Differentially expressed genes reflect disease-induced rather than disease-causing changes in the transcriptome

Abstract

Comparing transcript levels between healthy and diseased individuals allows the identification of differentially expressed genes, which may be causes, consequences or mere correlates of the disease under scrutiny. We propose a method to decompose the observational correlation between gene expression and phenotypes driven by confounders, forward- and reverse causal effects. The bi-directional causal effects between gene expression and complex traits are obtained by Mendelian Randomization integrating summary-level data from GWAS and whole-blood eQTLs. Applying this approach to complex traits reveals that forward effects have negligible contribution. For example, BMI- and triglycerides-gene expression correlation coefficients robustly correlate with trait-to-expression causal effects (r_BMI = 0.11, P_BMI = 2.0 × 10^-51 and r_TG = 0.13, P_TG = 1.1 × 10^-68), but not detectably with expression-to-trait effects. Our results demonstrate that studies comparing the transcriptome of diseased and healthy subjects are more prone to reveal disease-induced gene expression changes rather than disease causing ones.

Fig. 1. TWMR and revTWMR.

Schematic representation of how TWMR and revTWMR dissect bidirectional causal and confounder contributions to the observed correlation between gene expression and phenotype.

Fig. 2. Partitioning the gene expression-trait observational correlation for BMI, HDL, and triglycerides.

Using all the genes tested by TWMR and revTWMR (N = 10,395 for BMI, N = 10,391 for HDL, and N = 10,390 for TG), for each bin of correlation (absolute value) we plotted the combined contributions of the forward (TWMR, blue dots) and reverse (revTWMR, red dots) effect of the gene expression on the trait, the contribution of confounders (black dots). Data were presented as estimated contributions and 95% confidence intervals.