the conflict about the pheatmap and multibigwigsummary results

[liuyu988]

hi, everyone!
I recently analyzed the chip-seq data for histone marks. I am very confused with the results for pheatmap and multibigwigsummary results.
I plotted the enrichment level in the same genomic regions for WT and KO condition. from the profile or heatmap, I can get the conclusion that the average signal in WT will be higher than the ones in KO for some regions. however, I calculated the average scores for these regions by multibigwigsummary function, but I got the opposite result. To check the result, I observed the signal by IGV software. the observations were same as the heatmap, but opposite with the multibigwigsummary.

my code is as follow:

multiBigwigSummary:

multiBigwigSummary BED-file
--bwfiles WT.bw KO.bw
--BED file.bed
-out all.scores.npz --outRawCounts all.scores.tab
-p 35

plotHeatmap:

computeMatrix scale-regions -p 35 -S WT.bw KO.bw \
     -R file.bed -b 0 -a 0 -m 10000 --missingDataAsZero  -o all.mat.gz

plotHeatmap -m all.mat.gz -o all.pdf --interpolationMethod bicubic

my results:

igv：

multiBigwigSummary:

chr start end KO WT
chr4 155293554 155315474 24.22877 10.359672

could any one help me ? thanks a lot !
best wishes!
Yu

[WardDeb]

Are you sure you haven't flipped any labels somewhere ? The order in your multibigwigsummary is reversed from what your original command implies, and it seems to be post processed in some matter. If not, please share some example files to replicate the problem.