Dependencies

Make sure those dependencies are installed before you try to install the metageneInputs package:

require(devtools)
devtools::install_github("CharlesJB/FantomEnhancers.hg19", ref = "v0.2.0")
devtools::install_github("CharlesJB/FantomTSS.hg19", ref = "v0.2.0")
devtools::install_github("CharlesJB/metagene",
                         ref = "05b1e7594a72901b8ccd37eead171f56017693bc")
install_packages("dplyr")
install_packages("tidyr")

All other dependencies should be installed automatically during metageneInputs installation.

Introduction

The goal of this package is to show how to reproduce the input files that will be used for the Plos Computationnal Biology paper:

• Groups of regions
• Scripts to download bam files (Makefile)
• Designs

Usage

First, you need to install the package. This could take ~10-30 minutes depending on your machine since the whole analysis will be done during the package installation (i.e.: none of the results of the analysis are saved on the github repository).

require(devtools)
# TODO: Add ref when available
devtools::install_github("CharlesJB/metageneInputs", build_vignettes = TRUE)

To avoid accidently overwritting files with similar names to those produced by this package, the results are saved in a temporary directory. The name of the directory changes each time the package is built. The name of the directory used for the analysis is saved in the vignettes. To view the vignettes, from a R session:

browseVignettes("metageneInputs")

Input files descriptions

splitted_regions.RData

This object contains 2 GRangesLists: one for enhancers and one for promoters. Enhancers are regions are obtained with the FantomEnhancers.hg19 package and promoters are obtained with the metagene package.

Each element of the GRangesList correpond to an expression group in GM12878 cell lines (based on Famtom5 datasets):

• (0: not expressed)
• (25: First quartile)
• (50: Second quartile)
• (75: Third quartile)
• (100: Fourth quartile)

We calculate the tresholds of expression of each group (in transcript per million (TPM)) using the non-zero values from every Fantom5 transcription start sites (TSS; using the FantomTSS.hg19 package) for every available cell lines except GM12878 and splitting them in quartiles. Regions from GM12878 cell lines are then splitted in each groups based on their TPM value.

Makefile

We need to download ENCODE datasets from the GM12878 cell line (everything but histones). Using the ENCODExplorer package, we extract the URLs for every files and save them in a Makefile. We also saved the URL for the bam files to use for the validation in the same Makefile.

To download the files, simply copy the Makefile in the directory where you want to download the files and type in a shell:

make

This will download all the bam files required to reproduce the metagene analysis (~300G).

designs.RData

This objects contains 2 elements:

• designs_encode: A list with the designs for every ENCODE experiments to be analyzed.
• design_validation: The design for the validation.

tpm_grangeslist

This file contains a GRangesList named TPM that consists of 2 GRanges, one for the promoters and one for the enhancers. Each GRanges contains the mean TPM value for each regions.