- Background: In livestock, residual variance has been studied because of the interest to improve uniformity of production. - Several studies have provided evidence that residual variance is partially under genetic control;. - The aim of this study was to identify genomic regions associated with within-family residual variance of yearling weight (YW. - For this, solutions from double hierarchical generalized linear models (DHGLM) were used to provide the response variables, as follows: a DGHLM assuming non-null genetic correlation between mean and residual variance (r mv ≠ 0) to obtain deregressed EBV for mean (dEBV m ) and residual variance (dEBV v. - Results: The dEBV m and dEBV v were highly correlated, resulting in common regions associated with mean and residual variance of YW. - More independent association results between mean and residual variance were obtained when null r mv was assumed. - Full list of author information is available at the end of the article. - 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0. - Such a phenomenon is called genetic heterogeneity of residual variance or genetic variance in micro-environmental sensitivity. - Up to now, several studies support the existence of a genetic component on residual variance and draw attention for its potential to improve uniformity through selection (e.g. - Unraveling the genetic basis of heterogeneity of residual variance through genome-wide association studies (GWAS) will help to understand the biology be- hind it and increase selection response by identifying candidate genes affecting uniformity and including them in genomic prediction.. - [7] and birth weight in pigs by Wang et al. - Residual variance per individual from double hierarchical generalized linear model (DHGLM;. - obtained according to Rönnegård et al. - [10]) were used as response variable by Mulder et al. - [11] to identify gen- omic regions related to residual variance of somatic cell score in dairy cattle. - Estimated breeding values (EBV) from a DHGLM, using an extension developed by Felleki et al. - [12], were deregressed (dEBV) and used to identify genomic regions associated with variability of litter size in pigs by Sell-Kubiak et al. - The choice for dEBV in the study by Sell-Kubiak et al. - In the case of Mulder et al. - In addition, the existence of a genetic component on residual variance of growth traits was previously observed [17–19]. - Thus, the aim of this study was to identify genomic re- gions associated with within-family residual variance of yearling weight (YW) in Nellore cattle through GWAS, using different response variables, and candidate genes to better understand the biology behind genetic control of uniformity.. - Our aim was to identify genomic regions (1-Mb win- dows among the top 20 that explained the largest pro- portion of genetic variance shared between response variables and 1-Mb windows with SNPs that showed a strong association by BF) associated with within-family residual variance of YW in Nellore cattle, using different response variables in GWAS. - For this, we used solutions from DHGLM assuming: i) non-null genetic correlation between mean and residual variance of YW (r mv ≠ 0) to obtain deregressed EBV for mean (dEBV m ) and residual variance (dEBV v. - As a result, eight out of the top 20 windows were shared between dEBV m and dEBV v : chromosome (Chr) 1 (92 Mb. - dEBV m and dEBV v : deregressed EBV for mean and residual variance of yearling weight, respectively. - dEBV v_r0. - and ln_ σ 2 ^ e : deregressed EBV for residual variance and log-transformed variance of estimated residuals, respectively, both assuming null genetic correlation between mean and residual variance. - The response variables for residual variance of YW had low to moderate correlations with each other (Table 1). - Box plots of dEBV v and ln_ σ ^ 2 e by genotype of the markers with higher BF were presented to give an overview about the signals that these SNPs are capturing (Fig. - In both cases, the first homozygous genotype (0) was more uni- form, with lower mean and lower dispersion of the cor- responding response variable, compared to the other genotypes, although AA was less frequent in relation to AB and BB. - The top 20 windows explained together and 20.0% of the genetic variance of dEBV m , dEBV v, dEBV v_r0 and ln_ σ 2 ^ e (data not shown).. - The range of the proportion of variance explained by individual windows and their sum suggest that uniform- ity of YW behaves as a polygenic trait determined by several genes, as well as its mean.. - Such findings highlight that most of the effects captured by dEBV v can be due to the strong r mv and could be potentially scale effects [19].. - However, all SNPs showed a greater standardized effect on dEBV v than dEBV m indicating that these re- gions affect the residual variance beyond a simple scale effect. - Table 2 Common 1-Mb windows shared by deregressed EBV for mean (dEBV m ) and residual variance (dEBV v. - For dEBV v , most of the genes in the regions harboring SNPs that showed strong association were re- lated to metabolism: DPYD (Chr3), BTG1 (Chr5), AK7 (Chr21), ADIRF (Chr28), GLUD1 (Chr28), IP6K1, GMPPB, APEH, AMT, USP4, USP19, IMPDH2, NDU- FAF3, SLC25A20, PRKAR2A, UQCRC1, PFKFB4 and SHISA5 all located on Chr22 (Table 5). - In summary, most of the regions associ- ated with residual variance of YW harbor interesting candidate genes related to metabolism, stress, inflamma- tory and immune responses, mineralization, neuronal activity and bone formation.. - dEBV v_r0 and ln_ σ 2 ^ e : deregressed EBV for residual variance and log-transformed variance of estimated residuals, respectively, both obtained from solutions of a DHGLM assuming null genetic correlation between mean and residual variance. - In this study, we identified genomic regions associated with within-family residual variance of YW in Nellore cattle by using different response variables (dEBV v , dEBV v_r0 and ln_. - Overlapping windows among the top 20 that explained most of the gen- etic variance were observed between dEBV v and the other response variables including dEBV m . - The dEBV v shared 8 out of the top 20 windows with dEBV m , one with dEBV v_r0. - RB1CC1 is involved in cell growth and differentiation, senescence, apoptosis and autoph- agy, and was one of the genes differentially expressed in heat stressed chickens [29, 30]. - dEBV v : deregressed EBV for residual variance of yearling weight. - dEBV v_r0 and ln_ σ 2 ^ e : deregressed EBV for residual variance and log-transformed variance of estimated residuals, respectively, both assuming null genetic correlation between mean and residual variance. - Table 5 Significant SNPs a associated with deregressed EBV for residual variance (dEBV v. - For dEBV v , most of the genes are involved in meta- bolic pathways or related processes to it, such as metab- olism of carbohydrate, energy and lipid, nucleotide and amino acid (GMPPB, UQCRC1, PFKFB4, AMT, GLUD1, AK7, IMPDH2, NDUFAF3, DPYD, SLC25A20, PRKAR2A and IP6K1). - In such situations, genes related to metabolism are required to recover homeostasis and consequently part of the energy is diverted from growth which can poten- tially affect performance and therefore explain the po- tential trade-off between energy for growth and energy for homeostasis.. - SLC18A1 acts in the final stage of the HPA-axis, transporting catechol- amines like dopamine, noradrenaline and adrenaline. - UCN2 is a member of the corticotropin-releasing hormone (CRH) family, which is a key mediator of the stress response by activating the HPA axis [51, 52].. - Previously, Coble et al. - Neuronal plasticity is the ability of the nervous system to adapt to environmental changes, which denotes another potential class of genes determin- ing uniformity.. - Therefore, to- gether with other tissues, the nervous system seems to be a potential modulator of the mechanisms underlying uniformity.. - The high correlation between dEBV m and dEBV v re- sulted in 8 common regions affecting simultaneously mean and residual variance. - Such findings are in line with Wolc et al. - [7] who found correlations ranging from 0.54 to 0.74 and a region explaining a large proportion of the genetic variance of the mean and standard devi- ation (SD) of egg weight at different ages. - The authors noted that it was not a simple scale effect, because the effect of the QTL on mean egg weight (measured be- tween 26 and 28 weeks of age) was about 4% of the mean egg weight and on SD egg weight at the same age was about 5% of the mean SD. - More recently, Sell-Kubiak et al. - The dEBV v not only captured common effects with the mean of YW but also its residual variance by sharing regions (among top 20 by variance explained) with dEBV v_r0 and ln_ σ 2 ^ e , Chr12_5 and Chr22_52, respect- ively. - Though some of the genes found on chromosome 22 and most of those in common between dEBV m and dEBV v are involved in metabolism (e.g. - The weak correlations among dEBV m and the response variables that assumed null r mv emphasize the role of the latter to find genomic regions that are specifically affecting residual variance and not the mean of YW.. - Knowing the importance of considering potential scale effects and confounding between mean and residual variance, the next step is to discuss the suitability of the different response variables to be used in a GWAS for uniformity. - Firstly, we need to take into account the nature of the trait because small genetic variance and low heritabilities are often found for uni- formity. - 20) when we analyzed the response variables for residual variance. - In this context, irrespective of the response variable, a larger sample size is required to increase the power of GWAS [87].. - However, some non-genetic effects, such as the contemporary group (CG) effect on the variance may have affected the within-family variance estimate ln_ σ ^ 2 e , while they are accounted for when we used dEBV since non-genetic ef- fects, were fitted in the DHGLM in the residual variance part of the model. - In addition, the assumptions of null and non-null r mv were assumed to measure the influence of the mean and/or potential scale effects on the residual variance. - Furthermore, considering different re- sponse variables can be beneficial, windows that appear in multiple approaches may have a higher credibility and may help in understanding the role of genes affecting the mean, the residual variance or both and whether genes controlling the variance beyond scale effects.. - Using solutions from a DHGLM that assumed null gen- etic correlation between mean and residual variance was a suitable strategy to identify genomic regions affecting uniformity less dependent on the mean. - The phenotypic data used in this study are described in Iung et al. - The DHGLM is an iterative approach that estimates simultaneously genetic parameters for the mean and residual variance.. - s s v þ e e v , where y and ψ are vectors of response variables for the mean and residual variance models (denoted with the subscript v), respectively, b and b v are vectors of fixed effects and covariates (contemporary group, linear and. - obtained following the deregression procedure proposed by Garrick et al. - For the residual variance of YW, different response variables were analyzed: i) dEBV v : deregressed EBV of the residual variance from a DHGLM assuming non-null genetic correlation between mean and residual variance (r mv ≠ 0). - ii) dEBV v_r0 : deregressed EBV of the residual variance from a DHGLM assuming r mv = 0. - The log-transformation was used to reduce mean-variance relationship and correct for non-normality of the re- sidual variances. - was intended to reduce the impact of the mean of YW may have on the EBV in the residual variance part of the model, given the strong estimate of r mv on YW previ- ously reported for this population [19]. - When using null r mv , it is expected to have a higher possibility to find regions only affecting the residual variance. - μ is the overall mean, 1 is a vector of ones, z i is the vector of genotypes of the ani- mals for the i th SNP, a i is the allele substitution effect of the i th SNP, δ i is an indicator variable set to 1 if the i th SNP has an non-zero effect on the trait and to 0 other- wise, e is the vector of random residual effects and N is the number of SNPs. - where σ 2 a is the variance of SNP effects, σ 2 e is the residual variance and R is a diagonal matrix whose elements account for heterogeneous residual variance across observations. - [88], whereas when ln_ σ ^ 2 e was the re- sponse, such elements were equal to the reciprocal of the number of progeny of each sire. - In our study, π was fixed at 0.999, which means that 0.1% of the SNPs fitted in the model, i.e. - Chains of 550,000 and 250,000 iterations, after discarding the first 150,000 and 50,000 as burn-in, were generated for analyses pertaining to the three response variables for residual variance and mean of YW,. - Additional file 1: Complete list of the genes located within 1-Mb windows harboring SNPs with strong association by trait. - Table 6 Descriptive statistics for mean and residual variance of yearling weight. - ln_ σ 2 ^ e dEBV m and dEBV v : deregressed EBV for mean and residual variance of yearling weight, respectively. - dEBV v_r0 and ln_ σ 2 ^ e : deregressed EBV for residual variance and log-transformed variance of estimated residuals, respectively, both assuming null genetic correlation between mean and residual variance N number of observations, SD standard deviation. - RC and HAM led the coordination of the study.. - No permissions were required for the use of the samples in this study.. - heterogeneity of residual variance - estimation of variance components using double hierarchical generalized linear models. - Genetic variability of residual variance of production traits in Nellore beef cattle. - Genetic and environmental heterogeneity of residual variance of weight traits in Nellore beef cattle.. - Genetic control of residual variance of yearling weight in Nellore beef cattle. - PLAG1, the main translocation target in pleomorphic adenoma of the salivary glands, is a positive regulator of IGF-II. - Guidelines for the nomenclature of the human heat shock proteins. - Molecular cloning and expression regulation of PRG-3, a new member of the plasticity-related gene family.. - Six genes expressed in bones and teeth encode the current members of the SIBLING family of proteins. - Sequence variants of the LCORL gene and its association with growth and carcass traits in Qinchuan cattle in China. - Extension of the bayesian alphabet for genomic selection. - A whole- genome assembly of the domestic cow
Xem thử không khả dụng, vui lòng xem tại trang nguồn hoặc xem
Tóm tắt