What is slcview?

slcview is a program I wrote so that I could create image files from .cdt and .gtr files en masse from a Unix command prompt.

Why would I need slcview?

Cluster and TreeView by Mike Eisen are great programs, giving you a lot of power to manipulate your data interactively. The tradeoff for the interactivity is difficulty in automation, though. If you don't need to dig down into the details of individual clustergrams but instead need to look overall at a lot of clustering results, you probably want to automate the process. XCluster by Gavin Sherlock provides a fast, scriptable program for doing clustering of many files from the command line, which helps half the problem. I couldn't find a program that satisfied these three criteria:

  1. Runs under Linux/Unix
  2. Lets you script the creation of clustergram image files, while allowing good control of the output.
  3. Produces good looking output, like TreeView.

(I found out later that Gavin Sherlock also wrote a program to generate clustergram and tree images, but I'd written slcview already, and I think it gives the user more options. If you want to make lots of image files, though, either one will probably work ok for you, and you might want to check out Gavin's program. It should be distributed with his SOM-viewing software. See the bottom of this page for a link to Gavin's site.)

What does slcview do?

slcview is a Perl program that will take in the .cdt, .gtr, and .atr files output by Cluster or XCluster and create the clustergram, gene tree, and array tree images. There are many options you can specify to control how the output looks. To my eye, the images are indistinguishable from those output by TreeView.

Since I wrote this myself, I wanted to be able to control the output in certain respects. So you can specify the resolution of the clustergram and tree diagrams separately. You can specify the colors you want for the positive, negative, and missing values. You can specify the background color and the color of the lines used for drawing the tree diagrams.

Additional bonuses include the ability to draw a legend which uses the same color scaling as your clustering data files, so you will know how to translate back from colors to numbers on your resulting clustergram. Also, one very big benefit of using ImageMagick is that you can output to any of a variety of graphics formats, just by specifying the filename extension. GIF, JPG, TIFF, PDF, PS, etc. etc. etc. are all supported just by changing a few letters. A full list of compatible output file formats is available at the ImageMagick web site.