RPKM vs Fold Change

[ppm]

Hi,

I have been trying to make TSS plots. I see that the RPKM values for both the input and enriched ChIPseq samples are close while when I plot them for fold change, the global enrichment in one of the samples around TSS is more than the other. Also, I have two different cell lines, therefore, two different inputs as well.

THank you,

[ppm]

Would anyone know what could be the reason for this?

[dpryan]

My guess would be that it has something to do with how you're calculating RPKM. Is the "M" from the total mapped reads or just the total mapped reads in the regions that you're looking at?

[jparsons]

Why are you using RPKM here?

I'm not too familiar with your study, but I'd think that you want to be looking at windows of very specific width (around the TSSs, +/- N nt) and therefore will be comparing sequences of identical length. No need to length normalize when everything is the same length.

[ppm]

So, I have ChIPseq data for two cell-lines, therefore, two inputs.

I am plotting TSS plots +/-1500 bp. I am using ngsplot to do this.

When I just use enriched bam file or total input bam file, it gives me RPKM plotted on the Y-axis. These look very similar in case of both the cell lines.

Next, when I plot enrichedvstotalInput for these cell lines. I see that one of the cell line is more enriched than the other. I am not sure why would this happen. Views will be much appreciated.

Thanks