Sufficient direction factor model and its application to gene expression quantitative trait loci discovery.

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Technological advancements have reduced the cost of genetic data collection.
• Secondary analysis of existing genetic data is a cost-effective alternative to new data collection.
• Understanding gene expression and single-nucleotide polymorphism (SNP) relationships is crucial in genomics.

Purpose of the Study:

• To develop a flexible and superior statistical method for analyzing the relationship between gene expression and SNPs.
• To enable the identification of statistically significant SNPs associated with gene expression.
• To advance the secondary analysis of genotype-tissue expression (GTEx) data.

Main Methods:

• Utilized a combination of factor analysis and dimension reduction estimation.
• Employed semiparametric sufficient dimension reduction methods for joint estimation.
• Developed a novel estimator that does not require normality assumptions for latent factors.

Main Results:

• The proposed semiparametric estimator demonstrated superior performance compared to existing methods.
• Quantified the asymptotic performance of the new parameter estimator.
• Successfully identified statistically significant SNPs associated with gene-SNP relations in lung tissue using GTEx data.

Conclusions:

• The developed method provides a flexible and powerful tool for secondary analysis of genetic data.
• This approach facilitates the discovery of novel gene-SNP associations.
• The findings have significant implications for understanding genetic regulation of gene expression in human tissues.