PARADIGM: Pathway Recognition Algorithm using Data Integration on Genomic Models

Current Version

1.0

Authors

Charles Vaske, Stephen Benz, Sam Ng, Kyle Ellrott, David Haussler and Joshua M. Stuart

Requirements

• python >= 2.7
  • scipy >= 0.12.0
  • numpy
  • pandas

Installation

Note: This repository contains code that allows users to parallelize PARADIGM on a large cohort of samples, the PARADIGM binary is separate and may need to be compiled from https://sbenz.github.com/Paradigm

• Install dependencies
• Download the paradigm-scripts repository to the desired location
• Run “make” in paradigm-scripts/ to generate source files
• Source init files for paradigm-scripts (init.sh for bash and init.csh for csh)
• Run code on example data in examples/ with “make”

Command-Line

galaxyParadigm.py [options] attachment file:path [attachment file:path ...]

attachment - dogma node to attach evidence to
file:path - type of and path to evidence file

-w work_directory - path to directory where work is to be done (default: ./)
-n null_size - number of null samples to be generated per sample
-d dogma_library - path or zip file describing the PARADIGM plate model
-p pathway_library - path, zip, or file containing PARADIGM pathway interactions
-b discretization_bounds - value cutoffs for data discretization
-t param_file - parameter file to use as initial state
-s - skip EM parameter training
-y - utilizing the public PARADIGM binary

Folders

• bin : executables
• dogmas : dogma plate models for PARADIGM
• examples : OV inputs for demonstration purposes
• exe : pre-compiled paradigm binaries for select architectures
• pathways : paradigm compatible pathway files