Differential loop analysis

[helenhuangmath]

Hello,

I'm trying to run differential loop analysis using the script Imp_Scripts/DiffAnalysisHiChIP.r on a 2-conditions but only 1 replicate per condition data, without ChIP bam files.
However I got 0 significant loops identified. By checking the all differential loop results, I found all tested loops FDR results are 1. Do you have any suggestions why?

Here's my code for running the differential analysis.

Rscript ~/FitHiChIP/Imp_Scripts/DiffAnalysisHiChIP.r \
--AllLoopList ${Ctrl_loop},${Case_loop} \
--ChrSizeFile hg19.genomesize \
--FDRThr 0.01 \
--OutDir DELoops_byFitHiChIP_rerun \
--CategoryList Ctrl,Case \
--ReplicaCount 1,1 \
--FoldChangeThr 2 \
--DiffFDRThr 0.1 

The loops I used are from upstream results: FitHiChIP_Peak2ALL_b5000_L20000_U2000000/P2PBckgr_0/Coverage_Bias/FitHiC_BiasCorr/FitHiChIP.interactions_FitHiC.bed
I'm wondering why the FDRs I got are all 1? Is there anything I need to change in the code? I appreciate your suggestions!
Thank you!

[ay-lab]

Hi @helenhuangmath
It may be due to the issue that HiChIP counts between different conditions may not be drastically different. Could you please check the output file "Loops_EdgeR_Default.bed" (containing complete edgeR results) and specifically the entries with p-value < 1e-5?