A change-point model for detecting 3'UTR changes by RNA-Seq

Usage

    perl change_point.pl -t <treatment.bam> -c <control.bam> 
                         -g <utr.bed> -d <s|l> [options]

    Required:

     -t     FILE    A bam file for treatment sample, which can be 
                    obtained from RNA-Seq aligner, such as Tophat. 

     -c     FILE    A bam file for normal sample, which can be 
                    obtained from RNA-Seq aligner, such as Tophat. 

     -g     FILE    A bed file for regions to be analyzed, such as 
                    3utr regions. The website provides 3utr bed files 
                    for mouse mm9 and human hg19, which contain unique 
                    regions parsed from RefSeq annotations. This step 
                    will be included in the pipeline in the future. 

     -d     STRING  Analysis type, currently, s for detecting shortening 
                    events or l for detecting lengthening events.

    Options:

     -o     STRING  The prefix for output file. The default value is 
                    change_point.

     -n     INT     At least how many reads support each region. The 
                    default is 20. In either condition, if any region 
                    has reads smaller than this threshold, the region 
                    will be dicarded. 

     -a     FLOAT   Mixed directional FDR level. The default is 0.05. 

     -r     INT     The hypothesized odds ratio in Fisher's Exact Test. 
                    The default is 2, meaning that the program will tend 
                    to select significant changes with odds ratio greater 
                    than 2. 

     -s     STRING  The maximum heap size for JVM. The default is 256m.

     -x     STRING  The minimum heap size for JVM. The default is 1024m.

    Example:

     perl change_point.pl -t treatment.bam -c normal.bam -g 3utr.bed \
                          -d s -o output
			

Understand Output

COLUMN MEANING
CHR   chromosome name
START   start position
END   end postion
GENE   gene name
STRAND   strand information
PM   reads count in proximal region of treatment sample
PN   reads count in proximal region of combined sample
DM   reads count in distal region of treatment sample
DN   reads count in distal region of combined sample
LAMBDA_MAX   lambda in change point position
TAU_HAT   estimated tau, the predicted change point
CP_PVAL   p-value generated by change-point model
FISHER_PVAL   p-value generated by fisher's exact test
ODDS   odds ratio in change point position
DEC   final decision based on the specified mdFDR level