Abstract
Background Developing novel therapies for complex disease requires better understanding of the causal processes that contribute to disease onset and progression. Although trans-acting gene expression quantitative trait loci (trans-eQTLs) can be a powerful approach to directly reveal cellular processes modulated by disease variants, detecting trans-eQTLs remains challenging due to their small effect sizes and large number of genes tested. However, if a single trans-eQTL controls a group of co-regulated genes, then multiple testing burden can be greatly reduced by summarising gene expression at the level of co-expression modules prior to trans-eQTL analysis.
Results We analysed gene expression and genotype data from six blood cell types from 226 to 710 individuals. We inferred gene co-expression modules with five methods on the full dataset, as well as in each cell type separately. We detected a number of established co-expression module trans-eQTLs, such as the monocyte-specific associations at the IFNB1 and LYZ loci, as well as a platelet-specific ARHGEF3 locus associated with mean platelet volume. We also discovered a novel trans association near the SLC39A8 gene in LPS-stimulated monocytes. Here, we linked an early-response cis-eQTL of the SLC39A8 gene to a module of co-expressed metallothionein genes upregulated more than 20 hours later and used motif analysis to identify zinc-induced activation of the MTF1 transcription factor as a likely mediator of this effect.
Conclusions Our analysis provides a rare detailed characterisation of a trans-eQTL effect cascade from a proximal cis effect to the affected signalling pathway, transcription factor, and target genes. This highlights how co-expression analysis combined with functional enrichment analysis can greatly improve the identification and prioritisation of trans-eQTLs when applied to emerging cell-type specific datasets.
Background
Genome-wide association studies have been remarkably successful at identifying genetic variants associated with complex traits and diseases. To enable pharmacological and other interventions on these diseases, linking associated variants to causal intermediate phenotypes and processes is needed. A canonical example is the causal role of circulating LDL cholesterol in cardiovascular disease [1]. However, discovering clinically relevant intermediate phenotypes has so far remained challenging for most complex diseases. At the molecular level, cis-acting gene expression quantitative trait loci (cis-eQTLs) can be used to identify putative causal genes at disease-associated loci, but due to widespread co-regulation between neighbouring genes [2] and poor understanding of gene function, these approaches often identify multiple candidates whose functional relevance for the disease is unclear.
A promising approach to overcome the limitations of cis-eQTLs is trans-eQTL analysis linking disease-associated variants via signalling pathways and cellular processes (trans-acting factors) to multiple target genes. Although trans-eQTLs are widespread [3], most transcriptomic studies in various cell types and tissues are still underpowered to detect them [4]. This is due to limited sample sizes of current eQTL studies, small effect sizes of trans-eQTLs, and the large number of tests performed (>106 independent variants with >104 genes). To reduce the number of tested phenotypes, co-expression analysis methods are sometimes used to aggregate individual genes to co-expressed modules capturing signalling pathways and cellular processes [5]. Such approaches have been successful in identifying trans-eQTLs in yeast [6] as well as various human tissues [7–9] and purified immune cells [10,11].
Gene co-expression modules can be detected with various methods. Top-down matrix factorisation approaches such as independent component analysis (ICA) [12], sparse decomposition of arrays (SDA) [7] and probabilistic estimation of expression residuals (PEER) [13] seek to identify latent factors that explain large proportion of variance in the dataset. In these models, a single gene can contribute to multiple latent factors with different weights. In contrast, bottom-up gene expression clustering methods such as weighted gene co-expression network analysis (WGCNA) [14] seek to identify non-overlapping groups of genes with highly correlated expression values. Recently, both matrix factorisation and co-expression clustering methods have been further extended to incorporate prior information about biological pathways and gene sets, resulting in pathway-level information extractor (PLIER) [9] and funcExplorer [15], respectively. Out of these methods, ICA, WGCNA, SDA and PLIER have previously been used to find trans-eQTLs for modules of co-expressed genes [7–11], but only a single method at a time. However, since different methods solve distinct optimisation problems, they can detect complementary sets of co-expression modules [5]. Thus, applying multiple co-expression methods to the same dataset can aid trans-eQTL detection by identifying complementary sets of co-expression modules capturing a wider range of biological processes.
Another aspect that can influence co-expression module detection is how the data is partitioned prior to analysis [5]. This is particularly relevant when data from multiple cell types or conditions is analysed together. When co-expression analysis is performed across multiple cell types or conditions, then the majority of detected gene co-expression modules are guided by differential expression between cell types [16,17]. As a result, cell-type-specific co-expression modules can be missed due to weak correlation in other cell types [16]. One strategy to recover such modules is to perform co-expression analysis in each cell type separately [5].
In this study, we performed comprehensive gene module trans-eQTL analysis across six major blood cell types and three stimulated conditions from five published datasets. To maximise gene module detection, we applied five distinct co-expression analysis methods (ICA, PEER, PLIER, WGCNA, funcExplorer) to the full dataset as well as individual cell types and conditions separately. Using a novel aggregation approach based on statistical fine mapping, we grouped individual trans-eQTLs to a set of non-overlapping loci. Extensive follow-up with gene set and transcription factor motif enrichment analyses allowed us to gain additional insight into the functional impact of trans-eQTLs and prioritise loci for further analyses. In addition to replicating two known monocyte-specific trans-eQTLs at the IFNB1 [11,17–19] and LYZ loci [10,20,21], we found that the trans-eQTL at the ARHGEF3 locus detected in multiple whole blood datasets [8–10,22] was highly specific to platelets in our analysis. Finally, we also detected a novel association at the SLC39A8 locus that controlled a group of genes encoding zinc-binding proteins in LPS-stimulated monocytes.
Results
Cell types, conditions and samples
We used gene expression and genotype data from five previously published studies from three independent cohorts [18,20,23–25]. The data consisted of CD4+ and CD8+ T cells [23,24], B cells [20,23], neutrophils [23,25], platelets [23], naive monoctyes [18,23] and monocytes stimulated with lipopolysaccharide for 2 or 24 hours (LPS 2h, LPS 24h) and interferon-gamma for 24 hours (IFNγ 24h) [18]. The sample size varied from n = 226 in platelets to n = 710 in naive monocytes (Figure 1A). After quality control, normalisation and batch correction (see “Methods”), the final dataset consisted of 18,383 unique protein coding genes profiled in 3,938 samples from 1,037 unique genotyped individuals of European ancestries (Figure 1B). Even though the samples originated from five different studies, they clustered predominantly by cell type of origin (Figure 1B).
Detecting trans-eQTLs regulating modules of co-expressed genes
We performed co-expression analyses with ICA, WGCNA, PLIER, PEER and funcExplorer on the full gene expression dataset (integrated approach) as well as on each cell type and condition separately (separate approach) (Figure 1C). In total, we obtained 482 gene modules from the integrated approach and 3,509 from the separate clustering of different cell types (Figure 1D; Additional file 1: Figure S1). For every module, the methods inferred a single characteristic expression pattern (‘eigengene’) that represents the expression profiles of the module genes across the samples. Although implementation details varied between methods (see “Methods”), these eigengene profiles were essentially linear combinations of expression levels of genes belonging to the modules.
For trans-eQTL analysis, we included 6,861,056 common (minor allele frequency > 5%) genetic variants passing strict quality control criteria. First, we used linear regression implemented in MatrixEQTL [26] package to identify all genetic variants nominally associated (P-value < 5×10-8) with the eigengenes of each of the 3,991 co-expression modules detected across 9 cell types and conditions. We performed trans-eQTL analysis in each cell type and condition separately. Next, we used SuSiE [27] to fine map all significant associations to 864 independent credible sets of candidate causal variants (Figure 1E). Since we applied five co-expression methods to both integrated and cell type specific (separated) datasets, we found a large number of overlapping genetic associations. We thus aggregated overlapping credible sets from 864 associations to 601 non-overlapping genomic loci (Additional file 1: Figure S2; see “Methods”). We observed that some, especially smaller, co-expression modules were driven by strong cis-eQTL effects that were controlling multiple neighbouring genes in the same module. To exclude such effects, we performed gene-level eQTL analysis for 18,383 protein-coding genes and the 601 lead variants identified above. We excluded co-expression modules where the module lead variant was not individually associated with any of the module genes in trans (> 5Mb away) (see “Methods”). This step reduced the number of nominally significant trans-eQTL loci to 303 (Figure 1F; Additional file 2). Finally, to account for the number of co-expression modules tested, we used both Benjamini-Hochberg false discovery rate (FDR) and Bonferroni correction (see “Methods”). The FDR 10% threshold reduced the number of significant associations to 140 and Bonferroni threshold retained only 4 significant loci, including loci near IFNB1 (Additional file 1: Figure S3) and LYZ (Additional file 1: Figure S4) genes that have been previously reported in several other studies [10,17–21] (Additional file 1: Table S1). While the strong trans-eQTL signals at the IFNB1 and LYZ loci were detected by all co-expression methods in both integrated and separate analyses, most associations were detected by only a subset of the analytical approaches (Additional file 2). Furthermore, almost all trans-eQTLs that we detected had highly cell type and condition specific effects (Additional file 1: Figure S5). We will now dissect two such cell type and condition specific associations in more detail.
Platelet specific trans-eQTL at the ARHGEF3 locus is associated with multiple platelet traits
We found that the rs1354034 (T/C) variant located within the ARHGEF3 gene is associated with three co-expression modules in platelets: one ICA module detected in integrated analysis (IC68, 1,074 genes) and two co-expression modules detected in a platelet-specific analysis by PLIER (X6.WIERENGA_STAT5A_TARGETS_DN, 918 genes) and funcExplorer (Cluster_12953, 5 genes) (Figure 2B, Additional file 1: Figure S6). The T allele increases the expression of the ARHGEF3 gene in cis and the two lead variants are the same (Figure 2A). Furthermore, both the cis and trans-eQTLs colocalise with a GWAS hit for mean platelet volume (cis PP4 = 0.99, trans PP4 > 0.99 for all modules), platelet count (cis PP4 = 0.99, trans PP4 > 0.99 for all modules) and plateletcrit (trans PP4 > 0.99 for all modules) (Figure 2A) [28]. Interestingly, ARHGEF3 itself is not in any of the three modules and the module eigengenes are not strongly co-expressed with ARHGEF3 (Pearson’s r ranging from 0.07 to 0.33 in platelets). While IC68 and X6.WIERENGA_STAT5A_TARGETS_DN share 74 overlapping genes (one-sided Fisher’s exact test P-value = 0.003), none of the genes in Cluster_12953 is in any of the other modules.
Although the ARHGEF3 trans-eQTL has been detected in multiple whole blood trans-eQTL studies [3,8,9,22] (Additional file 1: Table S1), our analysis demonstrates that this association is highly specific to platelets and not detected in other major blood cell types (Figure 2B). Furthermore, even though ARHGEF3 is expressed in multiple cell types, the cis-eQTL effect is also only visible in platelets (Additional file 1: Figure S6). Reassuringly, the trans-eQTL effect sizes in our small platelet sample (n = 216) are correlated (Pearson’s r = 0.68, P-value = 5.1×10-12) with the effects from the largest whole blood trans-eQTL meta-analysis [3] (n = 31,684) (Additional file 1: Figure S7). The platelet specificity of the ARHGEF3 association is further supported by functional enrichment analysis with g:Profiler [29] which found that both the PLIER module X6.WIERENGA_STAT5A_TARGETS_DN and target genes from the gene-level analysis were strongly enriched for multiple terms related to platelet activation (Figure 2E; https://biit.cs.ut.ee/gplink/l/rtyNe5M2R4). Cluster_12953, however, was enriched for cellular response to iron ion, suggesting that ARHGEF3 might be involved in multiple independent processes [9,30]. Altogether, these results demonstrate how a trans-eQTL detected in whole blood can be driven by a strong signal present in only one cell type.
SLC39A8 locus is associated with zinc ion homeostasis in LPS-stimulated monocytes
One of the novel results in our analysis was a locus near the SLC39A8 gene that was associated (P-value = 1.2×10-9) with a single co-expression module detected by funcExplorer (Cluster_10413) in monocytes stimulated with LPS for 24 hours (Figure 3A-C). The module consisted of five metallothionein genes (MT1A, MT1F, MT1G, MT1H, MT1M) all located in the same locus on chromosome 16 (Figure 3D). Although the trans-eQTL lead variant (rs75562818) was significantly associated with the expression of the SLC39A8 gene (Figures 3A and 3D), the two association signals did not colocalise and the credible sets did not overlap (Figure 3A; Additional file 1: Figure S8), indicating that the cis-eQTL detected in naive and stimulated monocytes in our dataset is not the main effect driving the trans-eQTL signal. Furthermore, the expression of SLC39A8 was only moderately correlated with the eigengene value of Cluster_10413 (Pearson’s r = 0.27). Since SLC39A8 is strongly upregulated (log2 fold-change = 3.53) in response to LPS already at two hours (Figure 4A), we speculated that there might be a transient eQTL earlier in the LPS response. To test this, we downloaded the eQTL summary statistics from the Kim-Hellmuth et al, 2017 study that had mapped eQTLs in monocytes stimulated with LPS for 90 minutes and six hours [31]. Indeed, we found that the cis-eQTL 90 minutes after LPS stimulation colocalised with our trans-eQTL (Figure 3A) and this signal disappeared by six hours after stimulation (Additional file 1: Figure S9).
To understand the function of the SLC39A8 locus, we turned to the target genes. Gene-level analysis identified two more metallothionein genes (MT1E and MT1X) from the same locus as likely target genes (Figure 3D). Enrichment analysis with g:Profiler revealed that these genes were enriched for multiple Gene Ontology terms and pathways related to zinc ion homeostasis (Figure 3E, full results at https://biit.cs.ut.ee/gplink/l/v7FhJRn4Rj). Furthermore, the promoter regions of the 7 genes were also enriched for the binding motif of the metal transcription factor 1 (MTF1) transcription factor (P-value = 2.1×10-4, Figure 3E). Taken together, these results suggest that a transient eQTL of the SLC39A8 gene 90 minutes after stimulation regulates the expression of 7 zinc binding proteins 24 hours later. Multiple lines of literature evidence support this model (Figure 4B). First, the ZIP8 protein coded by the SLC39A8 gene is a manganese and zinc ion influx transporter [32]. Secondly, SLC39A8 is upregulated by the NF-κB transcription factor in macrophages and monocytes in response to LPS and this upregulation leads to increased intracellular Zn2+ concentration [33]. Third, Zn2+ influx increases the transcriptional activity of the metal transcription factor 1 (MTF1) [34] and metallothioneins, which act as Zn2+-storage proteins, are well known target genes of the MTF1 transcription factor [35]. Finally, SLC39A8 knockdown in mice leads to decreased expression of the metallothionein 1 (MT1) gene [33].
To see if the SLC39A8 trans-eQTL might be associated with any higher level phenotypes, we queried the GWAS Catalog [36] database with the ten variants from the trans-eQTL 95% credible set. We found that a lead variant for red blood cell distribution width (rs7692921) was one of the variants in our credible set and in high LD (r2 = 0.991) with the trans-eQTL lead variant (Figure 4C) [37]. However, neither of the eQTL variants was in LD with a known missense variant (rs13107325) in the SLC39A8 gene that has been associated with schizophrenia, Parkinson’s disease and other traits (Figure 4C) [38].
Discussion
Given that trans-eQTLs have been more difficult to replicate between studies and false positive associations can easily occur due to technical issues [40,41], it is increasingly important to effectively summarise and prioritise associations for follow up analyses and experiments. We found that aggregation of credible sets of eigengene profiles from multiple co-expression methods (Additional file 1: Figure S2) successfully reduced the number of independent associations, but this still retained more than 300 loci that we needed to evaluate. To further prioritise associations, we used gene set and transcription factor motif enrichment analysis of the trans-eQTL target genes. Although motif analysis is often underpowered, it can provide directly testable hypotheses about the trans-eQTL mechanism such as the MTF1 transcription factor that we identified at the SLC39A8 locus. Similar approaches have also been successfully used to characterise trans-eQTLs involving IRF1 and IRF2 transcription factors [18,42].
A major limitation of co-expression based approach for trans-eQTL mapping is that many true co-expression modules can remain undetected by various co-expression analysis methods. We sought to overcome this by aggregating results across five complementary co-expression methods. We found that while all methods were able to discover strong co-expression module trans-eQTLs such as those underlying the IFNB1 (Additional file 1: Figure S3) and LYZ (Additional file 1: Figure S4) associations, most co-expression module trans-eQTLs were only detected by a subset of the analysis methods. For example, the ARHGEF3 association was detected by three of the five methods (Figure 2B) and SLC39A8 co-expression module was found only by funcExplorer and only when samples from LPS-stimulated monocytes were analysed separately (Figure 3B). While it is always possible to include additional analysis methods, this should be appropriately weighed against the increase in the number of phenotypes tested. In our analysis, we decided to first use a relaxed nominal significance threshold of P-value < 5×10-8 and subsequently focus on associations that we could either replicate in independent datasets or find significant support from the literature. Finally, if the trans-eQTL locus controls a single or a small number of genes then co-expression-based approaches are probably not well suited to detect such associations and gene-level analysis is still required.
Since eQTL datasets from purified cell types are still relatively small and single cell eQTL datasets are even smaller [43], it is tempting to perform trans-eQTL analysis on whole tissue datasets such as the brain or whole blood [3]. However, it remains unclear what fraction of cell type and condition specific trans-eQTLs can be detected in whole tissue datasets collected from healthy donors. Although we were able to replicate the ARHGEF3 association in the eQTLGen whole blood meta-analysis, because our fine mapped lead variant happened to be one of the 10,317 variants tested in eQTLGen, systematic replication requires genome-wide summary statistics that are currently lacking for trans-eQTL analyses. Secondly, tissue datasets can be biased by cell type composition effects. These can lead to spurious trans-eQTL signals, because genetic variants associated with cell type composition changes would appear as trans-eQTLs for cell-type-specific genes [3]. Furthermore, multiple studies have demonstrated that also the co-expression signals in tissues are largely driven by cell type composition effects [44–46]. Thus, even though PLIER detected the ARHGEF3 trans-eQTL in whole blood, this could have been at least partially driven by the change in platelet proportion between individuals [9]. Our analysis in purified cell types enabled us to verify that this was a truly platelet-specific genetic association.
Although both in the case of ARHGEF3 and SLC39A8, we could be reasonably confident that the expression level of the cis gene mediated the observed trans-eQTL effect, there was only a modest correlation (Pearson’s r between 0.07 and 0.33) between the cis gene expression and the corresponding trans co-expression module expression. In case of SLC39A8 there seemed to be a temporal delay with the cis-eQTL being active early in LPS response and trans-eQTL appearing much later after proposed accumulation of the ZIP8 protein and increase in intracellular zinc concentration. Temporal delay has also been reported for the trans-eQTLs at the INFB1 [18] and IRF1 [42] loci. Similarly, stimulation-specific cis-eQTL at CR1 locus in whole blood is not able to explain the full extent of the trans-eQTL observed at the same locus [47]. This suggests that if cis and trans effects are separated from each other either in time (early versus late response) or space (different cell types that interact with each other), then this might limit the power of methods that rely on genetically predicted gene expression levels to identify regulatory interactions [22,48,49] and infer causal models. This can also have a negative impact on mediation analysis [50–52], which seeks to estimate the proportion of trans-eQTL variance explained by the expression level of the cis gene. Altogether, these results indicate that limiting trans-eQTL analysis to missense variants and to variants that have been detected as cis-eQTLs in the same cell type might miss some true associations, because the cis effect might be active in some other, yet unprofiled, context.
Conclusions
We have performed a large-scale trans-eQTL analysis in six blood cell types and three stimulated conditions. We demonstrate that co-expression module detection combined with gene set enrichment analysis can help to identify interpretable trans-eQTLs, but these results depend on which co-expression method is chosen for analysis and how the input data is partitioned beforehand. We find that the detected trans-eQTLs are highly cell type specific and we use these approaches to perform in-depth characterisation of two cell type specific trans-eQTL loci: platelet-specific trans-eQTL near the ARHGEF3 gene and monocyte-specific associations near the SLC39A8 locus. In both cases, the co-expression modules were enriched for clearly interpretable Gene Ontology terms and pathways, which directly guided literature review and more detailed analyses. We believe that applying co-expression and gene set enrichment based approaches to larger eQTL datasets has the power to detect many more additional associations while simultaneously helping to prioritise trans-eQTLs for detailed experimental or computational characterisation.
Methods
Datasets used in the analysis
CEDAR
The CEDAR dataset [23] contained gene expression and genotype data from CD4+ T-cells, CD8+ T-cells, CD19+ B-cells and CD14+ monocytes, CD15+ neutrophils and platelets from up to 323 individuals. The raw gene expression data generated with Illumina HumanHT-12 v4 arrays were downloaded from ArrayExpress [53] (accession E-MTAB-6667). The raw IDAT files were imported into R using the readIdatFiles function from the beadarray v2.28 [54] Bioconductor package.
The raw genotype data generated by Illumina HumanOmniExpress-12 v1_A genotyping arrays were also downloaded from ArrayExpress (accession E-MTAB-6666). Genotype calling was performed with Illumina GenomeStudio v2.0.4, after which the raw genotypes were exported in PLINK format.
Kasela, 2017
Kasela et al, 2017 [24] generated gene expression and genotype data from CD4+ and CD8+ T cells from 297 unique donors. The raw gene expression data generated with Illumina HumanHT-12 v4 arrays were downloaded from Gene Expression Omnibus (accession GSE78840). The genotype data generated by Illumina HumanOmniExpress-12 v1_A genotyping arrays were obtained from the Estonian Genome Center, University of Tartu (https://www.geenivaramu.ee/en/biobank.ee/data-access). Ethical approval was obtained from the Research Ethics Committee of the University of Tartu (approval 287/T-14).
Fairfax, 2012, Fairfax, 2014 and Naranbhai, 2015
Fairfax et al, 2012 [20] profiled gene expression in CD19+ B cells from 282 individuals (ArrayExpress accession E-MTAB-945). Fairfax et al, 2014 [18] profiled gene expression in naive CD14+ monocytes as well as in cells stimulated with lipopolysaccharide (LPS) for 2 or 24 hours and interferon-gamma for 24 hours from up to 414 individuals (accession E-MTAB-2232). Naranbhai et al., 2015 [25] profiled gene expression in CD15+ neutrophils from 93 individuals (accession E-MTAB-3536). The genotype data for all three studies were generated by Illumina HumanOmniExpress-12 genotyping arrays and were downloaded from European Genome-phenome Archive (accessions EGAD00010000144 and EGAD00010000520).
Genotype data quality control and imputation
We started with raw genotype data from each study in PLINK format and GRCh37 coordinates. Before imputation, we performed quality control independently on each of the three datasets. Briefly, we used Genotype harmonizer [55] v1.4.20 to align the alleles with the 1000 Genomes Phase 3 reference panel and exclude variants that could not be aligned. We used PLINK v1.9.0 to convert the genotypes to VCF format and used the fixref plugin of the bcftools v1.9 to correct any strand swaps. We used ‘bcftools norm --check-ref x’ to remove any remaining variants where the reference allele did not match the GRCh37 reference genome. Finally, we excluded variants with Hardy-Weinberg equilibrium P-value > 10-6, missingness > 0.05 and MAF < 0.01. We also excluded samples with more than 95% of the variants missing. Finally, we merged genotype data from all three studies into a single VCF file.
After quality control, we included 580,802 autosomal genetic variants from 1,041 individuals for imputation. We used a local installation of the Michigan Imputation Server v1.2.1 [56] to perform phasing and imputation with EAGLE v2.4 [57] and Minimac4 [56]. After imputation, we used CrossMap.py v2.8.0 [58] to convert genotype coordinates to GRCh38 reference genome. We used bcftools v1.9.0 to exclude genetic variants with imputation quality score R2 < 0.4 and minor allele frequency (MAF) < 0.05. We used PLINK [59] v1.9.0 to perform LD pruning of the genetic variants and LDAK [60] to project new samples to the principal components of the 1000 Genomes Phase 3 reference panel [61]. The Nextflow pipelines for genotype processing and quality control are available from GitHub (https://github.com/eQTL-Catalogue/genotype_qc).
Detecting sample swaps between genotype and gene expression data
We used Genotype harmonizer [55] v1.4.20 to convert the imputed genotypes into TRITYPER format. We used MixupMapper [62] v1.4.7 to detect sample swaps between gene expression and genotype data. We detected 155 sample swaps in the CEDAR dataset, most of which affected the neutrophil samples. We also detected one sample swap in the Naranbhai, 2015 dataset.
Gene expression data quality control and normalisation
As a first step, we performed multidimensional scaling (MDS) and principal component analysis (PCA) on each dataset separately to detect and exclude any outlier samples. This was done after excluding the replicate samples and the samples that did not pass the genotype data quality control. Additional outliers were detected after quantile normalisation and adjusting for batch effects. The normalisation was performed using the lumiN function from the lumi v.2.30.0 R package [63]. Batch effects, where applicable, were adjusted for with the removeBatchEffect function from the limma v.3.34.9 R package [64]. After quality control to exclude outlier samples, the quantile normalised log2 intensity values from all datasets were combined. This was followed by regressing out dataset specific batch effects. Only the intensities of 30,353 protein-coding probes were used. Finally, the probe sets were mapped to genes. For genes with more than one corresponding probe set, the probe with the highest average expression was used. 18,383 protein-coding genes with unique Ensembl identifiers remained for co-expression analysis. We did not regress out any principal components from the gene expression data, as this can introduce false positives in trans-eQTL analysis due to collider bias [41]. In total, 3,938 samples remained after the quality control (Table 1).
Co-expression analysis
We applied five different methods to identify modules of co-expressed genes from the gene expression data. We used an expression matrix where rows correspond to genes and columns to individuals/samples as input for the methods. The gene expression profiles were centred and standardised prior to analysis. All methods infer gene co-expression modules, each of which can be described by a single expression profile (‘eigengene’) that captures the collective behavior of corresponding genes in the module. The approaches for defining eigengenes differ across the methods (see below). These eigengenes are treated as quantitative traits in the trans-eQTL analysis. To detect potential cell type and condition specific modules, we applied the same methods also to the expression matrices from each of the nine cell types and conditions separately. Summaries of the co-expression analysis results from both integrated and cell-type-specific expression data are shown in Additional file 1: Figure S1.
Co-expression clustering methods
Weighted gene co-expression network analysis (WGCNA)
The WGCNA method [14] identifies non overlapping co-expressed gene modules. Each of the modules is represented by its first principal component of expression values of genes in the module termed as module eigengene. We used the function blockwiseModules for automatic block-wise network construction and module identification with default parameters from the dedicated R package WGCNA (v.1.66). The number of modules was detected automatically by the algorithm.
funcExplorer
FuncExplorer [15] is a web tool that performs hierarchical clustering on gene expression values which is followed by automated functional enrichment analysis to derive the most biologically meaningful gene modules from the dendrogram. The expression data were uploaded to funcExplorer and the modules were detected using the following parameters: best annotation strategy, P-value threshold 0.01 for enrichment of Gene Ontology, KEGG and Reactome annotations. Every funcExplorer gene module is characterised by the eigengene profile which, like in WGCNA, is the first principal component of module expression values calculated in the same way as in WGCNA. The number of modules is detected automatically by funcExplorer and the different modules consist of non-overlapping sets of genes. The co-expression analysis results are available for browsing from https://biit.cs.ut.ee/funcexplorer/user/2a29dfa6de6b8b733f665352735adaf5 where the option ‘Dataset’ includes the full selection of expression data used in this analysis. Dataset ‘Merged_ENSG_expression’ incorporates integrated samples from all cell types and conditions, ‘CL_0000233_naive’ stands for platelets, ‘CL_0000236_naive’ for B-cells, ‘CL_0000624_naive’ for CD4+ T-cells, ‘CL_0000625_naive’ for CD8+ T-cells, ‘CL_0000775_naive’ for neutrophils, ‘CL_0002057_naive’ for monocytes and ‘CL_0002057_IFNg_24h’, ‘CL_0002057_LPS_24h’, ‘CL_0002057_LPS_2h’ include gene expression matrices from corresponding stimulated monocyte samples.
Matrix factorisation methods
Matrix factorisation methods, such as ICA, PLIER and PEER, deconvolve the input gene expression matrix into two related matrices [5]. One of the matrices is the matrix of factor loadings for each sample and the other describes the gene-level weights of the factors. In the case of ICA, PLIER and PEER, we used the factor loadings as module eigengene profiles. For g:Profiler enrichment analysis we used the gene-level weights to define the genes that characterise the modules by choosing the ones that are the most influenced, i.e. the genes at both extremes of gene weight values (2 standard deviations from the mean weights in this module). Thus, different modules can include overlapping sets of genes.
Independent component analysis (ICA)
The ICA [12] method attempts to decompose gene expression measurements into independent components (factors) which represent underlying biological processes. The fastICA [65] algorithm in R was run using the wrapper package picaplot v.0.99.7 (https://github.com/jinhyunju/picaplot). The number of components to be estimated was automatically detected by the implementation using a 70% variance cut-off value. The ICA algorithm was run 15 times and only the components that replicated in every run were used.
Pathway-level information extractor (PLIER)
PLIER [9] is a matrix decomposition method that uses prior biological knowledge of pathways and gene sets to deconvolve gene expression profiles as a product of a small number of latent variables (factors) and their gene weights. We performed PLIER analysis using the dedicated R package (v.0.99.0; downloaded from https://github.com/wgmao/PLIER) with the collection of 5,933 gene sets available in the package comprising canonical, immune and chemgen pathways from MSigDB [66], and various cell-type markers from multiple sources. PLIER was run with 100 iterations. Only the 16,440 genes appearing in both gene expression data and the pathway annotation matrix were used. The initial number of latent variables was set using the num.pc function provided by the package.
Probabilistic estimation of expression residuals (PEER)
PEER [13,67] is a factor analysis method that uses Bayesian approaches to infer hidden factors from gene expression data that explain a large proportion of expression variability. We applied PEER method for co-expression analysis using the peer R package (v.1.0; downloaded from https://github.com/PMBio/peer) with default parameters, accounting also for the mean expression. The initial number of factors was determined using the num.pc function from the PLIER package.
Functional enrichment analysis
We used the g:GOSt tool from the g:Profiler toolset [29] via dedicated R package gprofiler2 (v.0.1.8) for functional enrichment analysis of gene modules. The short links to the full enrichment results were automatically generated using the parameter as_short_link = T in the function gost. The results shown in this paper were obtained with data version e99_eg46_p14_55317af.
Cis-eQTL analysis and fine mapping
We performed cis-eQTL analysis using the qtlmap (https://github.com/eQTL-Catalogue/qtlmap) Nextflow [68] workflow developed for the eQTL Catalogue project [69]. Briefly, we performed cis-eQTL analysis in a +/- 1 Mb window centered around each gene. We used the first six principal components of both the gene expression and genotype data as covariates in the analysis. The eQTL analysis was performed using QTLtools [70].
For cis-eQTL fine mapping, we used the Sum of Single Effects (SuSiE) model [27] implemented in the susieR v0.9.0 R package. We performed fine mapping on a +/- 1 Mb cis window centered around the lead eQTL variant of each gene. We performed fine mapping on individual-level genotype and gene expression data. Prior to fine mapping, we regressed out six principal components of the gene expression and genotype data from the gene expression data. To identify significant eQTLs for QTL mapping, we performed Bonferroni correction for each gene to account for the number of variants tested per gene and then used Benjamini-Hochberg FDR correction to identify genes with FDR < 0.1. The fine mapping Nextflow workflow for cis-eQTLs is available from GitHub (https://github.com/kauralasoo/susie-workflow).
Gene module trans-eQTL analysis and fine mapping
The MatrixEQTL [26] R package (v2.2) was used for trans-eQTL analysis to fit a linear model adjusted for sex, batch (where available) and the first three principal components of the genotype data. Before the analysis, the module eigengene profiles were transformed using the inverse normal transformation to reduce the impact of outlier eigengene values produced by some clustering methods. A total of 6,861,056 autosomal genetic variants with minor allele frequency (MAF) > 0.05 were tested. Due to the partial sharing of individuals between cell types and conditions, the eQTL analysis was performed in each cell type and condition separately. To achieve this, the eigenvectors from the integrated approach were split into cell type specific sub-eigenvectors before the analysis. The results from every analytical setting (data partitioning approach (n = 2), co-expression method (n = 5), cell type (n = 9), 90 trans-eQTL analyses in total), were then individually filtered to keep nominally significant variant-module associations (P-value < 5×10-8).
Next, we applied SuSiE [27] to fine map the nominally significant associations to independent credible sets of variants. For every gene module, we started fine mapping from the lead variant (variant with the smallest association P-value for this module) and used a +/- 500,000 bp window around the variant to detect the credible sets. We continued fine mapping iteratively with the next best nominally significant variant outside the previous window to account for LD and continued this process until no variants remained for the gene module. This procedure resulted in a total of 864 credible sets across all cell types, co-expression analysis methods and data partitioning approaches (integrated and separate).
To aggregate and summarise overlapping associations, we combined all credible sets into an undirected graph where every node represents a credible set of a module from a triplet (data partitioning approach, co-expression method, cell type) and we defined an edge between two nodes if the corresponding credible sets shared at least one overlapping variant (Additional file 1: Figure S2). The graph was constructed using the igraph R package. After obtaining the graph, we searched for connected components, i.e. subgraphs where every credible set is connected by a path, to combine the vast number of results into a list of non-overlapping loci (n = 601). For every component we defined the lead variant by choosing the intersecting variant with the largest average posterior inclusion probability (PIP) value across all the credible sets in the component.
Genes in physical proximity often have correlated expressions levels and could thus manifest as co-expression modules in our analysis. Consequently, if one or more genes in such modules have cis-eQTLs, then these cis variant-module associations would also be detected by our approach. To differentiate cis-acting co-expressions module eQTLs from true trans associations, we decided to add an additional filtering step based on gene-level analysis. We performed gene-level eQTL analysis for individual gene expression traits of the 18,383 protein-coding genes and the 601 lead variants. The gene-level eQTL analysis was performed using the MatrixEQTL R package with the same settings and data transformations as in the module-level analysis described above. From every credible set component, we excluded the variant-module pairs together with corresponding credible sets where no trans associations (variant-level Benjamini-Hochberg FDR 5%) were included in the module. As trans-eQTLs we consider variants that act on distant genes (> 5 Mb away from the lead variant) and genes residing on different chromosomes. After this filtering step we repeated the process of aggregating credible sets and 303 non-overlapping loci remained (Additional file 1: Figure S10; Additional file 2).
To further account for the number of co-expression modules tested, we applied both Benjamini-Hochberg false discovery rate (FDR) and Bonferroni correction at the level of each analytical setting (Additional file 1: Figure S10). We applied the FDR 10% threshold to every module - lead variant pair from each of the 90 analytical settings (data partitioning approach, co-expression analysis method, cell type) and if a pair did not pass the threshold we excluded it together with the corresponding credible set(s) from the results. Bonferroni correction was applied in a similar manner with a threshold P-value , where ni,i, = 1,…,90, stands for the number of modules from the corresponding co-expression method and data partitioning approach. We repeated the graph-based aggregation process on the remaining credible sets individually from both correction methods and as a result the FDR 10% threshold reduced the number of significant associations to 140 and Bonferroni threshold to only 4 significant trans-eQTLs.
Colocalisation
We downloaded GWAS summary statistics for 36 blood cell traits [28] from the NHGRI-EBI GWAS Catalog [36]. We downloaded coloc [71] R package v3.1 from bioconda [72]. The cis-eQTL colocalisation Nextflow workflow is available from GitHub (https://github.com/kauralasoo/colocWrapper). The same workflow was adjusted for trans-eQTL colocalisation.
Replication of genetic associations
Kim-Hellmuth et al, 2017 [31] profiled gene expression in monocytes before and after stimulation with LPS, muramyl-dipeptide (MDP) and 5′-triphosphate RNA for 90 minutes and 6 hours. We downloaded the eQTL summary statistics from ArrayExpress [53] (accession E-MTAB-5631). Individual-level genotype data were not available for this study. The eQTLGen Consortium [3] performed trans-eQTL analysis for 10,317 trait-associated genetic variants in 31,684 whole blood samples. We downloaded the summary statistics from https://www.eqtlgen.org/trans-eqtls.html.
Declarations
Ethics approval and consent to participate
Gene expression and genotype data from the CEDAR study were available for download without restrictions from ArrayExpress. For the Fairax_2012, Fairfax_2014 and Naranbhai_2015 studies we applied for access via the relevant Data Access Committee. For the Kasela_2017, we obtained approval from the Data Access Committee of the Estonian Biobank. Ethical approval for the project was obtained from the Research Ethics Committee of the University of Tartu (approval 287/T-14).
Consent for publication
Not applicable.
Availability of data and materials
The gene expression matrix, detected gene modules, eigenvectors and trans-eQTL credible sets are available in Zenodo (https://doi.org/10.5281/zenodo.3759693). The analysis source code has been deposited to the GitHub repository https://github.com/liiskolb/coexpression-transEQTL.
Competing interests
The authors declare that they have no competing interests.
Funding
LK and HP were supported by the Estonian Research Council grant PSG59. KA was supported by the European Regional Development Fund and the programme Mobilitas Pluss (MOBJD67). KA also received funding from the European Union’s Horizon 2020 research and innovation programme (grant number 825775) and Estonian Research Council (grants IUT34-4 and PSG415). KA, LK and NK were also supported by the Estonian Centre of Excellence in ICT Research (EXCITE) funded by the European Regional Development Fund.
Authors’ contributions
LK and KA designed the study. LK performed gene expression data quality control and normalisation, co-expression analysis and trans-eQTL analysis. KA performed genotype data quality control and imputation and cis-eQTL analysis. KA and NK wrote Nextflow workflows for the cis-eQTL analysis. KA and HP supervised the research. LK and KA wrote the manuscript with input from all co-authors.
Supplementary information
Additional file 1. Supplementary Figures S1-S10 and Table S1 (PDF)
Additional file 2. Co-expression trans-eQTL analysis results together with links to g:Profiler enrichment analysis (XLSX)
Additional file 1
Acknowledgements
We would like to thank Urmo Võsa, Silva Kasela, Kaido Lepik and Sina Rüeger for helpful comments on the manuscript. The computational analyses were performed at the High Performance Computing Center, University of Tartu.