HaMMLET

HaMMLET – Fast Bayesian HMM segmentation for big data

This software implements Forward-Backward Gibbs sampling for Bayesian segmentation in Hidden Markov Models (HMM). It uses dynamic wavelet compression to drastically improve convergence and memory consumption, making inference possible on large-scale data.

For instance, HaMMLET can be used on a regular laptop for segmentation of genomic data, such as array-CGH or depth-of coverage from whole-genome sequencing (WGS), to find candidates for copy-number variants (CNV). For details, please refer to the doc/ directory.

For implementation details and the theory behind this approach, please refer to my thesis (DOI: 10.7282/t3-4e1k-ph18).

Hammlet

Fast Bayesian Hidden Markov Model with Wavelet Compression

Hammlet Info

⭐ Stars 10
🔗 Homepage wiedenhoeft.github.io
🔗 Source Code github.com
🕒 Last Update a year ago
🕒 Created 7 years ago
🐞 Open Issues 3
➗ Star-Issue Ratio 3
😎 Author wiedenhoeft