Impact of natural genetic variation on gene expression dynamics

Abstract

DNA sequence variation causes changes in gene expression, which in turn has profound effects on cellular states. These variations affect tissue development and may ultimately lead to pathological phenotypes. A genetic locus containing a sequence variation that affects gene expression is called an "expression quantitative trait locus" (eQTL). Whereas the impact of cellular context on expression levels in general is well established, a lot less is known about the cell-state specificity of eQTL. Previous studies differed with respect to how "dynamic eQTL" were defined. Here, we propose a unified framework distinguishing static, conditional and dynamic eQTL and suggest strategies for mapping these eQTL classes. Further, we introduce a new approach to simultaneously infer eQTL from different cell types. By using murine mRNA expression data from four stages of hematopoiesis and 14 related cellular traits, we demonstrate that static, conditional and dynamic eQTL, although derived from the same expression data, represent functionally distinct types of eQTL. While static eQTL affect generic cellular processes, non-static eQTL are more often involved in hematopoiesis and immune response. Our analysis revealed substantial effects of individual genetic variation on cell type-specific expression regulation. Among a total number of 3,941 eQTL we detected 2,729 static eQTL, 1,187 eQTL were conditionally active in one or several cell types, and 70 eQTL affected expression changes during cell type transitions. We also found evidence for feedback control mechanisms reverting the effect of an eQTL specifically in certain cell types. Loci correlated with hematological traits were enriched for conditional eQTL, thus, demonstrating the importance of conditional eQTL for understanding molecular mechanisms underlying physiological trait variation. The classification proposed here has the potential to streamline and unify future analysis of conditional and dynamic eQTL as well as many other kinds of QTL data.

Figure 1. eQTL classification.

Schematic representation of static, conditional and dynamic eQTL. For the sake of simplicity only two conditions are considered, but the concept is extensible to any number of cell types. The top part of each panel shows in which condition the eQTL influences a gene's expression (A, B) or if it affects expression changes between cell types (C). The lower parts of the panels show exemplary mRNA expression profiles of the gene in six samples. The genotype of the eQTL in each sample is indicated by the color, assuming homozygous diallelic markers. A A static eQTL impacts expression in all cell types. The ranking of gene expressions per genotype is the same in all conditions, as is the slope of expression change between cell types. B A conditional eQTL influences gene expression in only one of the two conditions. Thus, gene expression is a function of genotype in one cell type but not in the other. The slopes of expression changes may or may not be dependent on the genotype at the eQTL. C A dynamic eQTL drives expression changes between cell types. This implies that the slopes of expression changes between conditions are dependent on the genotype at the eQTL.

Figure 2. Simultaneous eQTL mapping.

Schematic of simultaneous eQTL mapping for two cell types. This approach combines the available information from the two cell types (red and green) in one eQTL analysis. To this end, the gene expressions measured in the different conditions are combined into one vector . Similarly, for each condition the genotype matrix is subset to all samples for which there are expression measurements in this cell type. The resulting two submatrices and are concatenated into one genotype matrix. In order to discriminate static and conditional eQTL, two additional predictors indicating the cell type from which a sample was derived, are added to the predictor matrix. The combined genotype and cell type indicator matrix is used to find the model which best predicts gene expression simultaneously in all conditions.

Figure 3. Number of cell types in which eQTL are active.

The bars show the number of eQTL conditional in one, two, three or four cell types. Results are obtained from post-hoc Wald tests in the linear model comprising the eQTL marker, the cell type and their interaction. Only models with a significant marker - cell type interaction are considered. eQTL that are conditionally active in exactly one cell type are further classified by cell type (S - stem, P - progenitor, E - erythroid and M - myeloid cells).

Figure 4. Venn diagram for the overlap between static, conditional and dynamic eQTL.

Static and conditional eQTL were obtained from the simultaneous eQTL mapping (red circles). Cell type-specific eQTL (eQTL that are detected in exactly one cell type) are shown as a subgroup of conditional eQTL (dark red circle). Dynamic eQTL were derived from mapping expression differences between pairs of cell types (black circle). Results are summarized over the three cell type transitions that were analyzed (S-P, P-E, P-M).

Figure 5. Number of cis- and trans-eQTL in different eQTL classes.

Numbers of significant eQTL with shown separately for cis-eQTL (left) and trans-eQTL (right). Static, conditional and dynamic eQTL are distinguished (see labels at the bottom). Further, the figure discriminates simultaneous and separate eQTL mappings, which represent alternative ways for distinguishing static and conditional eQTL. Simultaneous mapping increases the statistical power leading to substantially more eQTL significant at the same level (). Even though both, cis- and trans-eQTL are increased when performing simultaneous mapping, trans-eQTL benefit more from the increase in power. See main text for exact definitions of the various eQTL types.

Figure 6. Comparison of eQTL analyses.

The bars compare fractions of different eQTL classes obtained in the original study by with our study. The leftmost bars show fractions of static and conditional eQTL, fractions of cis- and trans-eQTL are shown in the center. The rightmost bars compare fractions of cell type-specific eQTL in the four hematopoietic lineages (color scheme as in Figure 3).

Figure 7. Simultaneous eQTL map.

Each dot represents an eQTL - target gene pair, where physical marker positions are shown on the x-axis, gene positions on the y-axis. Significant static eQTL () are shown in gray, cell type-specific eQTL ( in exactly one cell type) are shown in the color scheme of Figure 3. Red triangles indicate two cell type-specific eQTL-rich regions (eQRR).

Figure 8. Examples of static, conditional and dynamic eQTL.

mRNA expression profiles of four exemplary genes over the four hematopoietic cell types (S - stem cells, P - myeloid progenitor cells, E - erythroid cells, M - myeloid cells). The colors represent the genotype at the eQTL marker (blue - B allele, red - D allele). Significant static eQTL are shown by a rectangle around the differentiation scheme, significant conditional and dynamic eQTL by the black color of the respective cell type letter or cell type transition arrow. A, Prdx2 is affected by a static eQTL in all four cell types. B, Sirt2 is influenced by a conditional eQTL in erythroid cells. C, the transition of Il12rb2 expression from progenitor to myeloid cells is driven by a dynamic eQTL. The expression of Il12rb2 increases in samples carrying the B allele at the eQTL, while it remains constant in samples carrying the D allele. D, the expression of Gadd45gip1 is conditionally affected in three of the four cell types (S, P and M) by an eQTL which at the same time also influences the gene's expression changes during the differentiation from progenitors to the erythroid and myeloid lineages.

Figure 9. GO enrichment for eQTL classes.

We tested for the enrichment of GO categories among eQTL loci and target genes in the different eQTL classes, separately for different cell types and transitions. Examples of enriched functional categories for cell type-specific and dynamic eQTL are shown next to the corresponding cell types or cell type transitions. Important GO categories that were enriched in static eQTL and their targets are shown outside the box. Terms that are significantly enriched () among eQTL loci are shown in italic, GO categories enriched among eQTL targets in regular font. See Tables S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, S11, S12 for a list of the top significant GO terms of each mapping.