Interpreting HOMER peak calling score and annotation file

Question

[Daniel]

Hello,

I used homer for peak calling and peak annotation for two different conditions. I want to count the differential peaks, which are the peaks that exist in one file but not the other. My question is: Do I simply see which peaks occur in one file but not the other? It seems a little too simple to be correct...

I know there is a peak score column, but I couldn't find explanations online on whether I should use this for filtering.

I also want to use the distance to TSS column, but am having trouble understanding how to use it for filtering, too. If homer defines the TSS as a range (num1, num2), what does a distance to TSS of 2000 mean? Is it the distance from the center of the peak to num2?

Thank you and apologies for super basic questions.

Answer 1

[rfran010]

You could identify peaks present in one or the other file, however, sometimes these differences can arise from technical reasons.

A common practice is to take only the consensus peaks (peaks present in both conditions) and then count reads over the peaks as if they were genes and run a differential analysis on the resulting count matrix.

The TSS as a range (num1, num2) would be called a promoter in other packages. The distance to TSS is the distance of the peak to the TSS that the range references. E.g. peaks are annotated as TSS if they are -1kb to +100bp from the actual TSS. So if an upstream peak has a distance to TSS less than 1,000 it will be annotated as TSS, or if it is downstream < 100 bp. If the peak is further upstream or downstream than those values, it will not be annotated as TSS.

[Daniel]

I see, thank you! So for example, a distance of 900 would be considered in the TSS?

[rfran010]

Should be.