RCADER

Recognition Code-Assisted Discovery of regulatory Elements by Regression

RCADER uses ChIP-seq data and the C2H2-ZF recognition code to identify motifs that represent the binding preferences of C2H2-ZF proteins. It uses an iterative regression-based approach to optimize the motifs that are predicted by the recognition code, so that the correlation of motif scores and the ChIP-seq peak heights are maximized.

Requirements

• Unix-compatible OS
• R version 3.0.1 or later
• R randomForest library
• GNU Scientific Library version 1.15 or later

Installation

From the command line, go to the root folder of RCADER, and run make. This will compile the codes and will create the RCADER executables in ./bin/.

Running RCADER

Input files

For running RCADER, you need the following files:

• C2H2-ZFP sequence file: The protein sequence of the C2H2-ZF protein that is targeted by ChIP-seq (FASTA format).
• Peak sequences: The DNA sequences of the peak regions (FASTA format). We recommend the use of ±250 bp region around the peak summits.
• Peak scores: The scores (e.g. fold-enrichment) associated with peaks (tabular format). The first column represents peak names (must match the names in the peak sequence file), and the second column represents the scores (must be >0).

Usage

To run RCADER, go to the main folder of the package, where the file RACDER.sh is located. Then simply use the following command:

bash RCADER.sh <jobID> <C2H2_ZFP.fasta> <ChIP_seq.fasta> <ChIP_seq_scores.txt>

The <jobID> parameter is a name that will be used to create the output folders.

Output

RCADER creates the following main output files in the ./out/<jobID> folder:

• results.report.txt: A table with a summary of optimization results for each of the recognition code seed motifs.
• results.ps: A PostScript file containing a graphical representation of each seed and optimized motif.
• results.PFM.txt: A text file containing the optimized motifs as position frequency matrices (PFMs).
• results.scores.txt: A table with the original peak scores, the nucleotide and dinucleotide frequences of each peak, and the score of each optimized motif for each peak.

Example datasets

Example datasets are provided at ./examples, including a subsampled set of peaks for CTCF ChIP-seq. Use the following command to run RCADER on this example dataset:

bash RCADER.sh example_CTCF ./examples/CTCF/CTCF.fasta ./examples/CTCF/summits.fa ./examples/CTCF/macs.subsampled.uniform.txt

This will create the ./out/example_CTCF folder, which will contain the RCADER output files.