Deeptools - ComputeMatrix: Why does 'referencePoint center' gives the same heatmap as 'referencePoint TSS'

Question

[salamandra]

Why am getting exakly the same heatmap when using as reference Point the center of the peak or the TSS of the peak with 'Deeptools ComputeMatrix'?

With 'center':

 computeMatrix reference-point -S file_4heatmap.bw -R file_peaks_sorted_clip.bed -out file_matrixcenter.gz --referencePoint center --outFileSortedRegions file_regions4otherheatmaps.bed -b 4000 -a 4000 --missingDataAsZero -bs 50 

 plotHeatmap -m file_matrixcenter.gz -out HDFd2refcenter.png --dpi 300 --colorList '#FFFFFF,#FDF0E3,#FCE1C7,#FBD3AB,#FAC48F,#F8B573,#F7A757,#F6983B,#F58A20' --dpi 300 --whatToShow 'heatmap only' --refPointLabel 'Peak center'

With 'TSS':

the same code just changed

--referencePoint center

to

 --referencePoint TSS

[Friederike]

can you show the results? either the images or gunzip -c file_matrix*.gz | head

[Friederike]

also, a peek into the peak files might be helpful

[salamandra]

i feel a bit silly, they're actually a bit different... but still don't understand why in the peak center heatmap the darker areas are not aligned among different regions. They should be right? https://imgur.com/a/5sPUa

[Friederike]

they don't need to align perfectly, after all, this will depend on the structure of the peak and how the peak center was actually being determined.

you may wan to decrease the bin size (perhaps to 10bp), which will make the plot a little more smooth, but basically what these plots reveal is that there are some peaks with multiple local maxima