IGV signal visualization, low and high signals
0
0
Entering edit mode
8 weeks ago
Orange ▴ 30

Hello

I want to create a coverage plot from ATAC-seq using IGV. Is there a good way to make the plot such that both low and high signal peaks are visible?

To provide more context, I want to highlight locations with significant changes (defined with adjusted p-value and LFC from DESeq2). Bedgraphs (later converted to bigwigs) are generated from macs2 bdgcmp. Let's say at one peak the difference is 15 vs 7 and the other is 150 vs 70. If I set the scale to 0-20, the second peak gets cut off but if I set the scale to 0-200, the first peak becomes almost invisible.

Is there a good way to visualize differences in both low and high signal regions? log scaling on IGV is possible but I am wondering if there are other better strategies.

Thank you!

ATAC-seq IGV • 459 views
ADD COMMENT
2
Entering edit mode

You could also plot the fold changes,

Since you already know about log scaling what is the rationale for not using that? Because that scaling seems to fit your needs just right.

ADD REPLY
2
Entering edit mode

You can also just change the scale from locus to locus. So long as it's comparable across your samples and properly labeled, it's fine.

ADD REPLY
1
Entering edit mode

Agreed. You could of course work some bash magic to transform the coverage in the bigwig to some sort of Z-score that is independent of the absolute values, but much simpler is to simply stitch individual screenshots together in powerpoint or inkscape to make a nice plot. IGV often enough needs this manual postprocessing anyway to look somewhat nice and "publication-quality", as defaults are rather ugly.

ADD REPLY
0
Entering edit mode

Thank you all for your suggestions.

Is it acceptable to do this within a single gene? Let's say there are two sites (-10kb and +5kb) for gene A that I am interested in. Can I have different scales for the two sites and stitch them together? (as long as I clearly indicate so somewhere)

ADD REPLY
1
Entering edit mode

It is not the scales that matter but whether the plot is clear for the reader.

ADD REPLY

Login before adding your answer.

Traffic: 2015 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6