Identifying causal genotype-phenotype relationships for population-sampled parent-child trios

Area of Science:

• Genetics
• Biostatistics
• Causal Inference

Background:

• Genetic transmission introduces randomization preceding phenotype determination.
• Parent-child trio data can model this randomization for causal discovery.
• Existing genotype-phenotype association methods may not fully capture causal genetic effects.

Purpose of the Study:

• To develop and validate novel statistical methods for identifying causal genetic loci using parent-child trio data.
• To model genetic transmission as a source of experimental randomization for causal inference.
• To compare newly developed methods with established genotype-phenotype analysis techniques.

Main Methods:

• Development of the transmission mean test (TMT) for causal effect estimation.
• Derivation of causal properties of the TMT within the potential outcomes framework.
• Proof that the transmission disequilibrium test (TDT) is a specific case of the TMT for affected-only designs.

Main Results:

• The TMT provides an unbiased estimator for average causal effects.
• The TMT is applicable to arbitrarily distributed traits and randomly sampled trios.
• The TMT and TDT offer complementary approaches for different trait types and study designs.

Conclusions:

• Parent-child trio data can be effectively utilized for causal genetic locus identification through transmission modeling.
• The TMT and TDT are valuable causal inference tools, each suited to specific study conditions.
• Integrating transmission-based methods with other genotype-phenotype analyses enhances causal discovery in genetics.