J. J., Sumiyama, K., & Amemiya, C. T., 2012. A living fossil in the genome of a living fossil: harbinger transposons in the coelacanth genome. Molecular Biology & Evolution, 29(3): 985. Teves, M. E., Sundaresan, G., Cohen, D. J., Hyzy, S. L., Kajan, I., & Maczis, M., et al., 2015. Spag17 deficiency results in skeletal malformations and bone abnormalities. PloS One, 10(5): e0125936.