deeptools reference-point center

Question

[modockesner]

Hi, For some reason using deeptools computematrix reference-point center I get instead of TSS and TSE positions I am getting one mark that says center. I'm not sure what is causing the issue?

computeMatrix reference-point -p 1 --referencePoint center -R bedfiles -S wigfiles --binSize 5000 -b 5000 -a 5000 --skipZeros --missingDataAsZero 

plotHeatmap -m matrix --missingDataColor "white" -out x.pdf

(base) bash-4.1$ deeptools --version

deeptools 3.5.1 and other verions.

[Friederike]

Can you explain the issue a bit more? What type of visualization/matrix are you after and what are you getting? The gallery has some example plots that use different reference points, maybe those are helpful to illustrate the problem you have.

[modockesner]

I have a run of deeptools that is working and it header in matrix looks like this:

@{"verbose":false,"scale":1,"skip zeros":true,"nan after end":false,"sort using":"mean","unscaled 5 prime":[0,0,0],"body":[5000,5000,5000],"sample_labels":["i","j","l"],"downstream":[5000,5000,5000],"unscaled 3 prime":[0,0,0],"group_labels":["X.bed","Y.bed","X.bed","W.bed"],"bin size":[5000,5000,5000],"upstream":[5000,5000,5000],"group_boundaries":[0,10,14,18,28],"sample_boundaries":[0,3,6,9],"max threshold":null,"ref point":["center","center","center"],"min threshold":null,"sort regions":"keep","proc number":1,"bin avg type":"median","missing data as zero":true}

However, the one from the header not working is using 1 less -R bed file as input and its header looks like and different -S. Its header looks like this:

#from normal
@{"verbose":false,"scale":1,"skip zeros":true,"nan after end":false,"sort using":"mean","unscaled 5 prime":[0,0,0],"body":[0,0,0],"sample_labels":["i","j","k"],"downstream":[5000,5000,5000],"unscaled 3 prime":[0,0,0],"group_labels":["X.bed","Y.bed","Z.bed"],"bin size":[5000,5000,5000],"upstream":[5000,5000,5000],"group_boundaries":[0,11,530,899],"sample_boundaries":[0,2,4,6],"max threshold":null,"ref point":["center","center","center"],"min threshold":null,"sort regions":"keep","proc number":1,"bin avg type":"median","missing data as zero":true}

You can see that the body here has values [0,0,0] where as the working one has body [5000,5000,5000]. I'm not sure why the headers different. I basically use the same parameters and version of deeptools.

Answer 1

[Devon Ryan]

This is the intended behavior. --referencePoint center makes the reference point the middle of each feature of interest. This is then not the TSS (the start of the feature) or the TES (the end of the feature), so we just label it "center". If you want to use the TSS or TES as the reference point instead, then specify that rather than --referencePoint center.