Reads detected/"spill over" on introns after using bamCoverage function from deepTools

Question

[kdcastillo]

Hello! We recently used the recent version of bamCoverage from the deepTools suite for our ribo-seq and RNA-seq bam files to generate bigwig files. The conversions ran smoothly, however when we loaded the bigwig files into IGV, we noticed some reads spilling over to some exon-intron junctions when we zoomed into each track. We believe these are not genome misannotations because the non-normalized/non-bigwig bam files don't show these junction reads. Any advice is appreciated. Thank you!

Best, Kat

[ATpoint]

Hello, it would help if you add code and screenshots to understand the problem. My first guess when it comes to local anomalies is the chosen bin size and extension parameter (-bs -e).

[kdcastillo]

Hello! I agree that it might be due to bin size, I tried the default and bin size of 10, and I am still getting the same issue. But I attached the script (for bin size 10) and the example image of the problem I'm seeing.

bamCoverage -b filename.bam -o filename.bin10rpkm.bw --normalizeUsing RPKM -bs 10

Thanks for the response!

Kat

[cpad0112]

Is this riboseq data? can you check the length of the retained bases? If it is riboseq, you may have to set offset.

[FCarranz]

Did changing the bin size end up fixing this issue for you?

Franco