Bay2Ctrls: Bayesian model to integrate two types of ChIP-seq controls for binding site detection

Two types of controls were widely used for ChIP-seq: DNA input controls and mock IP controls. This software was developed to combine both for better binding site detection.

This R package is based on SPP, and the wrapper script (run_Bay2Ctrls.R) for the R package is adapted from the ENCODE ChIP-seq pipeline.

DATA INPUT

The software accecpt four files as input:
The first set is the DNA input controlled IP experiment. This DNA control (di) and the IP (i) are the results from conventional ChIP-seq.
The second set is the mock IP experiment. This includes the mock IP (m) and its corresponding DNA input (dm).

DATA OUTPUT

Binding peaks are in the narrow peak file format

DATA DOWNLOAD

Download the test data here
All raw and processed data are available here

DEPENDENCE

• SPP
• caTools

Tested with

• R 3.6.1
• SPP 1.16.0
• caTools 1.18.0

Please note that you might encounter a lot of issues if you want to install SPP from source. We recommend you install it via R install.package("spp").

INSTALLATION

devtools::install_github("gersteinlab/Bay2Ctrls")

USAGE

Run the wrapper script run_Bay2Ctrls.R as the following:
Rscript run_Bay2Ctrls.R -ip=<IP_file> -mock=<mockIP_control_file> -input4ip=<DNA_input_control_file> -input4mock=<DNA_input_control_for_mockIP_file> [Bay2Ctrls_parameters, ...] [SPP_parameters, ...]

Example

Rscript run_Bay2Ctrls.R -ip=test_data/a.rep0.tagAlign.gz -mock=test_data/EMb1.rep0.tagAlign.gz -input4ip=test_data/ap.rep0.tagAlign.gz -input4mock=test_data/EMb1p.rep0.tagAlign.gz -npeak=30000 -x=-500:85 -s=0:5:1200 -odir=./ -filtchr='.*_[CD].*' -savr -savp -rf -out=test.cc -npeak=30000 -totReads=10000000 -mcstep=1000000

You might also use the R package Bay2Ctrls in your scripts. See man pages for individual functions(TODO).

Paramters specific for Bay2Ctrls

-ip            a *.tagAlign.gz file containing mapped reads from the IP experiment (i)  
-mock          a *.tagAlign.gz file containing mapped reads from the mock IP experiment (m)  
-input4ip      a *.tagAlign.gz file containing mapped reads from the DNA input control for the IP experiment (di)  
-input4mock    a *.tagAlign.gz file containing mapped reads from the DNA input control for the mock IP experiment (dm)  
-totReads      scaling the numbers of reads of the four experiments to the same level, e.g. -totReads=10000000
-mcstep        simulation steps for the Bayesian model, e.g. -mcstep=1000000

Intrinsic parameters of SPP

-npeak         the number of total narrow peaks to be output, e.g. -npeak=30000  
-x             avoid phantom peaks, e.g. -x=-500:85  
-s             genomic distance for signal correlation, e.g. -s=0:5:1200  
-out           output log file  
-rf            remove previous results  
-odir          output directory  
-p             number of processors to be used

More options for output format please refer to the spp R package github or CRAN