Why IVG coverage track from bam and bigwig files looks different?

Question

[biplab]

Hi I am studying polyadenylation of ribosomal RNA in different yeast mutants. I am interested in knowing how does the coverage of ribosomal RNA locus in my oligo-dT purified samples changes in various mutants. After alignment by HISAT2, I converted bam file to bigwig file by deeptools with a default setting and viewed in IGV . I also viewed bam file directly in IGV they looks different. From bigwig file it seems to I have lots of read in rRNA locus. I will be really happy to know whey they took different? Looking forward to hearing from you.

[ATpoint]

I assume you have paired-end data? In that case, the bam file will make IGV print each read by itself, so the region between the two mates will not be covered. The bigwig joins the two mates and will therefore span the entire range between the two mates.

Example: Given you have 2x150bp reads and an average fragment size of 500, say an alingnment from chr1:10000-10500. The bam will be printed as 1) 10000-10150 and 2) 10350-10500 because that is the position of the individual reads. The bigwig will cover the entire 10000-10500 region.

[biplab]

Thank you. This is how they look like. Seeing no coverage along rRNA, still does not make sense to me. Both tracks came from same bam file. ![link for igv track image][https://ibb.co/furfaS]

Answer 1

[dariober]

Some wild guess...

bigwig file it seems to I have lots of read in rRNA locus

This may be due to some rescaling when converting bam to bigwig. For example, rescaling to have library size 10M reads will appear as having more reads compared to rescaling to have 1M reads. (Such rescaling can be useful to compare libraries sequenced with different read depth.)

I also viewed bam file directly in IGV they looks different.

This may be because deeptools applies some filtering on the raw reads? For example, on mapping quality, duplicates, etc...

Again, just guessing...