13d78f3b0149c4d9216fb78f28fbd9f91c088237 chmalee Tue Jun 13 15:50:10 2023 -0700 Remove sample level high repro region beds and replace with a single overlapped item bed, refs #31122 diff --git src/hg/makeDb/doc/hg38/problematic.txt src/hg/makeDb/doc/hg38/problematic.txt index 27dd8cf..7b22557 100644 --- src/hg/makeDb/doc/hg38/problematic.txt +++ src/hg/makeDb/doc/hg38/problematic.txt @@ -1,69 +1,75 @@ cd /hive/data/genomes/hg38/bed/problematic/ # note that the hg19 directory is /hive/data/genomes/hg19/bed/specialRegions/, the track name used to be different # add the ENCODE blacklist wget 'https://github.com/Boyle-Lab/Blacklist/blob/master/lists/hg38-blacklist.v2.bed.gz?raw=true' -O hg38-blacklist.v2.bed.gz gunzip hg38-blacklist.v2.bed.gz mkdir bb bedSort hg38-blacklist.v2.bed hg38-blacklist.v2.bed bedToBigBed hg38-blacklist.v2.bed /hive/data/genomes/hg38/chrom.sizes bb/encBlacklist.bb -type=bed4 -tab # bring over the manual comments track, mostly manual work cd ../manualComments ln -s ../../../../hg19/bed/specialRegions/manualComments mv manualComments hg19Comments cat hg19Comments/manual.bed | cut -f1-4 > hg19.bed ~/bin/x86_64/liftOver -tab hg19.bed /hive/data/genomes/hg19/bed/liftOver/hg19ToHg38.over.chain.gz manual.lift.bed unmapped.lift.bed # -- manual step! -- fixed up manual.lift.bed manually adding PAR regions cat hg19Comments/manual.bed | cut -f4- | sort > annots.tsv tabJoin annots.tsv 0 manual.lift.bed 3 > manual.bed # -- manual step -- fixed up manual.bed hgsql hg38 -e 'select * from chromInfo' | grep chrUn | cut -f-2 |tawk '{$3=$2; $2=0; $4="unplaced"; $5="ChrUn contains clone contigs that cannot be confidently placed on a specific chromosome."; $6="none"; print}' > chrUn.bed # fixes hgsql hg38 -e 'select * from chromInfo' | grep _fix | cut -f-2 |tawk '{$3=$2; $2=0; $4="fix"; $5="The chr_fix chromosomes, such as chr1_KN538361v1_fix, are fix patches currently available for the hg19 and hg38 assemblies that represent changes to the existing sequence. These are generally error corrections (such as base changes, component replacements/updates, switch point updates or tiling path changes) or assembly improvements (such as extension of sequence into gaps). These fix patch scaffold sequences are given chromosome context through alignments to the corresponding chromosome regions, in the \"Mapping and Sequencing > GRC Patch Release\" track. See also <a href=\"https://genome.ucsc.edu/FAQ/FAQdownloads#downloadFix\" target=_blank>the FAQ</a> for more information."; $6="none"; print}' > chrFix.bed # alts hgsql hg38 -e 'select * from chromInfo' | grep alt | cut -f-2 |tawk '{$3=$2; $2=0; $4="alt"; $5="The chr_alt chromosomes, such as chr5_KI270794v1_alt, are alternative sequences that differ from the reference genome currently available for a few assemblies including danRer11, hg19, and hg38. These are regions of the genome that exhibit sufficient variability to prevent adequate representation by a single sequence. UCSC labels these haplotype sequences by appending "_alt" to their names. These alternative loci scaffolds (such as KI270794.1 in the hg38 assembly, referenced as chr5_KI270794v1_alt in the browser), are mapped to the genome and provide suppemental genomic information on these variable locations. To find the regions these alternate sequences correspond to in the genome you may use the Alt Haplotypes track if one is available. See also <a href=\"https://genome.ucsc.edu/FAQ/FAQdownloads#downloadAlt\" target=_blank>the FAQ</a> for more information."; $6="none"; print}' > chrAlt.bed # various other gene clusters bigBedToBed /gbdb/hg38/ncbiRefSeq/ncbiRefSeqOther.bb stdout | less | grep -v pseudo | grep -v "T cell" -i | grep -v tRNA | grep -v immuno | grep -v constant | grep -v miR | grep -v UGT1A | grep -v PCDHA | grep -v PCDHB | cut -f1-4,18 | tawk '{$5=$5" HGNC ID:"$4" This is a cluster of many very similar genes based on the <b>Genes and Gene Predictions > NCBI RefSeq > RefSeq Other</b> Track"; $4="cluster"; print}' > chrClusters.bed # put everything together and make trix files cat manual.bed chr*.bed | sort -k1,1 -k2,2n | tawk '{desc=$5; desc2=$6; $5="0"; $6="+"; $7=$2; $8=$3; $9="0,0,0"; $10=desc; $11=desc2; print}' > all.bed bedToBigBed all.bed /hive/data/genomes/hg38/chrom.sizes comments.bb -tab -as=manual.as -type=bed9+ -extraIndex=name cut -f4,10 all.bed > notes.txt ixIxx notes.txt notes.ix notes.ixx # Tue Apr 11 02:12:18 PDT 2023 # add the GRC exclude list, from MarkD cat ~markd/public_html/browser/grc-bad/GCA_000001405.15_GRCh38_GRC_exclusions.bed | grep -v description > grcExclusions.bed bedSort grcExclusions.bed grcExclusions.bed bedToBigBed grcExclusions.bed ../../chrom.sizes bb/grcExclusions.bb -tab -type=bed4 # Add highly reproducible regions (#31122): # https://zenodo.org/record/5275189#.ZEhKRezMI-S cd /hive/data/genomes/hg38/bed/problematic/ mkdir highRepro # The zenodo link has a bunch of links with "?download=1" on the url, copy and paste them # into a text file to remove that and then download them vim urls.txt wget --quiet --input-file=urls.txt # all of the beds are just bed3's head *.bed ==> CQ-56.bed <== chr1 12857 12904 chr1 12932 13028 chr1 13129 13367 chr1 13520 13633 ... # so turn those into bigBeds and link everything to /gbdb for f in $(ls *.bed); do base=$(basename -s .bed $f); bedSort ${f} ${f}.sorted; bedToBigBed ${f}.sorted../../../chrom.sizes ${base}.bb; done # for some reason tabix doesn't like to be run in a for loop from ls, needs find for f in $(find . -name "*.vcf.gz"); do tabix --verbosity 3 -p vcf "${f}"; done mkdir -p /gbdb/hg38/problematic/highRepro ln -s `pwd`/*.bb /gbdb/hg38/problematic/highRepro/ ln -s `pwd`/*.vcf.gz* /gbdb/hg38/problematic/highRepro/ + +# turns the beds into one single bed with the overlapped regions: +for f in $(ls *.bed); do echo $f; n=${f/.bed/}; tawk -v name=${n} '{print $0,name}' $f > ${n}.bed4; done +cat *.bed4 | sort -k1,1 -k2,2n > highRepro.allRegions +bedOverlapMerge highRepro.allRegions > highRepro.merged +bedToBigBed -type=bed3+ -tab -as=highRepro.as highRepro.merged ../../../chrom.sizes highRepro.bb