https://github.com/nvictus/lavaburst

Chromatin domains bursting with flavor
https://github.com/nvictus/lavaburst

Last synced: 15 days ago
JSON representation

Chromatin domains bursting with flavor

• Host: GitHub
• URL: https://github.com/nvictus/lavaburst
• Owner: nvictus
• License: mit
• Created: 2015-05-04T04:17:09.000Z (over 9 years ago)
• Default Branch: master
• Last Pushed: 2019-12-29T06:01:34.000Z (almost 5 years ago)
• Last Synced: 2024-10-03T12:17:49.504Z (about 1 month ago)
• Language: Jupyter Notebook
• Homepage: http://nezar-compbio.github.io/lavaburst
• Size: 13.4 MB
• Stars: 12
• Watchers: 2
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.rst
  • Changelog: CHANGES.md
  • License: LICENSE

Awesome Lists containing this project

README

        
Lavaburst

=========

Chromatin domains bursting with flavor!

Let's get started! See `IPython

Notebook `__.

Optimal domain segmentation

~~~~~~~~~~~~~~~~~~~~~~~~~~~

Produce the highest scoring domain segmentation according to the model.

Marginal domain boundary probabilities

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Produce the marginal probabilities for each bin edge being a domain boundary. The

marginal boundary probability of a single bin edge is the total

probability over all possible domain segmentations for which that bin edge serves as a

border between domains. The output is a 1D array.

Marginal domain probabilities

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Produce the marginal probabilities of every unique segment. The output

is a 2D array where ``prob[a,b]`` describes the overall frequency of the

domain spanning bin edges ``a`` and ``b`` to occur in the ensemble of all

possible domain segmentations.

Domain boundary co-occurence probabilities

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Produce the marginal co-occurrence probabilities of every pair of bin

edges as domain boundaries. The output is a 2D array where ``prob[a,b]``

describes the overall frequency all segmentations in which both bin

edges ``a`` and ``b`` occur as domain borders.

Locus co-segmentation probabilities

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Produce the marginal probability of every pair of bins to co-occur

within the same domain. The output is a 2D array where ``prob[i,j]``

describes the overall probability of genomic bins ``i`` and ``j`` being

part of the same domain, over all possible segmentations.

Exact statistical sampling of domain segmentations

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Sample individual segmentations from the probabilistic model.