H1-hESC cooler based on combination of 2 replicates R1-R2 is actually just R2, only hg38

[sergpolly]

Just want to bring it up here, to those who are using these coolers and/or dot-calls based on these before it would cause more confusion ...
@nkrietenstein @betulakgol @itsameercat @demerson368 let anyone relevant know this as well.

I'll reference @burakalver as well - and yes we need to redo dot-calls and such using 4DN-mapped coolers, and probably stick to 4DN processed data going forward - at least for 4DN-related purposes.

The title says it all: cooler-files that we generated for H1-hESC using hg38 assembly were not combined properly, such that the R1-R2 cooler is actually identical to R2 ...
just check the screenshot with cooler info output for different libraries/combinations mapped to hg19/hg38:

here is the screenshot from 4dn portal - everyhting is just fine there - the proper 2.5 billion reads for R1-R2 combination:

I'll of course address that issue, but only after the 4DN meeting ...

[sergpolly]

confirmation about the # of dot-calls:

sv49w@ghpcc06 ➜  2018-09-09-Snake-dotcalling-test for f in U54-ESC4DN-FA-DpnII-2017524-R*/combineddots/*.bedpe; do dirname $(dirname $f); echo "# of dots called:  " $(wc -l $f|cut -f1 -d' '); done
U54-ESC4DN-FA-DpnII-2017524-R1_hg19
# of dots called:   3599
U54-ESC4DN-FA-DpnII-2017524-R1_hg38
# of dots called:   3638
U54-ESC4DN-FA-DpnII-2017524-R1-R2_hg19
# of dots called:   5804
U54-ESC4DN-FA-DpnII-2017524-R1-R2__hg38
# of dots called:   3019
U54-ESC4DN-FA-DpnII-2017524-R2_hg19
# of dots called:   2940
U54-ESC4DN-FA-DpnII-2017524-R2__hg38
# of dots called:   3019

[sergpolly]

at the same time bedpe-s provided in this repo seems fine for H1-hESC - i.e. R1-R2 is indeed based on a combination of 2 replicates. And they are hg38 at the same time as well ...

so, stuff that we sent to dcic should be fine ...

That means that H1-hESC has been remapped at some point on our side and things went wrong when, combining R1-R2.

@hakanozadam , were you mapping these ? could you please, at least, confirm that H1-hESC-s were remapped a couple of times, to clear confusion a little bit ?

• Right now in the U54-Deep folder , R1-R2__hg38 is the wrong one (just R2 really).
• So I'm confused about how come we have combined H1-hESC data on hg38 in this 4dn_jawg repo ...