MultiRepMacsChIPSeq - run_DESeq2

run_DESeq2.R

This script will run a basic DESeq2 differential analysis between two conditions to identify differential (or enriched in the case of ChIP and Input) ChIPseq regions.

It requires an input text file with chromosome, start, and stop columns (or a coordinate name string), along with columns of fragment counts for both ChIP1 and ChIP2 (or Input reference) sample replicates. DESeq2 requires at least 2 (ideally more) replicates per condition to estimate variance for normalization and significance testing.

Count data may be already normalized to a uniform genomic depth, in which case it should indicated as such and SizeFactors will be set to 1. Otherwise counts will be normalized by DESeq2. WARNING: Default normalization by DESeq2 may potentially erase any enrichment signal.

A sample condition file is required, consisting of two columns, unique sample identifiers and group names (conditions). If desired, a third column could be added as an additional batch factor. Only two groups are used in the contrast, defined with the --first and --second options; any additional groups and samples are ignored.

Results are filtered at the indicated threshold or alpha level (False Discovery Rate). Significant regions are written to a text file and respective bed files. Log2 Fold Changes may be shrunken to reduce the effect of low count peaks. If additional filtering on Log2FC will be performed, the values should be shrunk.

USAGE:

run_DESeq2.R [options]

OPTIONS:

-c COUNT, --count=COUNT
	Input file containing count data

-a SAMPLE, --sample=SAMPLE
	Sample condition file containing identifiers and conditions

-o OUTPUT, --output=OUTPUT
	Output file basename, default 'first_second' names

-f FIRST, --first=FIRST
	Name of first ChIP condition

-s SECOND, --second=SECOND
	Name of second ChIP condition or Input reference

-b, --batch
	Use third column in customized sample table as batch factor

-t THRESHOLD, --threshold=THRESHOLD
	Threshold adjusted p-value (alpha) for significance (FDR), default 0.1

--norm
	Input counts are already normalized

--all
	Report all windows, not just significant

--shrink
	Shrink Log2FC values for low count peaks

-h, --help
	Show this help message and exit