Efficient and accurate causal inference with hidden confounders from genome-transcriptome variation data.

Area of Science:

• Systems genetics
• Genomics
• Bioinformatics

Background:

• Gene expression quantitative trait mapping reveals genetic variation's functional impact.
• Current causal inference methods struggle with hidden confounders and weak regulatory effects.

Purpose of the Study:

• To develop a novel, highly accurate, and ultra-fast method for causal inference between gene expression traits.
• To improve upon existing methods by incorporating hidden confounders and weak regulations.

Main Methods:

• Developed Findr, a novel software utilizing cis-regulatory DNA variations as causal anchors for inference.
• Applied whole genome sequencing and transcriptome analysis for quantitative trait mapping.

Main Results:

• Findr demonstrated superior accuracy in causal inference compared to existing methods.
• Achieved significant speed improvements, being nearly a million times faster than previous approaches.
• Successfully predicted microRNA and transcription factor targets in human lymphoblastoid cells.

Conclusions:

• Findr offers a powerful and efficient tool for understanding the functional consequences of genetic variation.
• The method advances systems genetics by accurately inferring causal relationships in gene expression data.