Stochastic imputation for integrated transcriptome association analysis of a longitudinally measured trait

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Mechanistic links between genetic variations and complex diseases are poorly understood.
• Transcriptome data offers a powerful resource for investigating disease mechanisms.

Purpose of the Study:

• To propose a novel two-stage imputation method for transcriptome-wide association studies (TWAS).
• To extend the method for analyzing longitudinal traits and gene expression.
• To evaluate the performance of the proposed method through simulations and a case study.

Main Methods:

• A two-stage strategy using stochastic regression imputation with error incorporation.
• Bootstrap procedure for flexibility in imputation.
• Linear mixed-effects modeling and composite test statistics for longitudinal data.
• Simulation studies and a case study involving inflammasome genes and inflammatory response.

Main Results:

• The proposed method provides a flexible framework for TWAS.
• The generalization to longitudinal data allows for dynamic analysis of gene expression's role in disease trajectories.
• Simulation studies assess the method's accuracy, power, and error rates.

Conclusions:

• The novel imputation strategy enhances the analysis of genetic associations with complex traits using transcriptome data.
• The method is adaptable for both cross-sectional and longitudinal studies, offering new insights into disease mechanisms.