Abstract:
|
The analysis of global gene expression and alternative splicing via high throughput mRNA sequencing (RNA-seq) has become a powerful tool in the study of cancer. Motivated by a large RNA-seq study from a recent clinical trial, we develop a scalable, efficient, and flexible approach for the detection of differential alternative splicing from RNA-seq data. Our regression-based approach is compatible with output from commonly used isoform-level expression quantitation programs such as RSEM and Sailfish, and is capable of detecting differential splicing events associated with arbitrary sets of covariates (continuous or discrete). We demonstrate our method's computational efficiency and performance against alternative methods using simulated data. We apply our approach to detect differential alternative splicing events from 400 multiple myeloma RNA-seq samples with respect to various clinical and genomic factors. Lastly, we discuss extensions to incorporate technical uncertainly in isoform-level expression estimates to improve detection of differential splicing events.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.