Calculate the ratio of H3K4me3 to H3K4me1 using output of deeptools computeMatrix

Question

[jia]

Hi, everyone！ I have a list of peaks in bed format and bigwig track of H3K4me3 and H3K4me1. I want to calculate the ratio of H3K4me3 to H3K4me1 in regions near these peaks. What came to my mind at first is deeptools computeMatrix. Here is the code:

computeMatrix reference-point \
-S                              H3K4Me1_Kcl_mm10.bw H3K4Me1_Veh_mm10.bw \
                                 H3K4Me3_Kcl_mm10.bw H3K4Me3_Veh_mm10.bw \
                           -R   Increasing.bed  --referencePoint center \
--binSize 1 --skipZeros -a 2000 -b 2000 -p 10 -o matrix_bin_1_ratio_inc.gz

now I have the output matrix of computeMatrix and here is the header of the matrix

@{"upstream":[2000,2000,2000,2000],"downstream":[2000,2000,2000,2000],"body":[0,0,0,0],"bin size":[1,1,1,1],"ref point":["center","center","center","center"],"verbose":false,"bin avg type":"mean","missing data as zero":false,"min threshold":null,"max threshold":null,"scale":1,"skip zeros":true,"nan after end":false,"proc number":10,"sort regions":"keep","sort using":"mean","unscaled 5 prime":[0,0,0,0],"unscaled 3 prime":[0,0,0,0],"group_labels":["genes"],"group_boundaries":[0,3076],"sample_labels":["H3K4Me1_Kcl_mm10","H3K4Me1_Veh_mm10","H3K4Me3_Kcl_mm10","H3K4Me3_Veh_mm10"],"sample_boundaries":[0,4000,8000,12000,16000]}

As I'm not familiar with 'awk' and 'expr' command in bash , I read the matrix in R

inc_raw = read.delim("matrix_bin_1_ratio_inc.gz", skip=1, header=F)
inc =inc_raw[,-c(1:6)]
inc_H3K4Me1=inc[,c(1:8000)]
inc_H3K4Me3=inc[,c(8001:16000)]
inc_ratio=inc_H3K4Me3/inc_H3K4Me1

Then I don't know how to export the 'inc_ratio' to the format like output matrix of computeMatrix because I want to use deeptools plotProfile to show the ratio in the regions. Could anyone give me some help about this? Thanks!

[Devon Ryan]

Have you considered using bigwigCompare? It's the more typical way of doing this and produces a bigWig file compatible with computeMatrix.

[jia]

wooow!! thank you so much!