Clustered Heat Maps (CHMs) were first used in the biological sciences in the early 1990’s for visualizing pharmacogenomic data (Weinstein, Stem Cells, 1994) and later for integrated visualization of genomic, transcriptomic, proteomic, and functional data (Weinstein, Science, 1997). As the ubiquitous first-order way of visualizing omic data, CHMs have appeared in many thousands of publications and have been used extensively to represent TCGA data. Their limitations have been summarized, along with cautions about improper use (Weinstein, Science, 2008).
One such limitation is that currently popular approaches to the generation of CHMs don’t scale well. They produce fundamentally static images. It’s difficult to use a single representation for both the entirety of a large data set and details of individual data points. What we wanted was a visualization tool suited to dynamic exploration of large omic data sets as a whole and also in detail. We therefore initiated the Next-Generation CHM project, focusing it on TCGA data. The new web-based visualization tool brings new levels of scalability and interactivity to this popular visualization paradigm (Broom, Ca Res, 2018).
If you’re going to create a clustered heat map for large-scale data, why not make it a highly interactive, shareable NG-CHM? It’s almost as easy.
The NG-CHM system offers new levels of scalability and interactivity necessary for exploring large data sets. The JavaScript-based approach works well to balance the competing requirements of high-level and detailed views of very large heat maps, and users can fluidly navigate between the two. There are suitable interfaces to incorporate a wide range of interactions and annotations. NG-CHMs are customizable to support many different data types.
To explore further please visit the project home page at http://www.ngchm.net/.