mecfs_bio.assets.reference_data.genetic_map.genetic_map_hg38
Reference asset: the Eagle hg38 (GRCh38) genetic map and its parsed parquet.
Provides the per-position recombination rate (cM/Mb) for recombination tracks of GRCh38 locus plots, such as the coloc locus stack plots.
The raw map (genetic_map_hg38_withX.txt.gz) is space-separated with columns "chr position COMBINED_rate(cM/Mb) Genetic_Map(cM)"; the header names carry characters we would only rename away, so the download's read-spec renames the four columns positionally. Parsing is a plain read-plus-sort, so it is expressed as a PipeDataFrameTask (read via the download's read-spec, sort by CHR/POS, write parquet) rather than a bespoke task.
Attributes:
GENETIC_MAP_HG38
module-attribute
GENETIC_MAP_HG38 = PipeDataFrameTask.create(
source_task=GENETIC_MAP_HG38_RAW,
asset_id="genetic_map_hg38",
out_format=ParquetOutFormat(),
pipes=[SortPipe(by=[GWASLAB_CHROM_COL, GMAP_POS_COL])],
backend="polars",
)
GENETIC_MAP_HG38_RAW
module-attribute
GENETIC_MAP_HG38_RAW = DownloadFileTask(
meta=ReferenceFileMeta(
group="genetic_map",
sub_group="hg38",
sub_folder=PurePath("raw"),
id=AssetId("genetic_map_hg38_eagle"),
extension=".txt.gz",
read_spec=DataFrameReadSpec(
DataFrameTextFormat(
separator=" ",
has_header=True,
column_names=_OUTPUT_COLUMNS,
schema_overrides={
GWASLAB_CHROM_COL: pl.Int64(),
GMAP_POS_COL: pl.Int64(),
GMAP_RATE_COL: pl.Float64(),
GMAP_CM_COL: pl.Float64(),
},
)
),
),
url="https://storage.googleapis.com/broad-alkesgroup-public/Eagle/downloads/tables/genetic_map_hg38_withX.txt.gz",
md5_hash="70f0be81e1ad1314ccc88e48091bcba1",
)