iTARGEX

This is git repo for iTARGEX (identification of Trait-Associated Regulatory Genes via mixture regression using EXpectation maximization). iTARGEX is a bioinformatics tool to help biological researchers for in silico prediction of novel regulators linked to biological traits. iTARGEX is mainly writen in R with embedded code in C for accelerating EM. In order to execute iTARGEX, the packages and its dependencies must be installed separately.

Install

First, you need to install R enviroment in your computer. iTARGEX utlized Rcpp to accelerate the numerical calculation, so you need to install C++ compiler (e.g. gcc, clang). Next, you can install R packages and its dependencies using the following command.

pkgs <- c("data.table", "docopt", "doParallel", "foreach", "mixtools", "Rcpp", "scales", "weights")
install.packages(pkgs)

Quick Start

The simplest method to run iTARGEX is run itargex.R script without any options.

./bin/itargex.R trait.txt deletome.txt output_dir

This may take a few minutes to hours depending on your machine spec and data size. After execution, you can find the result under output_dir. The siginicant regulator can be found in cor_sig_local.csv. You can use any csv reader such as Excel or Libreoffice to open it. The first column is the name of signifiant regulator, followed by several columns.

Column name	Description
cor	Weighted pearson correlation coefficient
beta	Slope of the significant component
pval_beta	Significant level of beta (or cor)
adjp_beta	Bonferroni-adjusted p-value
ratio	Proportion of datapoints estimated in the significant component
iter	Iteration times during EM convergence

Advance usage

iTARGEX can be executed with serval options.

iTARGEX (Identify trail associated regulator via expectation–maximization).

Usage:
  itargex.R [options] <input-trait> <deletome-data> <output-folder>
  itargex.R em [--iter=<int> --ncpus=<int>] <input-trait> <deletome-data> <output-folder>
  itargex.R analysis [--ratio=<float> --ncpus=<int>] <input-trait> <deletome-data> <output-folder>
  itargex.R -h | --help
  itargex.R --version

Options:
  -h --help        Show this screen.
  --version        Show version.
  --ratio=<float>  Min ratio of regulator [default: 0.2].
  --iter=<int>     Max iteration time of EM algorithm [default: 1000].
  --ncpus=<int>    Number of CPU for parallel execution [default: 4].

If you want to execute iTARGEX for many times to search for the global optima of mixture models. You can run iTARGEX in two-step style to save the runtime.

# run EM step to esitmate the mixture weight of each datapoints
./bin/itargex.R em trait.txt deletome.txt output_dir
# run analyis step to identify significant regulators
./bin/itargex.R analysis --ratio 0.2 trait.txt deletome.txt output_dir

In addition, You can execute the second command for many time with different --ratio. Please notice that for any time the model runs it will overwrite the original result. We suggest you copy the ouput folder to other place to prevent overwrite.