Activity Number:
|
513
- Gene Expression Analysis
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Section on Statistics in Genomics and Genetics
|
Abstract #322534
|
View Presentation
|
Title:
|
dtangle: a fast and accurate deconvolution estimator
|
Author(s):
|
Gregory Hunt* and Johann Gagnon-Bartsch and Saskia Freytag and Melanie Bahlo
|
Companies:
|
University of Michigan and University of Michigan and Walter and Eliza Hall Institute of Medical Research and Walter and Eliza Hall Institute of Medical Research
|
Keywords:
|
deconvolution ;
microarray ;
RNA-seq ;
gene expressions
|
Abstract:
|
The efficacy of gene profiling techniques is often jeopardized by heterogeneous samples. Detecting differences in expressions across samples is confounded by disparities in the constituent cell types of each sample. In order to separate the effects of sample heterogeneity from other biological factors methods of cell type deconvolution have been developed. Here we propose new methodology called dtangle. Our approach directly models the interaction between the quantity of oligonucleotides in a sample and expression measurements from the sample. Using real data we demonstrate that our algorithm has a competitive prediction error with existing methods while being much faster to compute.
|
Authors who are presenting talks have a * after their name.