PROMISE-ME: A Robust Method for Integrated Analysis of DNA Methylation, Gene Expression, and Multiple Biologically Related Clinical and Pharmacological Outcomes

• Activity Number: 428
• Type: Contributed
• Date/Time: Tuesday, August 2, 2016 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Stanley Pounds* and Xueyuan Cao and Tong Lin
• Companies: St. Jude Children's Research Hospital and St. Jude Children's Research Hospital and St. Jude Children's Research Hospital
• Keywords: omic data ; integrated data analysis ; molecular oncology

Abstract

Projection onto the Most Interesting Statistical Evidence (PROMISE) is a robust method to perform integrated analysis of one form of genomic data with multiple biologically related pharmacologic and clinical outcome variables. Here, we extend that framework to develop PROMISE for methylation and expression (PROMISE-ME) as a robust method to perform integrated analysis of DNA methylation, gene expression, and multiple pharmacologic and clinical outcome variables. For each gene, PROMISE-ME evaluates the association of methylation with expression, the association of methylation with each outcome variable, and the association of expression with each outcome variable. Previous knowledge of the biological relationships among outcome variables is used to define a test statistic that is a linear combination of each of the pairwise association statistics described above. Permutation is used to determine the significance of this test statistic. The advantages of PROMISE-ME in terms of enhancing statistical power for meaningful biological discoveries are seen in simulation studies and examples from pediatric oncology research.