Abstract
DATE: |
April 30, 2009 |
TIME: |
1:15 - 3:00 pm |
LOCATION: |
Center for Clinical Sciences Research (CCSR), Rm 4205 |
TITLE: |
Computational Deconvolution of Gene Expression from Mixed Tissue: The Sequel |
SPEAKERS: |
Rob Tibshirani |
Accurate analysis of gene expression patterns from many tissues
is hampered by the variation in relative cell subset frequency from
one sample to another and the different expression patterns of each
cell type. Blood is a mixed tissue containing many different cell
subsets, which vary in relative frequency between individuals, both
in humans and in animal models. Despite this, and at a loss of
sensitivity for differential expression, analysis of gene expression
from peripheral blood is frequently used in both basic and clinical
research as it is easily accessible and is thought to be reflective
of immune state.
Here, we will describe a statistical methodology aimed at harnessing
quantitative information on mixed tissue composition to control for
tissue heterogeneity and yield increased sensitivity and specificity
from gene expression studies. We apply the method to a study of rejection
in kidney transplants.
This work is in collaboration with Dale Bodian, Atul Butte,
Mark Davis, Trevor Hastie, Purvesh Khatri, Balasubramanian
Narasimhan, Nicholas Perry, Minnie Sarwal, Lihua Ying.
Internal Staff Login