|
Activity Number:
|
126
- SPEED: New Methods in Statistical Genomics and Genetics Part 1
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 29, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Genomics and Genetics
|
|
Abstract #304359
|
Presentation
|
|
Title:
|
A Generalized Multi-Response Permutation Procedure to Evaluate Associations of Multivariate Data with Quantitative and Censored-Event Time Variables
|
|
Author(s):
|
Stanley Pounds* and Natasha Sahr and Xueyuan Cao
|
|
Companies:
|
St. Jude Children's Research Hospital and St. Jude's Children's Hospital and University of Tennessee Health Science Center
|
|
Keywords:
|
genomic;
gene-set analysis;
multivariate associations;
permutation
|
|
Abstract:
|
The multi-response permutation procedure (MRPP) is a robust tool to identify both simple and complex patterns of multivariate differentiation across two or more groups. We propose a generalized MRPP (GMRPP) to evaluate associations of multivariate data with quantitative or censored event-time variables. The GMRPP repeatedly applies the traditional MRPP to evaluate the association of multivariate data with groups defined by all possible dichotomizations of the quantitative or event-time endpoint and combines the results to obtain a final assessment. An adaptive permutation procedure greatly improves computational efficiency while maintaining statistical rigor. In a simulation study, the GMRPP shows much better statistical power than other methods to identify complex associations of multivariate data with a categorical, quantitative, or censored event-time variable. Additionally, the GMRPP provides biologically meaningful results in a pediatric leukemia genomics example. GMRPP is available as an R package from GitHub.
|
Authors who are presenting talks have a * after their name.
