For background information about the GC bias assessment and correction, see computeGCBias.

Usage example


correctGCBias requires the output of computeGCBias and a genome file in 2bit format. Most genomes can be found here: Search for the .2bit ending. Otherwise, FASTA files can be converted to 2bit using faToTwoBit, which is available here:

$ correctGCBias -b H3K27Me3.bam
   --effectiveGenomeSize 2695000000
   --genome genome.2bit
   --GCbiasFrequenciesFile freq_test.txt # output of computeGCBias
   -o gc_corrected.bam


The GC-corrected BAM file will most likely contain several duplicated reads in regions where the coverage had to increased in order to match the expected read density. This means that you should absolutely avoid using any filtering of duplicate reads during your downstream analyses!

deepTools Galaxy. code @ github.