known unmappable or poorly sequenced regions in hg38

Question

[Paul]

Dear all,

is there any similar file to this, which known unmappable or poorly sequenced regions for hg38? Or any way to download it from ucsc table browser?

Thank you for any link or comments.

Answer 1

[igor]

The blacklist files are available from ENCODE directly (there is a drop-down menu to select the proper genome build): https://www.encodeproject.org/annotations/ENCSR636HFF/

I don't know if using LiftOver for these regions is a good idea, since their coordinates are uncertain in the first place.

[Paul]

Thank you Igor. Did you validate this regions? I can see, that for build GRCh38 is there only 38 regions. For Hg19 it is 411 regions. Also when I check this version of BED, there are 1649 regions.

[igor]

There are actually two different BED files there. There are DAC and Duke lists. More info: Difficult to Sequence Regions for WGS

Answer 2

[harold.smith.tarheel]

You can use a combination of blacklisted regions from hg37 with the UCSC liftOver utility (or maybe someone else can recommend a blacklist for hg38).

[Paul]

Thank you so much. I was thinking to use "wgEncodeDukeMapabilityRegionsExcludable.bed.gz" from here and then use UCSC liftOver, but is this good procedure? Isnt those regions "out of date" for HG38?