Efficient and effective control of confounding in eQTL mapping studies through joint differential expression and...

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Expression quantitative trait loci (eQTLs) mapping is crucial for understanding gene expression variation.
• Controlling confounding factors in eQTL studies is essential for accurate results.
• Probabilistic estimation of expression residual (PEER) analysis is a common method, but determining the optimal number of PEER factors is computationally intensive.

Purpose of the Study:

• To introduce ECCO (Effect size Correlation for COnfounding determination), a computationally scalable alternative for determining the optimal number of PEER factors.
• To enable efficient and optimized eQTL discovery in large-scale genetic studies.

Main Methods:

• ECCO combines differential expression analysis and Mendelian randomization analysis to determine the optimal number of PEER factors.
• This approach avoids repetitive eQTL mapping procedures, significantly reducing computational cost.

Main Results:

• ECCO is two orders of magnitude faster than the standard approach for determining PEER factors.
• ECCO identifies a similar number of optimal PEER factors as the standard method.
• ECCO facilitated optimized eQTL discovery across 48 GTEx tissues, achieving a 5.89% power gain in discovering eQTL harboring genes (eGenes).

Conclusions:

• ECCO provides a computationally efficient and scalable solution for optimizing PEER factor selection in eQTL mapping.
• The ECCO software and associated results enable enhanced eQTL discovery, particularly in large-scale genomic datasets.
• This advancement aids in a deeper understanding of the genetic determinants of gene expression.