HOMER HiC interactions

Question

[srhic]

Hello,

I am using HOMER to analyze my HiC data and looking for significant interactions. I used HOMER to identify significant interactions in my sample which has two replicates and got ~100,000 significant interactions in each replicate (using the command analyzeHiC interactions).

However, since HOMER does not have the option to compare interactions between conditions with replicates, I also performed the same analysis after merging the two replicates (by merging their tag directories). I got ~ 200,000 significant interactions using the same parameters on merged replicates. Now I am confused because I was not expecting merging the replicates to double the number of significant interactions. My understanding is that merging the two replicates should produce results closer to the average of the two?

Does anyone have any comments on what merging the tag directories of two replicates actually does and if this is the expected behavior?

Thanks

[Gjain]

Are you sure you got different interactions and not duplicates?

[srhic]

They dont seem to be duplicates though it is hard to check that systematically.

[Gjain]

Maybe you can try with bedtools to check if they are overlapping or not?

[srhic]

Thanks, I found this paper which shows that the number of interactions detected by HOMER is directly related to the number of reads in a sample.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5493985/

I assume this is why I am getting more interactions in merged samples but I am still unsure what would be the best way to go about this analysis in this case.

[Gjain]

Thanks for the paper and the link. I was not aware of that so that is good to know.

I think you can try the standard HiC pipelines and a nice tutorial can be found here: https://hms-dbmi.github.io/hic-data-analysis-bootcamp/#1

I hope this helps.