Strand specification problem at deeptools (computeMatrix)

Question

[miso5103]

Hello,

I trying to quantify the ChIP-seq data using deeptools, computeMatrix and plotHeatmap.

But, I got only symmetrical graph which has double peaks, both TSS and TES.

I confirmed that this kind of graph is resulted from strand problem by using "+ strand sorted bed file" (The only one peak remain)

The question is, how can I separate two kind of strand, Watson and Crick strand, and merge them with right direction?

Or, is there any step that distinguish Watson and Crick strand in processing computMatrix?

I attempted to solve the problem by computeMatrixOperation - filterStrand but there is another error,

(ValueError: need at least one array to concatenate)

[colin.kern]

What are the regions you're defining in your bed file that you use for the -R parameter? Can you edit your question and add a few lines from this file?