Genome-Wide Causation Studies of Complex Diseases

Area of Science:

• Genetics
• Complex Disease Research
• Causal Inference

Background:

• Genome-Wide Association Studies (GWAS) have advanced complex disease genetics but often identify variants lacking direct pathological relevance.
• A significant portion of disease-causing genetic variants remains undiscovered due to the limitations of association-based analyses.
• Association analysis, while useful, provides limited insight into the causal mechanisms underlying diseases.

Purpose of the Study:

• To propose Genome-Wide Causation Studies (GWCS) as a novel approach to identify causal genetic structures.
• To introduce and evaluate Additive Noise Models (ANMs) for genetic causation analysis.
• To investigate the overlap between association and causation signals in complex diseases, using schizophrenia as a case study.

Main Methods:

• Development and application of Additive Noise Models (ANMs) for testing genetic causation.
• Conducting Genome-Wide Causation Studies (GWCS) as an alternative to traditional GWAS.
• Simulation studies and real-data analysis (schizophrenia) to assess ANM performance and signal overlap.

Main Results:

• The study presents the Type 1 error rates and statistical power of ANMs for causation testing.
• Analysis of schizophrenia data revealed a small proportion of overlap between association and causation signals.
• Simulation results corroborate the findings from real-data analysis regarding signal overlap.

Conclusions:

• GWCS, utilizing ANMs, offers a promising avenue for discovering underlying causal genetic structures beyond GWAS capabilities.
• The limited overlap between association and causation signals highlights the need for advanced methods like GWCS.
• This research calls for a re-evaluation of the utility of GWAS and encourages the adoption of causation-based approaches in complex disease genetics.