Related Experiment Videos
Quantifying the relationship between gene expressions and trait values in general pedigrees.
Yan Lu1, Peng-Yuan Liu, Yong-Jun Liu
1Osteoporosis Research Center, Creighton University, Omaha, Nebraska 68131, USA.
Genetics
|September 17, 2004
Summary
We developed new statistical methods to analyze gene expression and clinical traits in families. Our variance-components approach improves accuracy and power in genetic association studies, especially with complex family structures.
Area of Science:
- Genetics
- Bioinformatics
- Statistical Genomics
Background:
- Quantitative genetics approaches analyze mRNA transcript abundances as quantitative traits in relation to clinical traits.
- The family expression association test (FEXAT) uses a family-based design but may not fully account for biological relationships, potentially affecting statistical accuracy.
- Existing methods may not optimally handle complex family structures and genetic covariances in expression-trait association studies.
Purpose of the Study:
- To propose novel statistical test methods for analyzing microarray data from general pedigrees.
- To develop methods that account for biological relationships and genetic effects among relatives.
- To improve the accuracy and power of expression-trait association studies.
Main Methods:
- Developed two new test statistics using a variance-components approach for general pedigrees.
- Incorporated covariance for unmeasured genetic effects among relatives.
- Modeled covariates of clinical importance directly within the statistical framework.
- Investigated method efficacy and validity using simulated data with varying sample and family structures.
Main Results:
- The proposed Likelihood Ratio (LR) method demonstrates correct Type I error rates for moderate to large sample sizes, irrespective of family structure.
- The LR method exhibits higher statistical power in complex pedigrees compared to existing methods.
- The FEXAT(R) method shows advantages with large family sizes, maintaining accuracy across different sample and family structures.
- Both proposed methods are robust to population stratification.
Conclusions:
- The new variance-components methods offer robust and powerful tools for expression-trait association studies in diverse family structures.
- These methods complement existing approaches like FEXAT, particularly in complex family designs.
- The developed statistics enhance the analysis of genetic influences on clinical traits through gene expression data.