Difference between "peaks" and "hotspots" in Output.type of ENCODE DNase-seq data sets

Question

[panyard]

I'm having trouble understanding the difference between two types of "Output.type" for DNase-seq data sets from ENCODE. About half of them are labeled as "peaks" and the other half as "hotspots". Looking through the ENCODE website (https://www.encodeproject.org/data-standards/dnase-seq/) (https://www.encodeproject.org/data-standards/terms/#enrichment), I can see that Hotspot is the name of a program that is used in the analysis pipeline, but I don't see any description of the difference between hotspots and peaks as output types. The only difference I did see was that all hotspots were for bed broadPeak files and all peaks were for bed narrowPeak files, so perhaps this is just inconsistency in the naming convention?

Does anyone know if there's any real difference between these Output.type designations? Thanks

Answer 1

[Alex Reynolds]

Hotspots are areas of enriched DNaseI cleavage called on full or subsampled read sets. Peaks are 150nt regions that overlap or are contained within hotspots, which are areas of very high density relative to background. These 150nt regions are also called DHSs (DNase-hypersensitive sites). This is for first-generation hotspot data. The hotspot2 method calls variable-sized DHSs.

• https://github.com/rthurman/hotspot
• https://github.com/Altius/hotspot2

Answer 2

[ben.kunfang]

The "hotspots" files are DNaseI sensitivezones, provided in two file formats (bed broadPeak, and bigBed broadPeak for visualization). "hotspots" are genomic regions enriched for cleavage by DNaseI, corrected to remove false positives. Narrow regions of local maxima from these "hotspots" are called as "peaks", provided in the "peaks" file. The "peaks" are the DNaseI hypersensitive sites (DHS) that you're looking for, and are also provided in two file formats (bed narrowPeak and bigBed narrowPeak for visualization).