ecc2331000637cbc2523653642d926935b5b83fc chmalee Sat Jun 13 00:42:03 2026 -0700 gnomAD v4.1.1 bigBed variant track for hg38, refs #37351 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> diff --git src/hg/makeDb/doc/hg38/gnomad.txt src/hg/makeDb/doc/hg38/gnomad.txt index f5c7829c9d4..5fadce03a90 100644 --- src/hg/makeDb/doc/hg38/gnomad.txt +++ src/hg/makeDb/doc/hg38/gnomad.txt @@ -737,15 +737,56 @@ # Creates 11 bigWig files: # gnomad.coverage.mean.bw, gnomad.coverage.median.bw, # gnomad.coverage.over_{1,5,10,15,20,25,30,50,100}.bw # Note: The script handles chromosome ordering (gnomAD uses chr1-22,X,Y,M order, # not the chrom.sizes order) and skips duplicate positions in the input data. # Processed 170,202,927 lines, skipped 2 duplicate positions. # Create symlinks in /gbdb mkdir -p /gbdb/hg38/gnomAD/coverage/v4-exome cd /gbdb/hg38/gnomAD/coverage/v4-exome for f in /hive/data/genomes/hg38/bed/gnomad/coverage/v4-exome/gnomad.coverage.*.bw; do ln -sf "$f" .; done # Add track to trackDb (gnomadCoverage.ra) and rebuild cd ~/kent/src/hg/makeDb/trackDb make DBS=hg38 + +############################################################################## +# gnomAD v4.1.1 Update to bigBed - June 5, 2026 - Claude (chmalee) +############################################################################## +# v4.1.1 is a minor release. The per-chromosome sites VCFs were mirrored from +# the public bucket with the download script (run once per dataset): +# src/hg/makeDb/scripts/gnomadV4.1.1/downloadGnomadV4.1.1.sh exomes +# src/hg/makeDb/scripts/gnomadV4.1.1/downloadGnomadV4.1.1.sh genomes +# Data landed in /hive/data/outside/gnomAD.4.1.1/{exomes,genomes}/. + +# Confirmed v4.1.1 is field-compatible with v4.1 before reusing the v4.1 +# recipe: every INFO field in the v4.1 build lists is present in v4.1.1, and +# the VEP Format string is identical for both datasets: +# for f in $(cat v4.1.1.vcfToBed.exomes.fieldList); do \ +# bcftools view -h .../gnomad.exomes.v4.1.1.sites.chr1.vcf.bgz \ +# | grep -Eo "^##INFO=<ID=[^,]+" | cut -d= -f3 | grep -qx "$f" \ +# || echo "MISSING: $f"; done +# (no output for either dataset). Because the layout matches v4.1, the build +# reuses the v4.1 vcfToBed field lists and the gnomadVcfBedToBigBed parser keys +# -v v4.1_exomes / -v v4.1_genomes unchanged. + +# Note: the in-tree src/hg/makeDb/gnomad/gnomadVcfBedToBigBed was stale (only +# supported up to v3.1.1); the v4.1 parsing work had only ever lived on the +# build host. The working version is now committed to the tree. + +# The whole VCF-to-bigBed pipeline (vcfToBed, gnomadVcfBedToBigBed, bgzip of the +# details file, bedJoinTabOffset, bedToBigBed, and the /gbdb symlinks) is driven +# by one resumable script, run once per dataset (long job, run in screen): +# src/hg/makeDb/scripts/gnomadV4.1.1/buildGnomadV4.1.1BigBed.sh exomes +# src/hg/makeDb/scripts/gnomadV4.1.1/buildGnomadV4.1.1BigBed.sh genomes +# Workdir /hive/data/inside/gnomAD/v4/v4.1.1, progress in build.{exomes,genomes}.log. +# Scripts, field lists and autosql: +# https://github.com/ucscGenomeBrowser/kent/tree/master/src/hg/makeDb/scripts/gnomadV4.1.1 + +# trackDb: the bigBed composite gnomadVariantsV4.1.1 (children +# gnomadExomesVariantsV4_1_1, gnomadGenomesVariantsV4_1_1) was added to +# human/hg38/gnomad.ra alongside v4.1, cloned from the v4.1 stanza. The +# experimental vcfTabix track in gnomadV4.1.1.ra (alpha only) was renamed to +# gnomadVariantsV4_1_1Vcf so the bigBed track could take the canonical name; its +# per-chrom lookup tables are built by loadGnomadV4.1.1ChromTables.sh. +