|
Recent advances in large-scale mRNA expression
measurements and proteomic technologies have opened the opportunity for
massively parallel biological data acquisition and thus have shifted our
attention towards an integrated understanding of the genetic networks
underlying complex biological phenotypes. Despite many statistical
developments for genomics research over the past several years, some basic
statistical questions have not been very well addressed primarily because a
very large number of genes are studied with a limited number of replicates.
Therefore, it is not a trivial task to identify genes with different
expression levels across a set of samples and to identify genes with
correlated expression patterns. It is even more challenging to understand
biological pathways from microarray data and to integrate data of various
types.
The Biostatistics Resources provides state-of-the-art statistical
analyses to identify genes with different expression
levels among a set of conditions, to cluster genes with co-regulated
expression patterns, and to identify biomarkers from microarray data for
disease diagnosis and prognosis. In addition to providing basic biostatistics
services, we are collaborating with the
Microarray, and
Affymetrix Resources and with many research labs which are conducting
large scale gene expression experiments. We are also part of the
Yale Microarray Database.
For statistical and computational methodology developments, we are (1)
developing statistical methods to estimate gene expression level changes
among a set of samples and to identify genes with different expression
levels for both cDNA microarrays and
Affymetrix microarrays;
(2) developing novel approaches for the analysis of clustering; and (3)
developing computational and visualization tools to dissect biological
pathways. Our active methodology and collaborative research programs will ensure that
we will be able to provide the most powerful statistical methods to extract
as much information as possible from microarray experiments.
|
