This genetic dataset is one of my favorites mainly for it’s size.  One of my colleagues let me use this dataset which he obtained by using radiation hybrids.  To make radiation hybrids, cells from a donor species are irradiated which causes their chromosomes to fragment killing the vast majority of cells.  Some cells are rescued by fusing with a recipient species.  In the fusing process, some of the fragments from the donor cells get inserted into the chromosomes of the recipient species.  These fused cells are then selectively bred and assayed.  Using this process, they created nearly 100 of these hybrids and measured over 200,000 markers along with over 20,000 gene expression levels.  Using this data, we regressed the gene expression levels on the markers.  We then placed an edge between two genes for each marker they showed statistically significant dependence on.  This means that this network has more than 20,000 nodes with each pair having at most 200,000 edges between them.  In the view below, all edges are significant, but red edges are the most significant.  For a full pdf of the network click here.

Comments closed.