Sjaracne

Highlights of this release

Fixed two bugs:
1. TF_run_000 was the same as TF_run_001;
2. The total number of edges calculation for a bootstrap run was incorrect, this causes the wrong calculation of the probability of the Bernoulli distribution.
Updated corresponding tests.
Improved README documentation.

SJARACNe 0.2.0 includes some major improvements and bug fixes compared to 0.1.0.

Highlights of this release
1. Implement the workflow using CWL that supports running on multiple platforms.
2. Allow users to control resource usage (runner depended on different platforms).
3. Merging consensus network creation and enhanced consensus network creation into one step. Refactoring to increase its speed/memory and reliability.
4. Acceptance tests and unit tests (python) using Travis CI.
5. Entry wrapper script that supports local and IBM LSF.
6. Improved documentation.

SJARACNe is a scalable solution of ARACNe that dramatically improves the computational performance, especially on the memory usage to allow even researchers with modest computational power to generate networks from thousands of samples. The algorithm uses adaptive partitioning mutual information to calculate the correlation between all pairs of genes and reconstruct the regulatory network.