Efficient inference of parent-of-origin effect using case-control mother-child genotype data

Area of Science:

• Genetics
• Developmental Biology
• Biostatistics

Background:

• Parent-of-origin effects significantly influence mammalian development and disease.
• Existing methods for detecting these effects often lack covariate incorporation, limiting their ability to control for confounding factors.
• Case-control mother-child pair genotype data offer a practical approach for studying parent-of-origin effects.

Purpose of the Study:

• To develop a robust statistical model for assessing parent-of-origin effects using logistic regression.
• To incorporate covariates into the model for controlling confounding factors.
• To enhance the inference efficiency of parental origins using linked genetic markers.

Main Methods:

• A logistic regression model was proposed, including maternal and child genotypes, parental origins, and covariates.
• Genotypes of markers tightly linked to the target marker were used to improve the inference of parental origins.
• A modified profile log-likelihood and a computationally feasible expectation-maximization (EM) algorithm were developed for robust statistical inference.
• The EM algorithm was adapted to handle missing child genotypes and its convergence was established.

Main Results:

• The proposed method effectively models parent-of-origin effects while accounting for covariates.
• The use of linked markers improved the efficiency of inferring parental origins.
• The developed EM algorithm demonstrated robustness and handled missing data effectively.
• Large sample properties, including consistency, asymptotic normality, and efficiency, were established for the proposed estimator.

Conclusions:

• The novel logistic regression model provides a powerful and flexible tool for detecting parent-of-origin effects in genetic studies.
• The integrated approach enhances the accuracy of parent-of-origin effect detection by incorporating covariates and utilizing linked marker information.
• The robust statistical inference procedure and efficient EM algorithm offer practical advantages for analyzing genetic data, including handling missing genotypes.