HMMSeg is a program for the scale-specific segmention of continuous genomic data using hidden Markov models (HMMs). It can segment multiple datasets simultaneously. Scale-specificity is achieved via an optional smoothing step using wavelets. HMMSeg is written in Java and may be downloaded for unrestricted use below. It has been tested on Windows and several Unix-type operating systems.

HMMSeg was written by Nathan Day and Andrew Hemmaplardh at the University of Washington.

Version release information may be found here.

  • Installation instructions. Since HMMSeg is distributed as a JAR archive, no installation should be necessary. Run as java -jar <path>HMMSeg.jar [options] [filelists]. The program requires Java 1.5 or newer to run. See the help file for more information.
  • Manuscript. See the article, "Unsupervised segmentation of continuous genomic data," Bioinformatics 2007 23:1424-1426
  • Essentials of wavelets (pdf) A short description of the wavelet analysis implemented in HMMSeg.
  • Example An example HMMSeg analysis with step-by-step instructions.
  • Validation and testing analysis