Data-driven RNA phenotyping captures genetically regulated dimensions of the transcriptome.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Transcriptomic diversity is shaped by RNA regulation, crucial for mapping genetic associations (xQTLs) and interpreting genome-wide association studies (GWAS).
• Existing multimodal frameworks like Pantry integrate various RNA data but require complete gene annotations and face statistical complexity.
• Limitations in current tools hinder comprehensive discovery of regulatory variations influencing complex traits.

Purpose of the Study:

• To introduce LaDDR (Latent Data-Driven RNA phenotyping), a novel mechanism-agnostic framework for enhanced xQTL discovery and GWAS integration.
• To enable xQTL discovery and GWAS interpretation without reliance on complete gene annotations.
• To broaden the landscape of detectable trait-relevant transcriptomic regulation.

Main Methods:

• LaDDR generates orthogonal, latent coverage features per gene, independent of specific RNA regulation mechanisms.
• The framework was applied to GTEx data for xQTL discovery and integrated with transcriptome-wide association studies (TWAS).
• LaDDR was evaluated against knowledge-driven phenotypes from the Pantry framework, with and without residualizing known modalities.

Main Results:

• LaDDR identified an average of 95% more independent xQTLs per tissue compared to Pantry's six knowledge-driven modes.
• Combining LaDDR with knowledge-driven phenotypes increased discovery by an additional 41% per tissue on average.
• In TWAS of 114 complex traits, LaDDR uncovered an average of 11,790 unique gene-trait pairs per tissue, surpassing knowledge-driven phenotypes (8,579).

Conclusions:

• LaDDR significantly expands the discovery of genetic associations by efficiently capturing regulatory variation missed by current pipelines.
• The framework provides a powerful, annotation-agnostic approach for xQTL discovery and GWAS interpretation.
• LaDDR broadens the detectable landscape of trait-relevant transcriptomic regulation, offering new insights into complex trait genetics.