Statistical Learning of Sparse and Structured Biological Networks
We know that genes, proteins, and other molecules operate not on their own, but as part of complex pathways or regulatory networks. Here we consider the task of uncovering such regulatory networks on the basis of gene expression data. In this context, we draw an edge between a pair of genes that are partially correlated — that is, correlated conditional on all of the other genes. We present some techniques for estimating such networks on the basis of high-dimensional gene expression data sets.
Daniella Witten,PhD , Assistant Professor, Biostatistics