I want to make a genome segmentation for a cell type of interest by using 6 different histone marks. Unfortunately, only 5 histone marks are available for that cell type so I decided to first use ChromImpute to impute the signals for the missing histone marks. After following the manual of ChromImpute (Manual), I obtained the imputed signal for histone mark of interest (one .wig file for each chromosme).
Next step is to binarize these signal values by using
BinarizeSignal from ChromHMM. According to the ChromHMM reference manual (ChromHMM manual) it can be done by using BinarizeSignal. I prepared the input files according to the given information for 2 chromosomes (input files looks like below) to first test if everything goes well:
HEK293 Chr10 H3K27me3 0 0.1 1.25 HEK293 Chr11 H3K27me3 0 0.52 2.37
I used the following command for binarization of signal files (the input files have _signal_ in their names as well, to be precise, filenames are: chr10_signal_HEK293_H3K27me3.wig and chr11_signal_HEK293_H3K27me3.wig):
java -jar ChromHMM.jar BinarizeSignal ./inputdir ./outputdir
I got the following output:
Writing to file ./outputdir//HEK293_chr10_binary.txt Writing to file ./outputdir//HEK293_chr11_binary.txt
But when I checked the binarized files they are just
0 values, there is not a single
1 in the entire files.
When I checked my input signal files which are binned into 25 bp windows, many of the bins have non zero values (e.g. 0.2, 1.3, 0.2 etc..). So, I think I'm making a mistake in using the
Can anybody guide me how to solve this issue? Thank you.