Cancer cell lines

The Broad Institute has been working with the NCI (National Cancer Institute) in an attempt to characterize a large number of the 949 human cancer cell lines currently available for study. Each of these cancer cell lines has been explored in terms of gene expression, gene copy number and somatic mutations. For more information on this project visit the Cancer Cell Line Encyclopedia.

Some of the computational work has been undertaken by Brinton Seashore-Ludlow and Matthew Rees. Brinton worked out a nice set of hierarchical relationships between the cell lines which I put into the form of a sunburst visualization. A static presentation of this hierarchy may be seen below.

The main motivation for reworking the interactive image into a static form comes from the labels, which are tough to read when they are all overlapping one another. To fix this problem Brinton took the raw SVG my code generated and used Adobe Illustrator to rework each of those labels, giving them all a nice, roomy place on the screen and the line linking the label to the appropriate arc. This is the process people commonly undertake to take svg image and make it suitable for publication.