Differences in DeepTools Heatmap Resolution Driven by Bed File

Question

[gcooper1245]

Hi! I'm using DeepTools to plot coverage over a list of significantly differential accessible peaks from ATAC-seq data. However, I'm running into a strange issue: the resolution of the heatmap seems to vary depending on which BED file I use, even though the underlying BigWig files and sequence data are exactly the same. Here is my code I am using to plot this:

>     computeMatrix reference-point \
>     -S 1_R1_final.bw \
>        1_R2_final.bw \
>        2_R1_final.bw \
>        2_R2_final.bw \
>        3_R1_final.bw \
>        3_R2_final.bw \
>       -R gain.bed
>       loss.bed \
>     --referencePoint center \
>      -a 2500 -b 2500 --outFileName matrix_compute.tab.gz
>     
>     plotHeatmap \
>      -m matrix_compute.tab.gz  \
>      -out heatmaps \
>      --heatmapHeight 15  \
>      --zMax 15 \
>      --missingDataColor 1 \

Both BED files seem properly formatted and I don’t see any glaring differences between them. However, the resulting heatmap from the second BED file looks much more pixelated and of lower resolution.

Has anyone encountered something like this before? Any suggestions on how I can make the resolution more consistent across different BED files? The current difference makes it challenging to publish the figures since they look so different.

Heatmap with high resolution: BED file used here (31,568 lines): chr2 53505719 53506219 chr5 95712381 95712881 chr7 22786177 22786677 chr6 163745824 163746324 chr3 46114243 46114743 chr16 77298533 77299033 chr6 154555094 154555594 chr7 7099406 7099906 chr14 35607752 35608252 chr3 141084100 141084600 ...

Heatmap with low resolution: BED file used here (36451 lines): chr3 23981877 23982377 chr19 27716797 27717297 chr8 6503164 6503664 chr9 96499044 96499544 chr5 103923705 103924205 chr17 39177255 39177755 chr3 170438004 170438504 chr19 27995789 27996289 chr12 65525854 65526354 chr6 129369187 129369687 chr19 53892424 53893417 chr15 77859264 77859764 ...

Any suggestion on how to correct this and make it consistent between the two different regions.