63653eb7f4fcccfedec87fe115ed0a954012c3ad max Thu Apr 30 05:46:04 2026 -0700 Adding new "m6A-Atlas v2 sites" track under rnaMod superTrack on hg38: 427760 high-confidence base-resolution m6A sites compiled by m6A-Atlas v2 (Liang 2024) from 12 detection technologies across 24 cell lines / tissues. Filters on region, technique, cell line, biotype, #techniques, #cell lines. Track lives in its own m6aAtlas.ra file, included alpha-only via trackDb.ra, refs #36613 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> diff --git src/hg/makeDb/doc/hg38/rnaMod.txt src/hg/makeDb/doc/hg38/rnaMod.txt index 65ba5cdd5c1..52db49d9ad7 100644 --- src/hg/makeDb/doc/hg38/rnaMod.txt +++ src/hg/makeDb/doc/hg38/rnaMod.txt @@ -1,21 +1,22 @@ # 2026-04-29 Claude / Max -- DRACH (m6A consensus) motif track on hg38 # The "rnaMod" supertrack groups RNA-modification annotations on hg38. # Currently it contains: # - fetalXiao2019 (built earlier, not documented here) # - drach (built below) +# - m6aAtlas (built below; high-confidence m6A sites from m6A-Atlas v2) ############################################################################### # DRACH motif sites in MANE Select transcripts (Claude, 2026-04-29) ############################################################################### # All DRACH 5-mers (D=A/G/T, R=A/G, A, C, H=A/C/T) are extracted from MANE # Select v1.5 mature transcript fasta and lifted to hg38 genome coordinates # via pslMap. The result is a bigBed 12+5 with empty `name` and gene/motif # metadata in extra columns. mkdir -p /hive/data/genomes/hg38/bed/rnaMod/drach cd /hive/data/genomes/hg38/bed/rnaMod/drach # the build is fully driven by one shell script; see the directory for helpers ~/kent/src/hg/makeDb/scripts/rnaMod/makeDrach.sh @@ -48,15 +49,48 @@ # Motifs mapped to genome (pslToBed): 1381109 # Final features in drach.bb: 1381109 # Multi-block (splice junction): 13174 (~0.95%) # Symlink into /gbdb (one-time): # ln -sf /hive/data/genomes/hg38/bed/rnaMod/drach/drach.bb \ # /gbdb/hg38/rnaMod/drach.bb # trackDb stanza was added to ~/kent/src/hg/makeDb/trackDb/human/hg38/rnaMod.ra # under the existing `track rnaMod` supertrack. Filters were added for `motif` # and `region`, with no defaults. # Off-by-one verification (single-block + and -, multi-block + spanning a # splice junction) was performed against /hive/data/genomes/hg38/hg38.2bit and # all three sample motifs round-tripped correctly. + +############################################################################### +# m6A-Atlas v2 high-confidence sites on hg38 (Claude, 2026-04-29) +# PMID 37587690 Liang Z et al, NAR 2024;52(D1):D194-D202 +############################################################################### + +# m6A-Atlas v2.0 distributes a "high-confidence" base-resolution m6A site +# list per species. We use the human (hg38) "Cell line / Technique" file, +# which already aggregates per-site evidence across GEO studies. + +mkdir -p /hive/data/genomes/hg38/bed/rnaMod/m6aAtlas +cd /hive/data/genomes/hg38/bed/rnaMod/m6aAtlas +aria2c -x10 http://rnamd.org/m6a/download/high/hg38/hg38_CL_Tech.txt.gz + +# Convert to bed9+15 (1-based -> 0-based half-open, region color, score from +# nTechniques/nCellLines, m6A-Atlas detail URL). +python3 ~/kent/src/hg/makeDb/scripts/rnaMod/m6aAtlasToBed.py \ + hg38_CL_Tech.txt.gz m6aAtlas.bed 2> m6aAtlas.convert.log + +sort -k1,1 -k2,2n m6aAtlas.bed > m6aAtlas.sorted.bed +bedToBigBed -tab -type=bed9+15 \ + -as=$HOME/kent/src/hg/makeDb/scripts/rnaMod/m6aAtlas.as \ + -extraIndex=name,geneName,ensemblId \ + m6aAtlas.sorted.bed /hive/data/genomes/hg38/chrom.sizes m6aAtlas.bb + +# Driver and helpers: +# ~/kent/src/hg/makeDb/scripts/rnaMod/m6aAtlasToBed.py +# ~/kent/src/hg/makeDb/scripts/rnaMod/m6aAtlas.as + +# Build summary (recorded after the run): +# Source rows (data lines in hg38_CL_Tech.txt.gz): 427760 +# Output features in m6aAtlas.bb: 427760 +# Skipped (unknown chrom or out-of-bounds): 0