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

• Published in: Nature communications 2021-09, Vol.12 (1), p.5647-9, Article 5647
• Main Authors: Porcu, Eleonora, Sadler, Marie C., Lepik, Kaido, Auwerx, Chiara, Wood, Andrew R., Weihs, Antoine, Sleiman, Maroun S. Bou, Ribeiro, Diogo M., Bandinelli, Stefania, Tanaka, Toshiko, Nauck, Matthias, Völker, Uwe, Delaneau, Olivier, Metspalu, Andres, Teumer, Alexander, Frayling, Timothy, Santoni, Federico A., Reymond, Alexandre, Kutalik, Zoltán
• Format: Article
• Language: English
• Subjects: 38; 38/43; 45; 631/114/2415; 631/208/199; 631/208/205; Algorithms; Biomarkers; Causality; Correlation coefficient; Correlation coefficients; Gene expression; Gene Expression Profiling - methods; Genes; Genetic Association Studies - methods; Genetic Predisposition to Disease - genetics; Genome-Wide Association Study - methods; Humanities and Social Sciences; Humans; Mendelian Randomization Analysis - methods; multidisciplinary; Phenotype; Phenotypes; Polymorphism, Single Nucleotide; Quantitative Trait Loci - genetics; Randomization; Science; Science (multidisciplinary); Transcription; Transcriptome - genetics; Transcriptomes; Triglycerides
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-021-25805-y

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. Identification of gene expression changes between healthy and diseased individuals can reveal mechanistic insights and biomarkers. Here, the authors propose a bi-directional transcriptome-wide Mendelian Randomization approach to assess causal effects between gene expression and complex traits.