Deeptools comparing heatmap zMin and zMax across different plotHeatmaps

Question

[connorjfausto]

Hello!

I am in the process of making different heatmaps of ChIPseq bigwig files from the same publication to check chromatin binding patterns, and had a question regarding the zMin and zMax. I am not using the entire mm10 genome as an input, but rather a list of genes my colleagues generated that are of interest to our biological question.

To make my example more general, my groups of bigwig files are group 1 and group 2, and I did the following:

1. computeMatrix with the following parameters, noting that the only differences are the input bw files

computeMatrix reference-point \
-b 10000 -a 10000 \
-R induction.gtf \
-S K27ac-5uM-rep1.bw K27ac-5uM-rep2.bw K27me3-5uM-rep1.bw K27me3-5uM-rep2.bw \
--skipZeros \
-o matrix-induction-chromreg.gz \
-p 6 \
--outFileSortedRegions regions-5uM-induction-chromreg.bed

computeMatrix reference-point \
-b 10000 -a 10000 \
-R induction.gtf \
-S Cttnb1-5uM-rep1.bw Cttnb1-5uM-rep2.bw  Lef1-5uM-rep1.bw Lef1-5uM-rep2.bw Tcf7-5uM-rep1.bw Tcf7-5uM-rep2.bw \
--skipZeros \
-o /project/nilslind_369/cfausto/distal_genes/output/5uM-CHIR/10kb/matrix-induction-bcattarget.gz \
-p 6 \
--outFileSortedRegions /project/nilslind_369/cfausto/distal_genes/output/5uM-CHIR/10kb/regions-5uM-induction-bcattarget.bed

2. Two separate plotHeatmaps using the different matrix files generated, but the same zMin and zMax

plotHeatmap -m matrix-induction-bcattarget.gz \
-out 5uM-induction-bcattarget-heatmap.png \
--colorMap RdBu \
--zMin -3 --zMax 5 \
--whatToShow 'heatmap and colorbar' \
-T '5uM Induction' \
--kmeans 6 \
--outFileSortedRegions 5uM-induction-bcattarget-clusters.bed

plotHeatmap -m matrix-induction-chromreg.gz \
-out 5uM-induction-chromreg-heatmap.png \
--colorMap RdBu \
--zMin -3 --zMax 5 \
--whatToShow 'heatmap and colorbar' \
-T '5uM Induction' \
--kmeans 6 \
--outFileSortedRegions 5uM-induction-chromreg-clusters.bed

By doing this, can I assume my zMin and zMax are comparable across the two different heatmap png generated?

[Marco Pannone]

By changing the arguments --zMin and --zMax you simply change the heatmaps intensity values. This is something you tune accordingly to the intensity of the signal you are plotting and it has a purely visual purpose. Naturally, if you want to compare two heatmaps of two different sample groups, you would go for same --zMin and --zMax values as you did in your command line.

[connorjfausto]

Hi Marco,

Thanks for the reply! So even if the visualization intensity comes from different matrix files, the intensity can be compared? Would this be similar to using the same intensity score for featureplots of different genes with scRNAseq data?

[Marco Pannone]

I would adjust --zMax based on the heatmap with the highest intensity among the ones to be compared. Plotting also a profile plot on the top of the heatmap helps you to have an idea of what is the highest intensity value in an heatmap.