minepy - Maximal Information-based Nonparametric Exploration

https://travis-ci.org/minepy/minepy.png?branch=master Documentation Status https://anaconda.org/bioconda/minepy/badges/version.svg https://badge.fury.io/py/minepy.svg

minepy provides a library for the Maximal Information-based Nonparametric Exploration (MIC and MINE family). Key features:

  • APPROX-MIC (the original algorithm, DOI: 10.1126/science.1205438) and MIC_e (DOI: arXiv:1505.02213 and DOI: arXiv:1505.02214) estimators;
  • Total Information Coefficient (TIC, DOI: arXiv:1505.02213) and the Generalized Mean Information Coefficient (GMIC, DOI: arXiv:1308.5712);
  • an ANSI C library
  • a C++ interface;
  • an efficient Python API (Python 2 and 3 compatibility);
  • an efficient MATLAB/OCTAVE API;

minepy is an open-source, GPLv3-licensed software.

The minerva R interface is available at CRAN.

MICtools

The `mine` command-line application is deprecated since version 1.2.2. We suggest to use MICtools, a comprehensive and effective pipeline for TICe and MICe analysis. TICe is used to perform efficiently a high throughput screening of all the possible pairwise relationships assessing their significance, while MICe is used to rank the subset of significant associations on the bases of their strength. Paper, code and documentation.

Docker image

The minepy library is preinstalled in the MICtools Docker image.

Indices and tables