- Location
- --
- Video
-
--
- Abstract
- A typical microarray experiment compares two or several conditions
(for example, cancer vs. normal cells) and produces a list of genes
that are differentially expressed. The standard approach is to score
each gene individually. However, real genes do not act alone: they
interact in various ways, and cellular processes involve pathways that
contain multiple genes. Many of those pathways and interactions are
known.
Hence an analysis that combines the microarray data with prior
biological knowledge has the potential to be both more biologically
relevant and statistically powerful. We survey some of the methods
that have been proposed to address this problem, with a focus on
algorithms based on interaction graphs.
- Supplements
-
--
|
