MEDICI.ReadMe.txt

MEDICI Documentation (Contact: lee.cooper@emory.edu)

Mining Essentiality Data to Identify Critical Interactions (MEDICI) is a method to identify essential 
protein-protein interactions in cancer. This documentation describes the software available in the MEDICI
Github repository (http://github.com/cooperlab/MEDICI)

If you find this code useful, please cite our paper ()

PAPER CITATION HERE

** Overview ******************************************************************************

MEDICI integrates pathway models and gene essentiality data to understand which network interactions 
are critical for the proliferation of cancer cells. Interactions such as MDM2/TP53 are gaining attention 
as potential targets for cancer therapy, as they present a highly specific means to interrupt oncogenic
activity that results from genetic alterations. Despite the significance of network interactions, there 
are currently no high-throughput experimental platforms to interrogate the essentiality of individual 
protein-protein interactions (PPIs). MEDICI aims to provide predictions of the essentiality of network 
interactions to aid in prioritizing them as potential therapeutic targets.

MEDICI is written in Matlab, and consists of several collections of modules that implement the following 
functionality:

-Superpathway Generation - creating a large topological model of network interactions from smaller pathways
-Generating Gene-Centric Essentialities - aggregating RNAi data from multiple probes into gene-centric 
											essentiality scores.
-Interaction Essentiality Inference - combining gene-centric essentialities with superpathway models
										to infer the essentialities of individual interactions.
-Correlative Analysis - data mining to identify associations between interaction essentialities,
						drug sensitivities, and genetic alterations
- Cytoscape Visualization - Cytoscape visualization of gene and interaction essentialies

** Running MEDICI ************************************************************************

1. Generate Superpathway Model
	
	Generate a generic, baseline superpathway model. This model will be tuned for each cell line based 
	on it's genetic alterations extracted in step 3.
	
	Run GenerateSuperPathway.m providing the pathway and filenames of the pathway gene symbols,
	symbol interactions, pathway inclusion list, HUGO database file, PPI screening file (optional).
	
2. Generate Gene-Centric Essentiality Scores

	Transform RNAi screening results into gene-centric essentiality scores that will be used for interaction 
	essentiality prediction.
	
	Run GeneCentricEssentiality.m providing the RNAi GCT file, the cell line description file, 
	HUGO database file, number of permutations for estimating score null distributions, and 
	an optional file containing pre-computed null models (to cut down on computation time).

3. Extract Cell-Line Genetic Profiles

	Extract mutations and copy-number variations from Cancer Cell Line Encyclopedia data to
	build context-specific superpathways for each cell line.
	
	Run CCLEBuildCNV.m and CCLEHybridCapture.m providing the CNV text file and hybrid capture MAF.


4. Prediction of Interaction Essentialities
	
	Run CalculateEssentialities.m providing the context-specific superpathways, Gene-Centric Essentiality Scores,
	Template Reactions, and parameters alpha and w.
	
	Run ScaleEssentialities.m providing the predicted essentialities.
	
	Run PrintTable.m providing Superpathway, Scaled Essentialies.


5. Correlative Analysis

	Run SensitivityAnalysis.m providing Scaled Essentilities and Drug Responds to generate the table of the correlations
	between PPI essentialities and drug sensitivities.
	

6. Generate Cytoscape Visualization Tables

	Run GenerateCytoscape.m to create a Cytoscape edge and node table for each cell line.
	Provide a link to a list of oncogenes and tumor suppressors to enhance the visualization.
	A Cytoscape .xml style file is provided that defines a style encoding interaction essentialities,
	gene essentialities, and oncogene / tumor suppressor status into the visualization.