Genome-wide association analyses of expression phenotypes

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genomics

Background:

• High-throughput microarray experiments generate vast datasets, posing challenges for genotype-phenotype association studies.
• Interpreting genome-wide association studies (GWAS) for expression phenotypes requires robust analytical strategies.
• The Genetic Analysis Workshop 15 (GAW15) provided a platform to address these data analysis challenges.

Purpose of the Study:

• To explore issues in analyzing large-scale genotype and expression data from microarray experiments.
• To evaluate diverse hypotheses related to quantitative trait association, including cancer and obesity.
• To assess various analytical techniques, including information theory-based methods and machine learning.

Main Methods:

• Contributions evaluated hypotheses using diverse analytical techniques, including information theory.
• Focus on association of quantitative expression data.
• Exploration of single nucleotide polymorphism (SNP) selection and individual inclusion criteria in study design.

Main Results:

• Careful consideration of the genetic model and early selection of SNPs and individuals are crucial for study design.
• Pathway-specific analyses yield more interpretable results than agnostic genome-wide approaches.
• Machine learning methods are practical for datasets with small sample sizes and numerous features.

Conclusions:

• Effective analysis of high-throughput genomic data necessitates strategic study design, including genetic model consideration and SNP/individual selection.
• Pathway-targeted analysis enhances the interpretability of GWAS findings.
• Machine learning offers a viable alternative to traditional methods for analyzing complex genomic datasets with limited sample sizes.