Quantification method affects replicability of eQTL analysis, colocalization, and TWAS

Area of Science:

• Genomics
• Bioinformatics
• Statistical Genetics

Background:

• Genome-wide association studies (GWAS) are powerful for identifying genetic variants associated with traits.
• eQTL mapping and transcriptome-wide association studies (TWAS) are commonly used to interpret GWAS findings by linking genetic variants to gene expression.
• The influence of RNA-sequencing (RNA-seq) data processing choices on these downstream analyses is not well understood.

Purpose of the Study:

• To investigate how different RNA-seq quantification methods and transcriptomic references impact eQTL detection and gene expression prediction.
• To assess the downstream consequences of these RNA-seq processing choices on genetic colocalization and TWAS results.
• To highlight the need for standardized RNA-seq processing protocols in genetic association studies.

Main Methods:

• Systematic evaluation of various RNA-seq quantification tools.
• Comparison of results using different transcriptomic references.
• Analysis of eQTL detection rates and gene expression prediction accuracy.
• Assessment of the impact on gene-trait colocalization and TWAS summary statistics.

Main Results:

• The choice of RNA-seq quantification method significantly alters eQTL detection and gene expression prediction.
• The selection of a transcriptomic reference also substantially impacts eQTL mapping and TWAS outcomes.
• These methodological variations lead to significant downstream effects on the interpretation of genetic association studies.

Conclusions:

• Seemingly minor decisions in RNA-seq data processing have a substantial impact on eQTL and TWAS results.
• Current practices in RNA-seq processing introduce variability that can affect the reproducibility of genetic association studies.
• Standardization of RNA-seq quantification and reference selection is crucial for robust and reproducible genetic research.