First release of epigraHMM on Bioconductor.
It is now possible to add normalizing offsets via addOffsets
.
epigraHMM now uses hdf5 files to store all intermediate data during computation of the EM algorithm. Intermediate data include window-based HMM and mixture model posterior probabilities, and forward-backward probabilities. This change leads to a better memory utilization of the package.
segmentGenome
that segments a given
genome (e.g. 'mm10') into non-overlapping genomic windows while considering
gap tracks and blacklisted regions.Adding function callPatterns
to exp[ort] combinatorial patterns (or posterior
probabilities) associated with a given set of genomic regions.
Adding function info
to print summary statistics from epigraHMM output. This
function will print the model's BIC, log-likelihood, and combinatorial patterns
associated with mixture model components.
Adding new example dataset helas3
with ENCODE ChIP-seq data from broad
epigenomic marks H3K27me3, H3K36me3, and EZH2.
Adding option to prune combinatorial patterns associated with rare states. See vignette for details.
In differential peak calling, epigraHMM now exports combinatorial pattern table. See vignette for details.
Improvement of the vignette to clarify epigraHMM's use of blacklisted regions and gap tracks.
Minor bug fix in callPatterns and info function (explict import of S4Vectors::mcols and utils::tail).
Exporting expStep function, which implements the E-step of EM algorithm (forward-backward & Viterbi algorithm) for a K-state HMM.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.