Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock

Biaty Raymond; Loic Yengo; Roy Costilla; Chris Schrooten; Aniek C Bouwman; Ben J Hayes; Roel F Veerkamp; Peter M Visscher

doi:10.1371/journal.pgen.1008780

. 2020 Sep 14;16(9):e1008780. doi: 10.1371/journal.pgen.1008780

Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock

Biaty Raymond ^1,^2,^*, Loic Yengo ³, Roy Costilla ⁴, Chris Schrooten ⁵, Aniek C Bouwman ¹, Ben J Hayes ⁴, Roel F Veerkamp ^1,^‡, Peter M Visscher ^3,^‡

Editor: Jennie Elizabeth Pryce⁶

PMCID: PMC7514049 PMID: 32925905

Abstract

Genome-Wide Association Studies (GWAS) in large human cohorts have identified thousands of loci associated with complex traits and diseases. For identifying the genes and gene-associated variants that underlie complex traits in livestock, especially where sample sizes are limiting, it may help to integrate the results of GWAS for equivalent traits in humans as prior information. In this study, we sought to investigate the usefulness of results from a GWAS on human height as prior information for identifying the genes and gene-associated variants that affect stature in cattle, using GWAS summary data on samples sizes of 700,000 and 58,265 for humans and cattle, respectively. Using Fisher’s exact test, we observed a significant proportion of cattle stature-associated genes (30/77) that are also associated with human height (odds ratio = 5.1, p = 3.1e-10). Result of randomized sampling tests showed that cattle orthologs of human height-associated genes, hereafter referred to as candidate genes (C-genes), were more enriched for cattle stature GWAS signals than random samples of genes in the cattle genome (p = 0.01). Randomly sampled SNPs within the C-genes also tend to explain more genetic variance for cattle stature (up to 13.2%) than randomly sampled SNPs within random cattle genes (p = 0.09). The most significant SNPs from a cattle GWAS for stature within the C-genes did not explain more genetic variance for cattle stature than the most significant SNPs within random cattle genes (p = 0.87). Altogether, our findings support previous studies that suggest a similarity in the genetic regulation of height across mammalian species. However, with the availability of a powerful GWAS for stature that combined data from 8 cattle breeds, prior information from human-height GWAS does not seem to provide any additional benefit with respect to the identification of genes and gene-associated variants that affect stature in cattle.

Author summary

Provided that there is a similarity in the biological control of complex traits in different species, the results of genome-wide association study for a given trait in a model species can be utilized as prior information to identify genes and gene-associated variants that affect the same trait in a non-model species. Here we report that the number of genes found to be affecting height in both humans and cattle is more than can be expected by chance, suggesting a similarity in the biology of height in both species. In addition, genes that are associated with height in humans were found to be highly enriched with variants that are strongly associated with height in cattle. Using height as a model trait, we demonstrated that across-species prior information cannot compete with a within-species prior information, with respect to the identification of genes and the gene-associated variants that affect a complex trait. Across-species prior information can be useful, however, when within-species prior information is either unavailable or poor.

Introduction

Genome-wide association studies (GWAS) have become popular and powerful tools for identifying genes involved in complex traits aetiology, and for investigating the genetic architecture of such traits in animal and plant species. In humans for example, GWAS have been used to identify thousands of risk loci for complex diseases such as heart diseases and psychiatric disorders, and have led to the discovery of potential drug targets [1, 2]. In plant and livestock species, the focus of GWAS has primarily been, to identify variants or genomic regions that underlie complex and economically important traits [3, 4]. Ultimately, the ambition, with respect to GWAS, is the identification of a subset of genomic variants that together explain the total genetic variance for complex traits.

The ability of GWAS to identify variants that are associated with complex traits depends on sample size. Large sample size is especially important for identifying rare variants with medium-sized effects and common variants with small effects. A classic example of the importance of sample size in GWAS is the trajectory of the number of loci that have been associated with human height over the past decade. That number rose from ~40 with a sample size between 13,000 to ~ 30,000 in 2008 [5–7], to 180 with a sample size of 183,727 in 2010 [8]. By 2014, the number of loci associated with human height had increased to 697 with a sample size of 253,288 [9], while the latest GWAS for human height from 2018 reported 3,290 near-independently associated loci, with a sample size of ~ 700,000 [10]. It should be noted that the increase noticed in the past five years also reflects improvements in data analysis methods [11]. It is predicted that as sample sizes continue to increase, the number of loci reliably associated with human height will continue to increase [2, 10].

While GWAS sample sizes have continued to increase in humans, the same is not true for livestock species. While there are millions of animals from livestock species that are regularly genotyped around the world, the data are rarely shared or combined for the purpose of GWAS, given that the data are commercially sensitive and proprietary. Up until recently, for example [12–15], most GWAS in livestock have been carried out with cohort level data, with samples sizes varying from a few hundreds to a few thousands [4]. While GWAS in livestock have resulted in a number of trait-associated loci, mainly those with large effect sizes, the joint effect of the discovered loci explain only small proportions of the total genetic variance for the traits of interest [16]. Given the sample-size limitation in most GWAS in livestock, it would be of scientific interest to know if the results of GWAS studies for complex traits in humans can be used as prior information to identify the genes and gene-associated variants that affect equivalent traits in livestock species.

The idea to utilize GWAS results for a given trait across species presupposes that the genes underlying the trait are conserved across species. There have been reports of genes and mutations affecting height in humans that are found to affect height in other mammalian species, see [17] for a review. In their study, Pryce et al. [18] found that variants within 55 cattle genes, whose orthologs are associated with height in humans, were significantly associated with stature in cattle. The result of Pryce et al. [18] is further evidence to suggest that the same genes may be underlying height in humans and cattle. Thus, instead of using a within-species GWAS to identify genes and gene-associated variants for complex traits as was done in [12], a GWAS results for the same trait but from a different species could be utilized. With the availability of large sample sizes for GWAS and whole genome sequence data in both humans and cattle, we hypothesised that the evidence for similarity in the biology of height in humans and cattle will become stronger, and that use of GWAS results for height in humans as prior information to identify stature genes in cattle will become feasible. The objective of this study was to test this hypothesis. We therefore investigate the usefulness of GWAS results on human height as a prior information for identifying genes and gene-associated variants that affect stature in cattle. For this, we used GWAS summary data on samples sizes of ~700,000 and 58,265 for humans and cattle, respectively. We find that prior information from human height GWAS is beneficial for identifying genes and gene-associated variants affecting stature in cattle, only when cattle prior information is not available.

Results

Orthologous genes in humans and cattle are highly correlated in size but not in LD score or gene-effect on height

From the UMD3.1 cattle genome build, we identified 13,742 genes that have orthologs in humans and passed our quality control (QC). Within each species, we estimated the correlation between gene size and gene LD score, defined as the mean LD score of variants in the vicinity of the gene (Methods). We observed no significant correlation between gene size and gene LD scores across the 13,742 orthologous genes selected within the human genome (r = 0.0029, p = 0.74), and a significant correlation between gene size and gene LD scores within the cattle genome (r = 0.028, p = 0.001). We observed a significant positive correlation in gene-size (r = 0.85, p = 2.2x10^-16) but no significant correlation in gene LD scores (r = 0.012, p = 0.16) between human genes and their cattle orthologs. Furthermore, we estimated the correlation between the effect of orthologous genes on height in humans and cattle, as a high correlation may be an indication that the genes perform similar function across species. To do this, we estimated the aggregated effect of all the SNPs within each orthologous gene on the phenotype (human height and cattle stature), which we refer to as gene-effect. The estimate of correlation between the effect of the orthologous genes on height was low, but significant (r = 0.12, p = 2.2x10^-16).

Height-associated genes in humans and cattle significantly overlap

To address if the same genes underlie height in humans and cattle, we estimated the proportion of orthologous genes in or close to GWAS peaks for human height that are also located in or close to GWAS peaks for stature in cattle, and performed Fisher’s exact test to check if the proportion of overlap is more than can be expected by chance. We identified 77 cattle stature-associated orthologous genes based on the results of [12]. Out of the 77 cattle orthologous genes, 10 were also found to be associated with human height based on the result of [9], Fig 1A. The proportion 10/77 is more than can be observed by chance based on the results of Fisher’s exact test (odds ratio = 5.5, p = 3.7×10⁻⁵). The contingency table for the Fisher’s exact test is provided in the S1 Table. Details about the 10 overlapping genes are provided in S2 Table The overlap was stronger (30/77) between the 77 cattle genes and human height-associated genes identified based on the results of [10], Fig 1B. The human height-associated genes were identified as those that were overlapping with or nearby the near-independent significant SNPs from the meta-GWAS. The proportion 30/77 is more than expected by chance based on the result of Fisher’s exact test (odds ratio = 5.1, p-value = 3.1×10⁻¹⁰). The contingency table for the Fisher’s exact test is provided in the S3 Table, and details about the 30 overlapping genes are provided in S4 Table. To check if the average marker density in the 10kbp windows of association with lead SNPs is higher than the average marker density of genes in both species, we plotted the distribution of average marker density (total number of SNPs divided by length in kbp) of the 10kbp windows against the distribution of average marker density of genes in both species (S1 Fig). In general, we observed very similar distributions of average marker density between the windows of interest and genes. Furthermore, the result fisher’s exact test (S5 Table) showed no evidence of non-random association in marker density of genes in cattle and humans (p = 0.87), indicating that the estimates of overlap of height genes in cattle and humans is less likely to be influenced by marker density.

Fig 1 — Overlap between orthologous genes associated with height in humans based on Wood *et al*. [9] that are also associated with stature in cattle based on Bouwman *et al*. [12] (Panel A), and overlap between orthologous genes associated with height in humans based on Yengo *et al*. [10] that are also associated with stature in cattle based on Bouwman *et al*. [12] (Panel B).

It is worth noting that the 77 orthologous cattle stature-associated genes identified from [12], are among the largest in size in both the cattle and human genomes (Fig 2), which could inflate the likelihood of overlap. Therefore, we implemented a gene size-stratified Cochran-Mantel-Haenszel test to correct our estimates of overlap for potential bias. Our result show that accounting for gene-size still results in a significant overlap (p-value = 2.4×10⁻⁹). The contingency table for the Cochran-Mantel-Haenszel test is provided in S6 Table.

Fig 2 — Genes were stratified into 10 quantiles based on gene size, with the shortest genes in 1 and the longest genes in 10.

Cattle stature-associated variants are enriched in human height-associated genes

To further check if human height-associated genes also affect stature in cattle, we tested if cattle genes prioritized based on information from a GWAS on human height are more enriched for GWAS signals for cattle stature than random sets of genes from the cattle genome. For this purpose, we utilized two sets of human height-associated genes. The first set comprised of 370 orthologous genes that were identified by [10] using the summary-based Mendelian Randomization (SMR) analysis [19]. The SMR analysis combines gene-expression data from multiple tissues and summary level GWAS data to identify genes whose expression level is associated with the trait. We identified the cattle orthologs of the SMR-based human genes that we refer to as SMR candidate (SMR-C) genes. The result of randomized sampling test (Methods) showed that the number of near-independently significant SNPs for cattle stature within and around the 370 SMR-C genes are not more than can be expected within and around random sets of cattle genes matched on size, SNP density and mean LD score (Fig 3). The second set of human height-associated genes we used comprised of 1,184 orthologous genes that overlap with the 3,290 near-independently significant SNPs affecting human height identified through a conditional and joint effect (COJO) analysis in [10]. The cattle orthologs of the COJO-based human genes were identified and referred to as COJO candidate (COJO-C) genes. We found that the number of near-independent SNPs for stature in cattle within and around the 1,184 COJO-C genes is more than can be observed within and around random sets of cattle genes matched on size, SNP density and mean LD score (Fig 4).

Fig 3 — In each test, 370 cattle genes were sampled from a total of 13,865 genes that have human orthologs, and near-independent significant SNPs within and 100kbp (Column 1 –panels A,D,G), 50kbp (Column 2 –panels B,E,H) or 5kbp (Column 3 –panels C,F,I) either sides of the sampled genes were identified. The genes were sampled using stratified random sampling, with the genes stratified based on gene-size (Row 1 –panels A,B,C), number of SNPs in a gene (Row 2 –panels D,E,F) or gene LD score (Row 3 –panels G,H,I). The vertical red lines represents the number of near-independent significant SNPs (P < 5x10-8) associated with stature, identified within and around the 370 cattle (SMR-C) candidate genes that have been shown to be associated with height in humans. The p value represents the proportion of samples (out of 1,000) that have equal or higher level of enrichment for GWAS signals than the SMR-C genes.

Fig 4 — In each test, 1,184 cattle genes were sampled from a total of 13,865 genes that have human orthologs, and near-independent significant SNPs within and 100kbp (Column 1 –panels A,D,G), 50kbp (Column 2 –panels B,E,H) or 5kbp (Column 3 –panels C,F,I) either sides of the sampled genes were identified. the genes were sampled using stratified random sampling, with the genes stratified based on gene-size (Row 1 –panels A,B,C), number of SNPs in a gene (Row 2 –panels D,E,F) or gene LD score (Row 3 –panels G,H,I). The vertical red lines represents the number of near-independent significant SNPs (P < 5x10-8) associated with stature, identified within and around the 1,184 cattle (COJO-C) candidate genes that have been shown to be associated with height in humans. The p value represent the proportion of samples (out of 1,000) that have equal or higher level of enrichment for GWAS signals than the COJO-C genes.

The difference in sample size (370 vs 1,184) between the SMR-C and the COJO-C genes may result in the observed difference in the significance of enrichment for cattle stature GWAS signals. However, we found that with the same sample size of 370, the COJO-C genes are more enriched for GWAS signals than the SMR-C genes or any random set of 370 genes matched based on gene-size (Fig 5). This indicates that the results observed for the SMR-C genes is not just an issue of power (small gene sample size).

Fig 5 — In each test, 370 out of 1,184 cattle (COJO-C) genes were sampled. The number of near-independent significant SNPs within and 100kbp either side of the sampled genes were identified and compared to the number obtained within and 100kbp either side of the 370 SMR-C genes. The vertical red dashed line represents the number of near-independent significant SNPs (P < 5x10-8) associated with stature, identified within and 100kbp either side of the 370 SMR-C genes, that have been shown to be associated with height in humans through an SMR analysis. The p value represent the proportion of samples (out of 100) that have equal or lower level of enrichment for GWAS hits than the SMR-C genes.

The SMR analysis uses gene-expression data from multiple tissues. Given that not all tissues are relevant for a given trait, it is critical that gene-expression data from a trait-relevant tissue is used. Hence, we tested if the level of enrichment for GWAS signals in the SMR-C genes will be more than can be expected from random sets of genes if the selection criteria were to be more stringent. In particular, we required that the human ortholog of the gene must show evidence of differential gene expression in three or more tissues. With this selection criterion, we identified a subset of 121 (SMR-MT-C) genes instead of 370. Our result show that the level of enrichment for GWAS hits nearby the 121 SMR-MT-C genes is more than can be expected from any set of 121 randomly sampled genes, with p = 0.02 (Fig 6). The p values in the randomized sampling test were calculated as the proportion of randomly sampled gene-sets that had equal or higher level of enrichment for GWAS hits as compared to the candidate gene sets.

Fig 6 — In each test, 121 cattle genes were sampled from a total of 13,865 genes that have human orthologs. Stratified random sampling was carried out, with genes stratified based on gene-size. The number of near-independent significant SNPs within and 100k either side of the sampled genes were identified. The vertical dashed line represents the number of near-independent significant SNPs (P < 5x10-8) associated with stature, identified within and 100kbp either side of the 121 cattle (SMR-MT-C) candidate genes, whose human orthologues show evidence of differential expression in 3 or more tissues. The p value represent the proportion of samples (out of 1,000) that have equal or higher level of enrichment for GWAS hits than the SMR-MT-C genes.

Accuracy of cattle variants selection using within- and across-species prior information

Next, we quantified the proportion of stature variance explained by pre-selected cattle SNPs based on GWAS results for height in humans and stature in cattle. First, we tested the hypothesis that cattle SNPs that are pre-selected based on two information sources 1) genes identified from GWAS results for human height, and 2) Cattle stature GWAS significance level, will explain more genetic variance for stature than SNPs that are pre-selected based on only cattle stature GWAS significance level. To test this hypothesis, we identified the near-independent significant SNPs within 100kbp either side of the SMR-C and COJO-C genes, that also segregate in a New Zealand Holstein (NZH) cattle population: (80 SMR-C SNPs, 391 COJO-C SNPs). The NZH population was not included in the cattle meta-GWAS analysis [12] and is therefore ideal as a validation population. It must be noted, however, that some of the bulls in the NZH population are descendants of bulls from the Dutch Holstein population which was included in the meta-GWAS. The result of randomized sampling test showed that the identified candidate SNPs do not explain more genetic variance for stature in NZH population than the most significant SNPs for cattle stature within randomly sampled cattle genes (Fig 7).

Fig 7 — In each test, the most significant SNPs for stature within a sampled gene-set, that had at least 20 copies of the minor allele in the population were used to construct the first GRM fitted in a GREML model. A second GRM was made with randomly sampled 50k SNPs (excluding SNPs in the first GRM) across the genome, to account for population structure. Figure A shows the result of control analysis for the SMR-C SNPs, in which 80 top SNPs in each gene set were used to construct the first GRM, while figure B shows the result of control analysis for the COJO-C SNPs, in which 391 top SNPs within each gene set was used to construct the first GRM. The red vertical lines represents the percentage of genetic variance explained by the candidate SMR-C and COJO-C SNPs, respectively.

Secondly, we tested the hypothesis that cattle SNPs pre-selected using GWAS result for human height will explain more genetic variance for stature, than random set of SNPs from random cattle genes. To test the hypothesis, we randomly sampled 5 SNPs per gene to avoid having a large number of SNP in a set. For the SMR-C genes, this resulted in 1,777 unique SNPs that were segregating in NZH. For the COJO-C genes, the number of unique SNPs extracted was 5,781. Although the results are not statistically significant, we found that in ~ 90% of time, the candidate SNPs explain more genetic variance for stature in NZH population than randomly sampled SNPs from random sets of genes from the cattle genome (Fig 8).

Fig 8 — In each test, SNPs were sampled randomly (matched by allele frequency to candidate SNPs) within a sampled gene-set (5 SNPs were sampled per gene, within each gene-set). SNPs had at least 20 copies of the minor allele in the validation population. Each sampled SNP set was used to construct the first GRM. A second GRM was made with randomly sampled 50k SNPs (excluding all SNPs in the first GRM) across the genome, to account for population structure. Figure A shows the result of control analysis for the SMR-C SNPs, in which 1,777 random SNPs in each gene set were used to construct the first GRM, while figure B shows the result of control analysis for the COJO-C SNPs, in which 5,781 random SNPs within each gene set was used to construct the first GRM. The red vertical lines represents the percentage of genetic variance explained by the candidate SMR-C and COJO-C SNPs, respectively.

Discussion

A general expectation with respect to conserved orthologous genes is that they perform the same or very similar functions in different species and thus the functions of the genes in one species could be inferred to another, especially when such functions are critical for evolutionary fitness [20–22]. Several experiments in the past have reported a similarity in the function of orthologous genes in different species, although these studies focused on genes causing mendelian disorders, see [23] for a review. For a highly polygenic trait such as height, it is not yet established if the many genes underlying the trait in different species are conserved in their function, and if height-associated genes in one species can be used as prior information to identify the genes affecting height in a different species. In this study, we sought to establish if human height-associated orthologous genes, with mutations affecting height identified through GWAS, can be used as prior information to identify the genes and gene-associated variants that affect stature in cattle.

Due to LD, GWAS does not always result in the identification of the causal variants and/or genes for a given trait. In many cases, variants identified as significantly associated with a trait are proxies of unknown causal variants. One way of validating candidate genes identified from GWAS is to test if the genes are also associated with the same trait in a different species, given that LD pattern and thus the impact of LD on GWAS is different in different species. In our study, we identified a significant proportion (~39%) of cattle stature-associated orthologous genes that are also found to be associated with human height. Given that they appear as candidate genes for height in two species that separated over ~90 million years ago [24], the 30 overlapping genes cannot be considered an artifact of cryptic population structure. We found that 6 out of the 30 genes: ADAM12, SIK3, AXL, INSR, SPRED1 and PAPPA2 are involved in the Signaling by Receptor Tyrosine Kinases pathway, a pathway that is responsible for signaling of the growth hormone (GH) [25]. Two of the overlapping genes ADAMTS17 and ADAMTSL4 participate in the pathway: Defective B3GALTL causes Peters-plus syndrome (PpS). The Peters-plus syndrome is a recessive disorder that results in, among other things, short stature and delay in growth [26]. Altogether, evidence suggests that height in humans and stature in cattle are biologically similar traits and that the orthologous genes associated with height in humans can be considered validated if they are also found to be associated with stature in cattle.

With respect to the identification of genes and gene-associated variants that affect stature in cattle, our results suggest that across-species prior information (results of GWAS for human height) is less informative than within-species prior information (results of GWAS for cattle stature). There are two reasons likely underpinning this finding. Firstly, the cattle (within-species) GWAS [12] was for a trait that is highly heritable. It is easier for GWAS to identify the signals underlying a highly heritable trait such as stature than for a lowly heritable trait e.g. fertility, given that at low heritability, there is larger environmental effect in the phenotype, which acts as noise around the genetic effects. In that sense, the benefit of using prior information from humans to identify genes and gene-associated variants for complex traits in cattle is expected to be higher for lowly heritable cattle traits than for highly heritable traits. For example, Costilla et al. [27] estimated the proportion of additive genetic variance for age at puberty in cattle explained by SNPs in and around cattle orthologous genes that were associated with age at menarche in humans. They showed that the SNPs in and around the candidate genes explained 11.2% of the total additive genetic variance for age at puberty in cattle, which was about twice the variance explained by SNPs in random gene-sets of similar size and SNP number.

Secondly, the within-species GWAS was large in terms of sample size and therefore was powerful enough such that the human prior information does not provide additional information. However, the GWAS study of [12] is an exceptional collaboration and data sharing in the livestock genetics community. An important factor for the success of the collaboration was that stature is generally not considered an economic trait, and in that sense, data on stature is less sensitive to share. Studies as large as that of [12] for economically important traits, especially those that are difficult or expensive to measure, are less common. For these kinds of traits, GWAS will most likely be underpowered, given that they are carried out at population level, where, in most cases, sample sizes are limiting. Identification of the genes and gene-associated variants that affects such traits will be aided by the integration of the huge resources, namely, GWAS results that are publicly available for complex traits in humans. Furthermore, for traits in cattle where powerful GWAS are not possible for whatever reason, using prior information from human GWAS from similar traits will likely aid the identification of the underlying genes and the associated variants.

The method of determining the genes that underlie height in humans is very important, as this may determine whether or not the orthologs affects height in other species. In this study, we focused on two sets of orthologous genes that were determined to have an effect on human height. The COJO gene set are those that were prioritized as encompassing or close to the top GWAS signals for height in humans from [10]. The cattle orthologs of these genes were more enriched for cattle stature signals and captured more genetic variance for stature than random sample of genes in the cattle genome. The second set of genes were those that were prioritized based on the summary-based mendelian randomization analysis (SMR) in [10]. These are genes whose expression levels and association with height are assumed to be controlled by the same causal variant, and in that sense, assumed to be the most functionally relevant genes for human height. In our analysis, however, the cattle orthologs of the SMR-based human genes were not more enriched for cattle stature GWAS signals than random sets of genes from the cattle genome. There are two possible explanations for this observation. First, some genes are known to show tissue specific gene expression [28]. For their SMR analysis, Yengo et al. [10] used the GTEx-v7 database [29] that contains expression data of several genes in multiple tissues, not necessarily height related tissues. It is possible, therefore, that some of the genes showed differential level of gene-expression only in specific tissues that are not related to human height. The orthologs of such genes are less likely to affect height in other species. In contrast, genes with consistent differential expression in multiple tissue are more likely to maintain their effect across species. Our result, although only marginally significant (p = 0.02), suggests that to be the case. Given that the specific tissues that are relevant for specific traits of interest are not always known, the results of SMR analysis that uses gene expression data in only a single tissue, should be taken with caution. The second possible explanation might just be that the cattle meta-GWAS analysis of Bouwman et al. [12] was not powerful enough to detect important cattle stature variants within the cattle orthologs of the SMR-based human genes.

Potentially, GWAS for complex traits in humans could be utilized to aid precision genomic prediction in livestock. For example, knowledge about the position of candidate genes from GWAS in humans or the proportion of genetic variance for the trait that is explained by variants in or around the candidate genes, can be used to specify priors for genomic prediction models in livestock. A non-linear genomic prediction model that can be used for this purpose is the BayesR by segment (BayesRS) model [30]. In BayesRS, variant effects are assumed to belong to a mixture of four normal distributions, with the different distribution reflecting the different proportions of genetic variance explained by different genomic segments. Effectively, the BayesRS model can be specified such that the prior probability for a given variant depends on whether it is within or nearby an important gene identified from a human GWAS. An extended BayesRS model is the BayesRC model [31], which also allows for the explicit incorporation of prior knowledge in GP. In BayesRC, information on genes or variants from a human based GWAS can be used to assign variants into two or more classes (C). Other methods of using information from GWAS in humans for genomic prediction in livestock include weighting variants differently based on information from human GWAS in the construction of GRMs e.g. [32, 33], and fitting multiple GRMs in BLUP, with the different GRMs made from different classes of markers e.g. [34–36].

Altogether, our results suggest that human height and cattle stature are biologically similar traits. For the purpose of identifying the genes and gene-associated variants that affect stature in cattle, prior information from current human-height GWAS, does not provide more benefit than what a powerful GWAS for cattle-stature already provides. We must state that our findings cannot easily be generalized to other complex traits. This is especially so for traits that have been under strong artificial selection in cattle, given that such traits might have evolved differently and rapidly than in humans. More research is needed to make such a generalization.

Materials and methods

Data

The summary level results from meta-GWAS for human height and cattle stature and other data sets used in the study are summarised in Table 1.

Table 1. A summary of all data sets used in the study.

Species	Data type	Sample size	# of SNPs	Reference
Human	GWAS summary-level result	253,288	2,550,858	[9]
Human	GWAS summary-level result	693,529^*	2,334,001	[10]
Cattle	GWAS summary-level result	58,265	25.4 million	[12]
Human	Individual-level imputed genotypes	8,552	16,652,994	[43]
Cattle	Individual-level whole genome sequence genotypes	1,147	30.3 million	[41]
Human	LD scores	10,000	17,613,557	[39]
Cattle	Individual-level imputed genotypes and phenotype for stature	975	14,341,737	[34]

Open in a new tab

*Mean sample size.

Comparison of gene characteristics in humans and cattle

Many genes are conserved across species, therefore, comparison of traits across humans and cattle can be made at the gene level. For this study, we utilized information on orthologous genes between humans and cattle. Out of the 24,617 cattle genes identified in the UMD3.1 cattle genome build [37] and using Biomart Ensemb [38], we identified 13,742 genes that have orthologs in humans and passed quality control (QC). The QC thresholds used were: identity of the cattle gene to human gene > 70%; cattle gene order conservation score > 70%; cattle gene orthology confidence = 1, orthology type: 1 to 1.

Within each species (humans and cattle), we estimated the correlation between gene size and LD score of orthologous genes to see if longer genes have higher or lower level of LD than shorter genes. Furthermore, we estimated the correlation between human gene-size and cattle gene-size for the orthologous genes, and did the same for gene-based LD score. We calculated gene-size by subtracting the start, in base pairs (bp), from the end position of the gene and divided the value by 1,000 to obtain gene size in kilo base pairs (kbp). The start and end positions of a gene were defined as the beginning of the 5’ UTR and the end of the 3’ UTR of the gene, respectively. For humans, we used already calculated LD scores on UK10K [39] whole genome sequence variants. The LD score for each SNP was calculated as the sum of LD (r²) between the SNP and all other SNPs in a 200kbp window, with 100kp overlap between two adjacent segments. The mean LD score of a gene was calculated as the mean LD score of all the SNPs within the gene. For cattle, we used GCTA [40] to calculated the LD scores for sequence variants on individuals from the Run4 of the 1,000 bulls genomes project dataset [41] that have 3 or more copies of the minor allele. The LD score for cattle SNPs was calculated with the same specifications as was used for human LD score calculations.

A high correlation between the effect of orthologous genes on human height and cattle stature may be an indication that the genes perform similar function across the species. Using the fastBat function in GCTA [42], and GWAS summary level results, we estimated the aggregated effect of all the SNPs within each orthologous gene on the phenotype (human height and cattle stature), which we refer to as gene-effect. We then calculated the correlation between the effect of the genes in both species. For the fastBat analysis to estimate the effect of the orthologous genes on human height, we used summary-level GWAS result from [10] and individual level imputed genotype data on individuals from the Health and Retirement Study [43] as LD reference. Similarly, for the fastBat analysis to estimate the effect of the orthologous genes on cattle stature, we used the summary-level result from [12] and the full sequence data on all individuals from Run4 of the 1,000 Bull Genomes Project [41] as LD reference. For human data, it was shown previously that the fastBat analysis loses power when SNPs are in very high or perfect LD, hence the recommendation to apply LD threshold of 0.9 [42]. We therefore used an LD threshold of 0.9 in the fastBat analysis of human data and set a lower LD threshold of 0.5 for the fastBat analysis of cattle data, given the high average LD between SNPs in the cattle genome. Outputs from the fastBat analysis were gene-based chi-squared (calculated as the sum of chi-squared values for all the SNPs within and 50kbp either side of a gene), and the corresponding gene-based p-value that indicate whether or not the effect of a gene on height (or stature in cattle) is significantly different from zero.

Overlap of height-associated genes in humans and cattle

We estimated the number of orthologous genes in or close to GWAS peaks for human height that are also located in or close to GWAS peaks for stature in cattle, and tested if the overlap is more than can be expected by chance. Using the Biomart Ensembl tool, we identified the genes potentially affecting stature in cattle as those whose start or end position in the genome are located within 10kbp of the physical positions of the 163 lead SNPs for stature as reported in [12]. We identified 77 orthologous genes in total. Similarly, we identified a total of 368 human orthologous genes whose start or end positions in the genome are within 10kbp of the physical positions of the 649 lead SNPs for height as reported in [9]. We then checked the number of the 77 potential cattle genes that are part of the 368 potential human height genes.

Given the larger sample size used by [10], we wanted to know if there is a stronger overlap between the 77 cattle genes from [12], and human height genes from [10], as against those from [9]. To do this, we used the Biomart Ensembl tool to identify all human orthologous genes whose start or end positions in the genome are within 10kbp of the physical positions of the 3,290 near independent significant SNPs from [10], 1,553 genes in total. We then checked the number of the 77 potential cattle stature genes that are part of the 1,553 potential human height genes and tested if the number is more than can be expected by chance. In both cases, we used Fisher’s exact test.

There is a chance that genes are enriched for GWAS signals and are identified as important genes for height in both humans and cattle, simply because they are large genes [44].Thus, we implemented the Cochran-Mantel-Haenszel test [45, 46] to test for the association between cattle stature genes and human height genes, while at the same time, accounting for a possible confounding effect of gene size. We implemented the test only for the overlap between the 77 cattle stature genes identified from [12], and human height genes identified from [10].

Enrichment for cattle stature GWAS signals in and around human height-associated genes

The aim in this section was to quantify the usefulness of information from human height GWAS in the prioritization of stature genes in cattle, an analysis similar to [18]. Specifically, we tested the hypothesis that cattle genes prioritized based on information from a GWAS on human height are more enriched for cattle stature GWAS signals than random sets of genes from the cattle genome. For the analysis, we used two different sets of human-height genes. First, we used the 610 human height genes identified by [10] using the Summary-based Mendelian Randomization (SMR) (Zhu et al. 2016) analysis [19]. The SMR analysis identifies genes whose expression level and association with the trait are assumed to be controlled by the same causal variant. We identified 370 cattle orthologs that will henceforth be referred to as SMR candidate (SMR-C) genes. The number of near-independent significant SNPs for cattle stature within and around the 370 SMR-C genes were then identified. To test if this number is more than can be observed with a random set of 370 cattle genes, we implemented a validation analysis, using randomized permutation test as follows: all 13,742 cattle genes that have human orthologs were stratified into 10 quantiles based on gene-size, number of SNPs within a gene or gene-based LD score, with the genes stratified separately for each criterion. In each quantile, we identified the number of SMR-C genes and subsequently sampled the same number of control genes from the quantile. Using this method, we sampled 370 control genes 1,000 times. We identified the number of near-independent SNPs within and 100kbp, 50kbp or 5kbp either side of the genes in each sampled control gene-set and checked if the number is more, equal or less than was observed within and 100kbp, 50kbp or 5kbp either side of the 370 SMR-C genes, respectively. For the GWAS signal enrichment analysis, near-independent significant SNPs for stature refers to SNPs obtained following a clumping analysis in plink 1.9 [47], using the summary level result from [12] and the fully sequenced individuals from Run4 of the 1000 Bull Genomes Project [41] as LD reference. The following parameters thresholds were set in the clumping analysis: significance level for index SNPs: 5x10-8; secondary significance level for clumped SNPs: 5x10-8; LD threshold for clumping: 0.2; physical distance threshold for clumping: 0.5mbp.

Secondly, we mapped all the 3,290 near-independent significant SNPs affecting human height identified through a conditional and joint effect (COJO) analysis by [10] to human genes, using Biomart Ensemb and the GRCh37 human reference genome. We identified 1,184 cattle orthologs of the COJO-based human height genes. These 1,184 cattle orthologous genes will henceforth be referred to as COJO candidate (COJO-C) genes. We carried out the same validation analysis as described for the SMR-C genes, with 1,184 control genes sampled 1,000 times.

In addition to possible differences in the usefulness of human prior information used to prioritize the SMR-C and the COJO-C cattle gene sets, respectively, difference in gene sample size (370 vs 1,184) can result in a difference in the level of enrichment for GWAS signals. To account for the difference in gene sample size, we took a sample of 370 genes from the 1,184 COJO-C genes 100 times, and checked the number of near-independent SNPs within and 100kbp either side of the genes in each sample. We then checked if the number of near independent SNPs for stature in each sample is more or less than was observed within and 100kbp either side of the 370 SMR-C genes.

Furthermore, we tested if the level of enrichment for GWAS hits within the SMR-C genes is stronger with more stringent selection criterium, namely, evidence of differential expression in multiple tissues. To do this, we identified 182 out of the 610 human height genes prioritized by [10] through SMR, that have differential gene expression evidence in 3 or more tissues. We located the cattle orthologs of the 182 genes, with a total of 121 remaining after QC. These 121 cattle orthologs genes will henceforth be referred to as: SMR evidence in multiple tissues candidate (SMR-MT-C) genes. We implemented a validation analysis as earlier described, to test if the number of near-independent SNPs within the 121 SMR-MT-C genes are more than can be observed within and around sampled control genes. In this case, we used stratified random sampling to sample 121 control genes 1,000 times, with genes stratified only based on gene size, as this gives similar results compared to when genes are stratified based on gene-based LD score or SNP density.

Accuracy of cattle variants selection using within- and across-species prior information

The aim in this section was to quantify the accuracy (in terms of the proportion of genetic variance explained for stature), of pre-selecting cattle SNPs based on GWAS results for height in humans. For the analysis, deregressed estimated breeding values (dEBVs) for stature and effective daughter contributions (EDCs), for 975 New Zealand Holstein (NZH) bulls was used. The EDCs were used as weights for the dEBVs in all analyses. To estimate the proportion of genetic variance explained by SNPs in the NZH population, a univariate GREML model with two genomic relationship matrices (GRMs) was fitted using the MTG2 software [48] follows:

y = 1 μ + W g_{1} + W g_{2} + e,

where y is a vector containing the phenotypes (dEBVs of the NZH bulls), μ is the mean, W is an incidence matrix linking phenotypes to genetic effects, g₁ and g₂ are vectors of genetic effects corresponding to the two GRMs fitted simultaneously in the model, and e is a vector of residual effects. In this study, the first GRM (GRM₁) was made using the SNPs of interest and the second GRM (GRM₂) was made from 50k SNPs that were randomly sampled across the genome, excluding all SNPs in GRM₁, and meant to account for possible population structure. Genetic effects in g₁ and g₂ are assumed to be normally distributed as $g_{1} \sim N (0, G R M_{1} σ_{a}^{2})$ and $g_{2} \sim N (0, G R M_{2} σ_{a}^{2})$ , where $σ_{a}^{2}$ is the genetic variance for stature. Residual effects were assumed to be normally distributed as $e \sim N (O, D σ_{e}^{2})$ , where $σ_{e}^{2}$ is the residual variance and D is a diagonal matrix that contains the inverse of EDCs, which were used as weights for the dEBVs. SNPs were included in the analysis that had at least 20 copies of the minor allele in NZH population. Thus, MAF threshold was set to $\frac{20}{975} * 2 = 0.04$ . We tested two hypotheses as follows:

Hypothesis 1: Cattle SNPs that are pre-selected based on a combination of two information sources: 1) genes identified from GWAS results for human height, and 2) Cattle stature GWAS significance level, will explain more genetic variance for stature than SNPs that are pre-selected based from random cattle genes based on their significance level in a GWAS for cattle stature. To test this hypothesis, we identified the near-independent significant SNPs for stature in and 100kbp either side of the SMR-C and COJO-C genes, that segregate in NZH (80 SMR-C SNPs, 391 COJO-C SNPs). We used all the identified candidate SNPs to make GRM₁ in GCTA [40]. For the control analysis, we sampled 1,000 control gene sets, and for each control gene set we ranked the SNPs based on significance level for association with stature and sampled the top 80 (SMR) or 391 (COJO) SNPs. So each set of control SNP set had the same number of SNPs as the candidate SNP set and was used to calculate GRM₁. The proportion of total genetic variance explained by each set of control SNPs were estimated and compared to the genetic variance explained by the candidate SNPs.

Hypothesis 2: Cattle SNPs that are pre-selected based on GWAS result for human height will explain more genetic variance for stature, than random set of SNPs from random cattle genes. In this case, the candidate SNPs were all the SNPs identified within and 100kbp either side of the SMR-C and COJO-C genes, respectively. We randomly sampled 5 SNPs per gene to avoid a large number of SNP in high LD with each other, given the long range LD in commercial cattle populations. For the SMR-C genes, this resulted in 1,777 unique SNPs that were segregating in NZH. For the COJO-C genes, the number of unique SNPs extracted was 5,781. Each of these SNP set was used to calculate GRM₁. For the control analysis, the same number of SNPs as the candidate SNPs (1,777 and 5,781 SNPs respectively) were randomly sampled from within and 100kbp around the genes in each of the 1,000 control gene-sets and the SNPs were matched for allele frequency to the candidate SNPs.

Supporting information

S1 Fig. Distribution of average marker density in the 10kbp windows of association with lead SNPs from Bouwman et al. [12], Wood et al. [9] and Yengo et al. [10] and the average marker density of genes in both cattle and humans.

Average marker density were computed as total number of markers in a window or gene divided by the length (in kbp) of the window or gene.

(TIF)

Click here for additional data file.^{(70KB, tif)}

S1 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Wood et al. [9].

*Cattle genes with 1 to 1 orthologs in humans (after QC); **Cattle genes within 10kb of the 164 lead SNPs from Bouwman et al. [12] that also have orthologs in humans; ***Human height genes prioritized as those that overlap with, or are within 10kkp either side of the 649 lead SNPs for height in Wood et al. [9], that also have orthologs in cattle. The proportion 10/77 is more than can be expected by chance: Fisher’s exact test (odds ratio = 5.5, p-value = 3.7e-05).

(DOCX)

Click here for additional data file.^{(20KB, docx)}

S2 Table. Names and coordinates of 10 orthologous that are associated with both human height based on results from Wood et al. [9] and cattle stature based on results from Bouwman et al. [12].

(DOCX)

Click here for additional data file.^{(20.4KB, docx)}

S3 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Yengo et al. [10].

*Cattle genes with 1 to 1 orthologs in humans (after QC); **Cattle genes within 10kbp of the 164 lead SNPs from Bouwman et al. [12] that also have orthologs in humans; ***Human height genes prioritized as those that overlap with, or are within 10kbp either side of the 3,290 lead SNPs for height in Yengo et al. [10], that also have orthologs in cattle. The proportion 30/77 is more than can be expected by chance: Fisher’s exact test (odds ratio = 5.1, p-value = 3.1e-10).

(DOCX)

Click here for additional data file.^{(19.7KB, docx)}

S4 Table. Names and coordinates of 30 orthologous that are associated with both human height based on results from Yengo et al. [10] and cattle stature based on results from Bouwman et al. [12].

(DOCX)

Click here for additional data file.^{(23KB, docx)}

S5 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Wood et al. [9], with the average marker density of the genes in both species.

⁺ Total number of markers in the genes divided by gene length (in 10kpb). 10kbps windows were used to avoid having values of less than 1 in the contingency table. *Cattle genes with 1 to 1 orthologs in humans (after QC); **Cattle genes within 10kb of the 164 lead SNPs from Bouwman et al. [12] that also have orthologs in humans; ***Human height genes prioritized as those that overlap with, or are within 10kkp either side of the 649 lead SNPs for height in Wood et al. [9], that also have orthologs in cattle. Fisher’s exact test of non-random association in gene marker density in cattle and humans (p-value = 0.87).

(DOCX)

Click here for additional data file.^{(21.3KB, docx)}

S6 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Yengo et al. [10], while accounting for gene size.

Mantel-Haenszel chi-squared (corrected for gene length) = 35.6, p-value < 2.4e-9, common odds ratio = 4.1.

(DOCX)

Click here for additional data file.^{(19.6KB, docx)}

Data Availability

The (GWAS) summary statistics data for human height and cattle stature have been published previously. Genotype and phenotype data for a small cattle population, which we used to validate some of the results in this study were obtained from CRV BV, a commercial cattle breeding company. As such, the company cannot make these data publicly available.

Funding Statement

The author(s) received no specific funding for this work.

References

1.Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS, et al. Potential etiologic and functional implications of genome-wide association loci for human diseases and traits. Proceedings of the National Academy of Sciences. 2009;106(23):9362–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Visscher PM, Wray NR, Zhang Q, Sklar P, McCarthy MI, Brown MA, et al. 10 years of GWAS discovery: biology, function, and translation. The American Journal of Human Genetics. 2017;101(1):5–22. 10.1016/j.ajhg.2017.06.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Brachi B, Morris GP, Borevitz JO. Genome-wide association studies in plants: the missing heritability is in the field. Genome biology. 2011;12(10):232 10.1186/gb-2011-12-10-232 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Sharma A, Lee JS, Dang CG, Sudrajad P, Kim HC, Yeon SH, et al. Stories and challenges of genome wide association studies in livestock—a review. Asian-Australasian journal of animal sciences. 2015;28(10):1371 10.5713/ajas.14.0715 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Gudbjartsson DF, Walters GB, Thorleifsson G, Stefansson H, Halldorsson BV, Zusmanovich P, et al. Many sequence variants affecting diversity of adult human height. Nature genetics. 2008;40(5):609 10.1038/ng.122 [DOI] [PubMed] [Google Scholar]
6.Lettre G, Jackson AU, Gieger C, Schumacher FR, Berndt SI, Sanna S, et al. Identification of ten loci associated with height highlights new biological pathways in human growth. Nature genetics. 2008;40(5):584 10.1038/ng.125 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Weedon MN, Lango H, Lindgren CM, Wallace C, Evans DM, Mangino M, et al. Genome-wide association analysis identifies 20 loci that influence adult height. Nature genetics. 2008;40(5):575 10.1038/ng.121 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Allen HL, Estrada K, Lettre G, Berndt SI, Weedon MN, Rivadeneira F, et al. Hundreds of variants clustered in genomic loci and biological pathways affect human height. Nature. 2010;467(7317):832 10.1038/nature09410 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Wood AR, Esko T, Yang J, Vedantam S, Pers TH, Gustafsson S, et al. Defining the role of common variation in the genomic and biological architecture of adult human height. Nature genetics. 2014;46(11):1173 10.1038/ng.3097 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Yengo L, Sidorenko J, Kemper KE, Zheng Z, Wood AR, Weedon MN, et al. Meta-analysis of genome-wide association studies for height and body mass index in∼ 700000 individuals of European ancestry. Human molecular genetics. 2018;27(20):3641–9. 10.1093/hmg/ddy271 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Yang J, Zaitlen NA, Goddard ME, Visscher PM, Price AL. Advantages and pitfalls in the application of mixed-model association methods. Nature genetics. 2014;46(2):100 10.1038/ng.2876 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Bouwman AC, Daetwyler HD, Chamberlain AJ, Ponce CH, Sargolzaei M, Schenkel FS, et al. Meta-analysis of genome-wide association studies for cattle stature identifies common genes that regulate body size in mammals. Nature genetics. 2018;50(3):362 10.1038/s41588-018-0056-5 [DOI] [PubMed] [Google Scholar]
13.Jiang J, Ma L, Prakapenka D, VanRaden PM, Cole JB, Da Y. A large-scale genome-wide association study in US Holstein cattle. Frontiers in genetics. 2019;10:412 10.3389/fgene.2019.00412 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Purfield D, Evans R, Berry D. Reaffirmation of known major genes and the identification of novel candidate genes associated with carcass-related metrics based on whole genome sequence within a large multi-breed cattle population. BMC genomics. 2019;20(1):720 10.1186/s12864-019-6071-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Xiang R, Berg Ivd, MacLeod IM, Hayes BJ, Prowse-Wilkins CP, Wang M, et al. Quantifying the contribution of sequence variants with regulatory and evolutionary significance to 34 bovine complex traits. Proceedings of the National Academy of Sciences. 2019;116(39):19398–408. 10.1073/pnas.1904159116 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Dekkers JC. Commercial application of marker-and gene-assisted selection in livestock: strategies and lessons. Journal of animal science. 2004;82(suppl_13):E313–E28. [DOI] [PubMed] [Google Scholar]
17.Kemper KE, Visscher PM, Goddard ME. Genetic architecture of body size in mammals. Genome biology. 2012;13(4):244 10.1186/gb4016 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Pryce JE, Hayes BJ, Bolormaa S, Goddard ME. Polymorphic regions affecting human height also control stature in cattle. Genetics. 2011;187(3):981–4. 10.1534/genetics.110.123943 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Zhu Z, Zhang F, Hu H, Bakshi A, Robinson MR, Powell JE, et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nature genetics. 2016;48(5):481 10.1038/ng.3538 [DOI] [PubMed] [Google Scholar]
20.Koonin EV. Orthologs, paralogs, and evolutionary genomics. Annu Rev Genet. 2005;39:309–38. 10.1146/annurev.genet.39.073003.114725 [DOI] [PubMed] [Google Scholar]
21.Descorps-Declère S, Lemoine F, Sculo Q, Lespinet O, Labedan B. The multiple facets of homology and their use in comparative genomics to study the evolution of genes, genomes, and species. Biochimie. 2008;90(4):595–608. 10.1016/j.biochi.2007.09.010 [DOI] [PubMed] [Google Scholar]
22.Gabaldón T, Koonin EV. Functional and evolutionary implications of gene orthology. Nature Reviews Genetics. 2013;14(5):360–6. 10.1038/nrg3456 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Dolinski K, Botstein D. Orthology and functional conservation in eukaryotes. Annu Rev Genet. 2007;41:465–507. 10.1146/annurev.genet.40.110405.090439 [DOI] [PubMed] [Google Scholar]
24.Murphy WJ, Pevzner PA, O'Brien SJ. Mammalian phylogenomics comes of age. TRENDS in Genetics. 2004;20(12):631–9. 10.1016/j.tig.2004.09.005 [DOI] [PubMed] [Google Scholar]
25.Carter-Su C, Schwartz J, Argetsinger LS. Growth hormone signaling pathways. Growth Hormone & IGF Research. 2016;28:11–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Heinonen TY, Mäki M. Peters’-plus syndrome is a congenital disorder of glycosylation caused by a defect in the β1, 3-glucosyltransferase that modifies thrombospondin type 1 repeats. Annals of medicine. 2009;41(1):2–10. 10.1080/07853890802301975 [DOI] [PubMed] [Google Scholar]
27.Costilla R, Warburton C, Hayes B, editors. Genetic control of fertility traits across species: variance in tropical beef heifers’age at puberty explained by genes controling age at menarche in women. Proceeding, Association for the Advancement of Animal Breeding and Genetics; 2020 [Google Scholar]
28.Nica AC, Parts L, Glass D, Nisbet J, Barrett A, Sekowska M, et al. The architecture of gene regulatory variation across multiple human tissues: the MuTHER study. PLoS genetics. 2011;7(2):e1002003 10.1371/journal.pgen.1002003 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Consortium G. Genetic effects on gene expression across human tissues. Nature. 2017;550(7675):204 10.1038/nature24277 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Brøndum RF, Su G, Lund MS, Bowman PJ, Goddard ME, Hayes BJ. Genome position specific priors for genomic prediction. BMC genomics. 2012;13(1):543. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.MacLeod I, Bowman P, Vander Jagt C, Haile-Mariam M, Kemper K, Chamberlain A, et al. Exploiting biological priors and sequence variants enhances QTL discovery and genomic prediction of complex traits. BMC genomics. 2016;17(1):144. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Su G, Christensen OF, Janss L, Lund MS. Comparison of genomic predictions using genomic relationship matrices built with different weighting factors to account for locus-specific variances. Journal of dairy science. 2014;97(10):6547–59. 10.3168/jds.2014-8210 [DOI] [PubMed] [Google Scholar]
33.Fragomeni BO, Lourenco DA, Masuda Y, Legarra A, Misztal I. Incorporation of causative quantitative trait nucleotides in single-step GBLUP. Genetics Selection Evolution. 2017;49(1):59. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Raymond B, Bouwman AC, Wientjes YC, Schrooten C, Houwing-Duistermaat J, Veerkamp RF. Genomic prediction for numerically small breeds, using models with pre-selected and differentially weighted markers. Genetics Selection Evolution. 2018;50(1):49. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Speed D, Balding DJ. MultiBLUP: improved SNP-based prediction for complex traits. Genome research. 2014;24(9):1550–7. 10.1101/gr.169375.113 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Fang L, Sahana G, Ma P, Su G, Yu Y, Zhang S, et al. Use of biological priors enhances understanding of genetic architecture and genomic prediction of complex traits within and between dairy cattle breeds. BMC genomics. 2017;18(1):604 10.1186/s12864-017-4004-z [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Elsik CG, Tellam RL, Worley KC. The genome sequence of taurine cattle: a window to ruminant biology and evolution. Science. 2009;324(5926):522–8. 10.1126/science.1169588 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Chen Y, Cunningham F, Rios D, McLaren WM, Smith J, Pritchard B, et al. Ensembl variation resources. BMC genomics. 2010;11(1):293. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.consortium UK. The UK10K project identifies rare variants in health and disease. Nature. 2015;526(7571):82–90. 10.1038/nature14962 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics. 2011;88(1):76–82. 10.1016/j.ajhg.2010.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Daetwyler HD, Capitan A, Pausch H, Stothard P, Van Binsbergen R, Brøndum RF, et al. Whole-genome sequencing of 234 bulls facilitates mapping of monogenic and complex traits in cattle. Nature genetics. 2014;46(8):858 10.1038/ng.3034 [DOI] [PubMed] [Google Scholar]
42.Bakshi A, Zhu Z, Vinkhuyzen AA, Hill WD, McRae AF, Visscher PM, et al. Fast set-based association analysis using summary data from GWAS identifies novel gene loci for human complex traits. Scientific reports. 2016;6:32894 10.1038/srep32894 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JW, Weir DR. Cohort profile: the health and retirement study (HRS). International journal of epidemiology. 2014;43(2):576–85. 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Mirina A, Atzmon G, Ye K, Bergman A. Gene size matters. PloS one. 2012;7(11):e49093 10.1371/journal.pone.0049093 [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Cochran WG. Some methods for strengthening the common χ 2 tests. Biometrics. 1954;10(4):417–51. [Google Scholar]
46.Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. Journal of the national cancer institute. 1959;22(4):719–48. [PubMed] [Google Scholar]
47.Purcell S, Chang C. PLINK 1.9. URL https://www cog-genomics org/plink2. 2015.
48.Lee SH, Van der Werf JH. MTG2: an efficient algorithm for multivariate linear mixed model analysis based on genomic information. Bioinformatics. 2016;32(9):1420–2. 10.1093/bioinformatics/btw012 [DOI] [PMC free article] [PubMed] [Google Scholar]

PLoS Genet. doi: 10.1371/journal.pgen.1008780.r001

Decision Letter 0

Gregory S Barsh, Jennie Elizabeth Pryce

10 Jun 2020

* Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. *

Dear Dr Raymond,

Thank you very much for submitting your Research Article entitled 'Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock' to PLOS Genetics. Your manuscript was fully evaluated at the editorial level and by an independent peer reviewer. The reviewer appreciated the attention to an important topic but identified some aspects of the manuscript that should be improved.

We therefore ask you to modify the manuscript according to the review recommendations before we can consider your manuscript for acceptance. Your revisions should address the specific points made by each reviewer.

In addition we ask that you:

1) Provide a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

2) Upload a Striking Image with a corresponding caption to accompany your manuscript if one is available (either a new image or an existing one from within your manuscript). If this image is judged to be suitable, it may be featured on our website. Images should ideally be high resolution, eye-catching, single panel square images. For examples, please browse our archive. If your image is from someone other than yourself, please ensure that the artist has read and agreed to the terms and conditions of the Creative Commons Attribution License. Note: we cannot publish copyrighted images.

We hope to receive your revised manuscript within the next 30 days. If you anticipate any delay in its return, we would ask you to let us know the expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments should be included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, you will need to go to the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

Please let us know if you have any questions while making these revisions.

Yours sincerely,

Jennie Pryce

Guest Editor

PLOS Genetics

Gregory Barsh

Editor-in-Chief

PLOS Genetics

I am happy to endorse the opinion of the reviewer of this manuscript. The paper requires only very minor editing before it is ready for publication. It's a very interesting read.

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: Minor points:

line 136 –“…. no significant (p<0.05) ….” The ‘p<0.05’ is redundant as the exact p value has been reported nearby. Please remove the (p<0.05).

line 386-390 – Please be specific about the coordinates of start and end of the gene. Assuming this includes beginning of 5’UTR till the end of 3’ UTR. If so, please include in the methods description.

Figure2&3 Please make sure the y axis range is the same for both cattle and human graphs.

Table S1. Is the density of markers in datasets used the same with regards the 10K window of association with lead SNPs? There is a chance that OR of orthologue genes on chips with different marker density and genome gaps, affect the total likelihood of associations with said phenotype. Adding the marker density of the 10Kbp associated window and its average per gene in both species could clarify this issue. If there is a significant density difference (could be assessed the same way by Fisher’s exact 2x1 table) per same orthologue gene, please include that in the result section lines 149-165. This point also lends itself to the gene size-stratification analysis that has already been done using Cochran-Mantel- Haenszel in the following section.

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Mazdak Salavati DVM, PhD, AFHEA

PLoS Genet. 2020 Sep 14;16(9):e1008780. doi: 10.1371/journal.pgen.1008780.r002

Author response to Decision Letter 0

6 Jul 2020

Attachment

Submitted filename: response to reviewer .docx

Click here for additional data file.^{(1.6MB, docx)}

PLoS Genet. doi: 10.1371/journal.pgen.1008780.r003

Decision Letter 1

Gregory S Barsh, Jennie Elizabeth Pryce

21 Jul 2020

Dear Dr Raymond,

We are pleased to inform you that your manuscript entitled "Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock" has been editorially accepted for publication in PLOS Genetics. Congratulations!

Before your submission can be formally accepted and sent to production you will need to complete our formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Please note: the accept date on your published article will reflect the date of this provisional accept, but your manuscript will not be scheduled for publication until the required changes have been made.

Once your paper is formally accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosgenetics@plos.org.

In the meantime, please log into Editorial Manager at https://www.editorialmanager.com/pgenetics/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production and billing process. Note that PLOS requires an ORCID iD for all corresponding authors. Therefore, please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

If you have a press-related query, or would like to know about one way to make your underlying data available (as you will be aware, this is required for publication), please see the end of this email. If your institution or institutions have a press office, please notify them about your upcoming article at this point, to enable them to help maximise its impact. Inform journal staff as soon as possible if you are preparing a press release for your article and need a publication date.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Genetics!

Yours sincerely,

Jennie Elizabeth Pryce

Guest Editor

PLOS Genetics

Gregory Barsh

Editor-in-Chief

PLOS Genetics

www.plosgenetics.org

Twitter: @PLOSGenetics

----------------------------------------------------

Comments from the reviewers (if applicable):

----------------------------------------------------

Data Deposition

If you have submitted a Research Article or Front Matter that has associated data that are not suitable for deposition in a subject-specific public repository (such as GenBank or ArrayExpress), one way to make that data available is to deposit it in the Dryad Digital Repository. As you may recall, we ask all authors to agree to make data available; this is one way to achieve that. A full list of recommended repositories can be found on our website.

The following link will take you to the Dryad record for your article, so you won't have to re‐enter its bibliographic information, and can upload your files directly:

http://datadryad.org/submit?journalID=pgenetics&manu=PGENETICS-D-20-00564R1

More information about depositing data in Dryad is available at http://www.datadryad.org/depositing. If you experience any difficulties in submitting your data, please contact help@datadryad.org for support.

Additionally, please be aware that our data availability policy requires that all numerical data underlying display items are included with the submission, and you will need to provide this before we can formally accept your manuscript, if not already present.

----------------------------------------------------

Press Queries

If you or your institution will be preparing press materials for this manuscript, or if you need to know your paper's publication date for media purposes, please inform the journal staff as soon as possible so that your submission can be scheduled accordingly. Your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version. PLOS Genetics may also choose to issue a press release for your article. If there's anything the journal should know or you'd like more information, please get in touch via plosgenetics@plos.org.

PLoS Genet. doi: 10.1371/journal.pgen.1008780.r004

Acceptance letter

Gregory S Barsh, Jennie Elizabeth Pryce

4 Sep 2020

PGENETICS-D-20-00564R1

Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock

Dear Dr Raymond,

We are pleased to inform you that your manuscript entitled "Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock" has been formally accepted for publication in PLOS Genetics! Your manuscript is now with our production department and you will be notified of the publication date in due course.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

Soon after your final files are uploaded, unless you have opted out or your manuscript is a front-matter piece, the early version of your manuscript will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting PLOS Genetics and open-access publishing. We are looking forward to publishing your work!

With kind regards,

Kaitlin Butler

PLOS Genetics

On behalf of:

The PLOS Genetics Team

Carlyle House, Carlyle Road, Cambridge CB4 3DN | United Kingdom

plosgenetics@plos.org | +44 (0) 1223-442823

plosgenetics.org | Twitter: @PLOSGenetics

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Average marker density were computed as total number of markers in a window or gene divided by the length (in kbp) of the window or gene.

(TIF)

Click here for additional data file.^{(70KB, tif)}

S1 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Wood et al. [9].

(DOCX)

Click here for additional data file.^{(20KB, docx)}

S2 Table. Names and coordinates of 10 orthologous that are associated with both human height based on results from Wood et al. [9] and cattle stature based on results from Bouwman et al. [12].

(DOCX)

Click here for additional data file.^{(20.4KB, docx)}

S3 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Yengo et al. [10].

(DOCX)

Click here for additional data file.^{(19.7KB, docx)}

S4 Table. Names and coordinates of 30 orthologous that are associated with both human height based on results from Yengo et al. [10] and cattle stature based on results from Bouwman et al. [12].

(DOCX)

Click here for additional data file.^{(23KB, docx)}

(DOCX)

Click here for additional data file.^{(21.3KB, docx)}

S6 Table. Contingency table showing the overlap of cattle stature genes from Bouwman et al. [12] with human height genes from Yengo et al. [10], while accounting for gene size.

Mantel-Haenszel chi-squared (corrected for gene length) = 35.6, p-value < 2.4e-9, common odds ratio = 4.1.

(DOCX)

Click here for additional data file.^{(19.6KB, docx)}

Attachment

Submitted filename: response to reviewer .docx

Click here for additional data file.^{(1.6MB, docx)}

Data Availability Statement

[pgen.1008780.ref001] 1.Hindorff LA, Sethupathy P, Junkins HA, Ramos EM, Mehta JP, Collins FS, et al. Potential etiologic and functional implications of genome-wide association loci for human diseases and traits. Proceedings of the National Academy of Sciences. 2009;106(23):9362–7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref002] 2.Visscher PM, Wray NR, Zhang Q, Sklar P, McCarthy MI, Brown MA, et al. 10 years of GWAS discovery: biology, function, and translation. The American Journal of Human Genetics. 2017;101(1):5–22. 10.1016/j.ajhg.2017.06.005 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref003] 3.Brachi B, Morris GP, Borevitz JO. Genome-wide association studies in plants: the missing heritability is in the field. Genome biology. 2011;12(10):232 10.1186/gb-2011-12-10-232 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref004] 4.Sharma A, Lee JS, Dang CG, Sudrajad P, Kim HC, Yeon SH, et al. Stories and challenges of genome wide association studies in livestock—a review. Asian-Australasian journal of animal sciences. 2015;28(10):1371 10.5713/ajas.14.0715 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref005] 5.Gudbjartsson DF, Walters GB, Thorleifsson G, Stefansson H, Halldorsson BV, Zusmanovich P, et al. Many sequence variants affecting diversity of adult human height. Nature genetics. 2008;40(5):609 10.1038/ng.122 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref006] 6.Lettre G, Jackson AU, Gieger C, Schumacher FR, Berndt SI, Sanna S, et al. Identification of ten loci associated with height highlights new biological pathways in human growth. Nature genetics. 2008;40(5):584 10.1038/ng.125 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref007] 7.Weedon MN, Lango H, Lindgren CM, Wallace C, Evans DM, Mangino M, et al. Genome-wide association analysis identifies 20 loci that influence adult height. Nature genetics. 2008;40(5):575 10.1038/ng.121 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref008] 8.Allen HL, Estrada K, Lettre G, Berndt SI, Weedon MN, Rivadeneira F, et al. Hundreds of variants clustered in genomic loci and biological pathways affect human height. Nature. 2010;467(7317):832 10.1038/nature09410 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref009] 9.Wood AR, Esko T, Yang J, Vedantam S, Pers TH, Gustafsson S, et al. Defining the role of common variation in the genomic and biological architecture of adult human height. Nature genetics. 2014;46(11):1173 10.1038/ng.3097 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref010] 10.Yengo L, Sidorenko J, Kemper KE, Zheng Z, Wood AR, Weedon MN, et al. Meta-analysis of genome-wide association studies for height and body mass index in∼ 700000 individuals of European ancestry. Human molecular genetics. 2018;27(20):3641–9. 10.1093/hmg/ddy271 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref011] 11.Yang J, Zaitlen NA, Goddard ME, Visscher PM, Price AL. Advantages and pitfalls in the application of mixed-model association methods. Nature genetics. 2014;46(2):100 10.1038/ng.2876 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref012] 12.Bouwman AC, Daetwyler HD, Chamberlain AJ, Ponce CH, Sargolzaei M, Schenkel FS, et al. Meta-analysis of genome-wide association studies for cattle stature identifies common genes that regulate body size in mammals. Nature genetics. 2018;50(3):362 10.1038/s41588-018-0056-5 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref013] 13.Jiang J, Ma L, Prakapenka D, VanRaden PM, Cole JB, Da Y. A large-scale genome-wide association study in US Holstein cattle. Frontiers in genetics. 2019;10:412 10.3389/fgene.2019.00412 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref014] 14.Purfield D, Evans R, Berry D. Reaffirmation of known major genes and the identification of novel candidate genes associated with carcass-related metrics based on whole genome sequence within a large multi-breed cattle population. BMC genomics. 2019;20(1):720 10.1186/s12864-019-6071-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref015] 15.Xiang R, Berg Ivd, MacLeod IM, Hayes BJ, Prowse-Wilkins CP, Wang M, et al. Quantifying the contribution of sequence variants with regulatory and evolutionary significance to 34 bovine complex traits. Proceedings of the National Academy of Sciences. 2019;116(39):19398–408. 10.1073/pnas.1904159116 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref016] 16.Dekkers JC. Commercial application of marker-and gene-assisted selection in livestock: strategies and lessons. Journal of animal science. 2004;82(suppl_13):E313–E28. [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref017] 17.Kemper KE, Visscher PM, Goddard ME. Genetic architecture of body size in mammals. Genome biology. 2012;13(4):244 10.1186/gb4016 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref018] 18.Pryce JE, Hayes BJ, Bolormaa S, Goddard ME. Polymorphic regions affecting human height also control stature in cattle. Genetics. 2011;187(3):981–4. 10.1534/genetics.110.123943 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref019] 19.Zhu Z, Zhang F, Hu H, Bakshi A, Robinson MR, Powell JE, et al. Integration of summary data from GWAS and eQTL studies predicts complex trait gene targets. Nature genetics. 2016;48(5):481 10.1038/ng.3538 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref020] 20.Koonin EV. Orthologs, paralogs, and evolutionary genomics. Annu Rev Genet. 2005;39:309–38. 10.1146/annurev.genet.39.073003.114725 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref021] 21.Descorps-Declère S, Lemoine F, Sculo Q, Lespinet O, Labedan B. The multiple facets of homology and their use in comparative genomics to study the evolution of genes, genomes, and species. Biochimie. 2008;90(4):595–608. 10.1016/j.biochi.2007.09.010 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref022] 22.Gabaldón T, Koonin EV. Functional and evolutionary implications of gene orthology. Nature Reviews Genetics. 2013;14(5):360–6. 10.1038/nrg3456 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref023] 23.Dolinski K, Botstein D. Orthology and functional conservation in eukaryotes. Annu Rev Genet. 2007;41:465–507. 10.1146/annurev.genet.40.110405.090439 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref024] 24.Murphy WJ, Pevzner PA, O'Brien SJ. Mammalian phylogenomics comes of age. TRENDS in Genetics. 2004;20(12):631–9. 10.1016/j.tig.2004.09.005 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref025] 25.Carter-Su C, Schwartz J, Argetsinger LS. Growth hormone signaling pathways. Growth Hormone & IGF Research. 2016;28:11–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref026] 26.Heinonen TY, Mäki M. Peters’-plus syndrome is a congenital disorder of glycosylation caused by a defect in the β1, 3-glucosyltransferase that modifies thrombospondin type 1 repeats. Annals of medicine. 2009;41(1):2–10. 10.1080/07853890802301975 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref027] 27.Costilla R, Warburton C, Hayes B, editors. Genetic control of fertility traits across species: variance in tropical beef heifers’age at puberty explained by genes controling age at menarche in women. Proceeding, Association for the Advancement of Animal Breeding and Genetics; 2020 [Google Scholar]

[pgen.1008780.ref028] 28.Nica AC, Parts L, Glass D, Nisbet J, Barrett A, Sekowska M, et al. The architecture of gene regulatory variation across multiple human tissues: the MuTHER study. PLoS genetics. 2011;7(2):e1002003 10.1371/journal.pgen.1002003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref029] 29.Consortium G. Genetic effects on gene expression across human tissues. Nature. 2017;550(7675):204 10.1038/nature24277 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref030] 30.Brøndum RF, Su G, Lund MS, Bowman PJ, Goddard ME, Hayes BJ. Genome position specific priors for genomic prediction. BMC genomics. 2012;13(1):543. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref031] 31.MacLeod I, Bowman P, Vander Jagt C, Haile-Mariam M, Kemper K, Chamberlain A, et al. Exploiting biological priors and sequence variants enhances QTL discovery and genomic prediction of complex traits. BMC genomics. 2016;17(1):144. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref032] 32.Su G, Christensen OF, Janss L, Lund MS. Comparison of genomic predictions using genomic relationship matrices built with different weighting factors to account for locus-specific variances. Journal of dairy science. 2014;97(10):6547–59. 10.3168/jds.2014-8210 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref033] 33.Fragomeni BO, Lourenco DA, Masuda Y, Legarra A, Misztal I. Incorporation of causative quantitative trait nucleotides in single-step GBLUP. Genetics Selection Evolution. 2017;49(1):59. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref034] 34.Raymond B, Bouwman AC, Wientjes YC, Schrooten C, Houwing-Duistermaat J, Veerkamp RF. Genomic prediction for numerically small breeds, using models with pre-selected and differentially weighted markers. Genetics Selection Evolution. 2018;50(1):49. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref035] 35.Speed D, Balding DJ. MultiBLUP: improved SNP-based prediction for complex traits. Genome research. 2014;24(9):1550–7. 10.1101/gr.169375.113 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref036] 36.Fang L, Sahana G, Ma P, Su G, Yu Y, Zhang S, et al. Use of biological priors enhances understanding of genetic architecture and genomic prediction of complex traits within and between dairy cattle breeds. BMC genomics. 2017;18(1):604 10.1186/s12864-017-4004-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref037] 37.Elsik CG, Tellam RL, Worley KC. The genome sequence of taurine cattle: a window to ruminant biology and evolution. Science. 2009;324(5926):522–8. 10.1126/science.1169588 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref038] 38.Chen Y, Cunningham F, Rios D, McLaren WM, Smith J, Pritchard B, et al. Ensembl variation resources. BMC genomics. 2010;11(1):293. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref039] 39.consortium UK. The UK10K project identifies rare variants in health and disease. Nature. 2015;526(7571):82–90. 10.1038/nature14962 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref040] 40.Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics. 2011;88(1):76–82. 10.1016/j.ajhg.2010.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref041] 41.Daetwyler HD, Capitan A, Pausch H, Stothard P, Van Binsbergen R, Brøndum RF, et al. Whole-genome sequencing of 234 bulls facilitates mapping of monogenic and complex traits in cattle. Nature genetics. 2014;46(8):858 10.1038/ng.3034 [DOI] [PubMed] [Google Scholar]

[pgen.1008780.ref042] 42.Bakshi A, Zhu Z, Vinkhuyzen AA, Hill WD, McRae AF, Visscher PM, et al. Fast set-based association analysis using summary data from GWAS identifies novel gene loci for human complex traits. Scientific reports. 2016;6:32894 10.1038/srep32894 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref043] 43.Sonnega A, Faul JD, Ofstedal MB, Langa KM, Phillips JW, Weir DR. Cohort profile: the health and retirement study (HRS). International journal of epidemiology. 2014;43(2):576–85. 10.1093/ije/dyu067 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref044] 44.Mirina A, Atzmon G, Ye K, Bergman A. Gene size matters. PloS one. 2012;7(11):e49093 10.1371/journal.pone.0049093 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pgen.1008780.ref045] 45.Cochran WG. Some methods for strengthening the common χ 2 tests. Biometrics. 1954;10(4):417–51. [Google Scholar]

[pgen.1008780.ref046] 46.Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. Journal of the national cancer institute. 1959;22(4):719–48. [PubMed] [Google Scholar]

[pgen.1008780.ref047] 47.Purcell S, Chang C. PLINK 1.9. URL https://www cog-genomics org/plink2. 2015.

[pgen.1008780.ref048] 48.Lee SH, Van der Werf JH. MTG2: an efficient algorithm for multivariate linear mixed model analysis based on genomic information. Bioinformatics. 2016;32(9):1420–2. 10.1093/bioinformatics/btw012 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Using prior information from humans to prioritize genes and gene-associated variants for complex traits in livestock

Biaty Raymond

Loic Yengo

Roy Costilla

Chris Schrooten

Aniek C Bouwman

Ben J Hayes

Roel F Veerkamp

Peter M Visscher

Roles

Abstract

Author summary

Introduction

Results

Orthologous genes in humans and cattle are highly correlated in size but not in LD score or gene-effect on height

Height-associated genes in humans and cattle significantly overlap

Fig 1.

Fig 2. Distribution of gene-size (in cattle and humans) for the 77 orthologous genes that overlap with or are within 10kbp of the lead SNPs from Bouwman et al. [12].

Cattle stature-associated variants are enriched in human height-associated genes

Fig 3. Frequency distribution of the number of SNPs with near-independent significant effect on stature (P < 5x10-8) in 1,000 control tests for SMR-C genes.

Fig 4. Frequency distribution of the number of SNPs with near-independent significant effect on stature (P < 5x10-8) in 1,000 control tests for COJO-C genes.

Fig 5. Frequency distribution of the number of SNPs with near-independent significant effect on stature (P < 5x10-8) in 100 control tests.

Fig 6. Frequency distribution of the number of SNPs with near-independent significant effect on stature (P < 5x10-8) in 1,000 control tests for SMR-MT-C genes.

Accuracy of cattle variants selection using within- and across-species prior information

Fig 7. Frequency distribution of the percentage of genetic variance in New Zealand Holstein population, explained by SNPs in 1,000 control tests for SMR-C SNPs.

Fig 8. Frequency distribution of the percentage of genetic variance in New Zealand Holstein population, explained by SNPs in 1,000 control tests for COJO-C SNPs.

Discussion

Materials and methods

Data

Table 1. A summary of all data sets used in the study.

Comparison of gene characteristics in humans and cattle

Overlap of height-associated genes in humans and cattle

Enrichment for cattle stature GWAS signals in and around human height-associated genes

Accuracy of cattle variants selection using within- and across-species prior information

Supporting information

Data Availability

Funding Statement

References

Decision Letter 0

Gregory S Barsh

Jennie Elizabeth Pryce

Roles

Author response to Decision Letter 0

Decision Letter 1

Gregory S Barsh

Jennie Elizabeth Pryce

Roles

Acceptance letter

Gregory S Barsh

Jennie Elizabeth Pryce

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases