MICtools

MICtools is an open source pipeline which combines the TIC_e and MIC_e measures [Reshef2016] into a two-step procedure that allows to identify relationships of various degrees of complexity in large datasets. TIC_e is used to perform an efficient high throughput screening of all the possible pairwise relationships and a permutation based appraoch is used to assess their significance. MIC_e is then used to rank the subset of significant associations on the bases of their strength.

Please cite: Davide Albanese, Samantha Riccadonna, Claudio Donati, Pietro Franceschi; A practical tool for Maximal Information Coefficient analysis, GigaScience, giy032, https://doi.org/10.1093/gigascience/giy032

The MICtools pipeline can be broken into 4 steps (see the figure above):

1. given M variables pairs x_i and y_i measured in n samples, the empirical TIC_e null distribution is estimated by permutation;
2. TIC_e statistics and the associated empirical p-values are computed for all variable pairs;
3. p-values are corrected for multiplicity in order to control the family-wise error rate (FWER) or the false discovery rate (FDR);
4. finally, the strengths of the relationships called significant are estimated using the MIC_e estimator.

Table of contents

• Install
  • Using pip (Linux and macOS/OS X)
  • Docker (Linux, macOS/OS X and MS Windows)
  • From source
• Usage
• Tutorial
  • Preparation
  • Empirical TIC_e null distribution
  • TIC_e p-values
  • Multiple testing correction
  • Strength of significant associations

Install

Using pip (Linux and macOS/OS X)

We suggest to install Python 3 (3.5+) and the GCC compiler through the package manager (In Mac OS X, we recommend to install them using Homebrew) (e.g. on Ubuntu/Debian):

sudo apt-get update
sudo apt-get install build-essential python3-dev

Then, upgrade pip and install setuptools:

pip install --upgrade pip
pip install 'setuptools >=14.0'

Finally, install mictools:

pip install mictools

Docker (Linux, macOS/OS X and MS Windows)

1. Install Docker for Linux, Mac OS X or Windows.
2. Run the Docker Quickstart Terminal (Mac OS X, Windows) or the docker daemon (Linux, sudo service docker start).
3. Follow the instructions at https://hub.docker.com/r/minepy/mictools/.

From source

If you are installing from source, the following dependences must be installed: Python >= 3.5, Click >= 5.1, numpy >= 1.7.0, scipy >= 0.13, pandas >= 0.17.0, matplotlib >= 1.2.0,<2, statsmodels >= 0.6.1, minepy >= 1.2. We suggest to install these dependences using the OS package manager (Linux), Homebrew (macOS/OS X) or pip.

Download the latest stable version from https://github.com/minepy/mictools/releases and complete the installation:

tar -zxvf mictools-X.Y.Z.tar.gz
python3 setup.py install

Usage

MICtools can be used to investigate variable associations in different types of experimental scenarios:

• single dataset X, with M variables and N samples: to evaluate the M+(M-1)/2 possible associations;
• two datasets, X (MxN) and Y (KxN) (parameter -y/--yvars): to evaluate all the pairwise relationships between the variables of the two datasets (for a total of MxK associations). Note that the number samples (N) in the datasets X and Y must be the same.
• two datasets, X (MxN) and Y (KxN): to evaluate all the rowwise relationships (see -r/--rowwise), i.e. only the variables pairs X_i and Y_i (for each i in min(M, K)) will be tested;

In all the abovementioned cases the analysis will be performed within each class independently if the sample classes are provided (see -l/--labels and -t/--target).

MICtools is implemented as a single command (``mictools'') with the following subcommands:

null

Compute the TIC_e null distribution.

mergenull

Merge multiple TIC_e null distributions.

pval

Compute TIC_e p-values.

adjust

Multiple testing correction.

strength

Compute the strength (MIC_e).

Run mictools SUBCOMMAND --help for the documentation of each specific step.

Tutorial

We analyze the "Datasaurus" synthetic dataset generated following the approach discussed at https://www.autodeskresearch.com/publications/samestats ([Matejka2017]). The dataset contains 26 variables linked by 13 relationships which have the same summary statistics (e.g. the Pearson's correlation), but are very different in appearance. The dataset was modified in order to destroy secondary associations. In this example we test the entire set of possible associations (for a total of 26*(26-1)/2 = 325 relationships).

Preparation

Go to the examples folder:

cd examples

Select the Datasaurus dataset and the output folder:

X=datasaurus.txt
ODIR=datasaurus_results
mkdir $ODIR

Empirical TIC_e null distribution

Compute the empirical TIC_e null distribution (with 200,000 permutations, default value):

mictools null $X $ODIR/null_dist.txt

The output file null_dist.txt is a TAB-delimited file which contains the null distrubution:

Class	BinStart	BinEnd	NullCount	NullCountCum
None	0.000000	0.000100	0	200000
None	0.000100	0.000200	0	200000
None	0.000200	0.000300	0	200000
...	...	...	...	...

The first column (Class) contains the class membership (in this particular case no sample classes were provided), BinStart and BinEnd define the TIC_e range and NullCount and NullCountCum are distribution and the cumulative distribution, respectively.

TIC_e p-values

Compute the TIC_e statistics and the associated empirical p-values for all variable pairs:

mictools pval $X $ODIR/null_dist.txt $ODIR

The command will return in the output directory the following:

obs_dist.txt

the observed TICe distribution in the same format of null_dist.txt

obs.txt

TAB-delimited file containing the observed TICe values for each variable pair tested:

Var1	Var2	None
away_x	bullseye_x	0.029476
away_x	circle_x	0.018211
away_x	dino_x	0.050720
...	...	...

pval.txt

TAB-delimited file containing the empirical p-values for each variable pair

pval_None.png

the p values distribution plot:

Multiple testing correction

Correct the p-values for multiplicity in order to control the false discovery rate (FDR, default method);

mictools adjust $ODIR/pval.txt $ODIR

The command returns in the OUTPUT directory the following files:

pval_adj.txt

adjusted p values for each variable pair tested, in the same format of pval.txt

pi0_None.png

since the correction method is the Storey's qvalue, the command returns a plot with the estimated pi_0 versus the tuning parameter lambda:

Strength of significant associations

Finally, the strengths of the relationships called significant are estimated using MIC_e. By default the significance level is set to 0.05:

mictools strength $X $ODIR/pval_adj.txt $ODIR/strength.txt

The output file strength.txt is a TAB-delimited file, containing for each significant association the (corrected) TIC_e p-values, the Pearson's correlations, the Spearman's coefficients and finally the strengths, i.e. the MIC_e values:

Class	Var1	Var2	TICePVal	PearsonR	SpearmanRho	MICe
None	bullseye_x	bullseye_y	3.833704e-02	-0.068586	-0.078734	0.424553
None	circle_x	circle_y	4.723013e-04	-0.068343	-0.077292	0.631458
None	dots_x	dots_y	1.983666e-02	-0.060342	-0.126174	0.500185
None	slant_up_x	slant_up_y	1.593666e-02	-0.068609	-0.086098	0.355019
None	star_x	star_y	4.723013e-04	-0.062961	-0.051445	0.633117
None	x_shape_x	x_shape_y	4.723013e-04	-0.065583	-0.020535	0.566703

[Reshef2016]

Yakir A. Reshef, David N. Reshef, Hilary K. Finucane and Pardis C. Sabeti and Michael Mitzenmacher. Measuring Dependence Powerfully and Equitably. Journal of Machine Learning Research, 2016.

[Matejka2017]

J. Matejka and G. Fitzmaurice. Same Stats, Different Graphs: Generating Datasets with Varied Appearance and Identical Statistics through Simulated Annealing. ACM SIGCHI Conference on Human Factors in Computing Systems, 2017.