"""Data registry for GVL tutorial datasets."""
from pathlib import Path
from typing import Literal
import pooch
N_RETRIES = 3
[docs]
def fetch(name: Literal["geuvadis_ebi", "1kgp"]) -> dict[str, Path]:
"""Download and cache data for constructing/opening a GVL dataset. Files are cached in the
user's home directory under :code:`~/.cache/genvarloader`.
Parameters
----------
name
The name of the dataset to fetch. Can be one of:
- "geuvadis_ebi": Geuvadis data for the original analyses by Lappalainen et al. 2013. Phased, normalized, and split into biallelic variants.
- "1kgp": 1000 Genomes Project, all 3,202 individuals. Phased, normalized, and split into biallelic variants.
Returns
-------
A dictionary of paths to the fetched data.
"""
if name == "geuvadis_ebi":
return _geuvadis_ebi()
elif name == "1kgp":
return _1kgp()
raise ValueError(f"Unknown dataset: {name}")
def _geuvadis_ebi():
huang = pooch.Pooch(
pooch.os_cache("genvarloader/geuvadis_ebi"),
"https://github.com/ni-lab/personalized-expression-benchmark/raw/refs/heads/main/",
registry={
"data/gene_list.csv": "md5:ed1b207b14313dbddef520413cbf9e40",
"consensus/samples.txt": "md5:0ba048e6dbb39dc6a6a9835e4528621f",
},
retry_if_failed=N_RETRIES,
)
basenji = pooch.Pooch(
pooch.os_cache("genvarloader/geuvadis_ebi"),
"https://github.com/calico/basenji/raw/refs/heads/master/manuscripts/cross2020/",
registry={"targets_human.txt": "md5:61c7fe3aa9da2f309124830ddb282ce3"},
retry_if_failed=N_RETRIES,
)
zenodo = pooch.Pooch(
pooch.os_cache("genvarloader/geuvadis_ebi"),
"https://zenodo.org/records/15596289/files/",
registry={
"GD462.GeneQuantRPKM.50FN.samplename.resk10.txt.gz": "md5:500bffeed8e0f770c157e0189e9e50ae",
"geuvadis.pgen": "md5:2815f8e20f8a0acbdead2b0c0d6a00dc",
"geuvadis.psam": "md5:3c33d631dceccb43222341485b8658f8",
"geuvadis.pvar.zst": "md5:51aad81267254695191d6284352b0523",
},
retry_if_failed=N_RETRIES,
)
genes = Path(huang.fetch("data/gene_list.csv", progressbar=True))
samples = Path(huang.fetch("consensus/samples.txt", progressbar=True))
basenji2_targets = Path(basenji.fetch("targets_human.txt", progressbar=True))
expr = Path(
zenodo.fetch(
"GD462.GeneQuantRPKM.50FN.samplename.resk10.txt.gz", progressbar=True
)
)
pgen = Path(zenodo.fetch("geuvadis.pgen", progressbar=True))
_ = zenodo.fetch("geuvadis.psam", progressbar=True)
_ = zenodo.fetch("geuvadis.pvar.zst", progressbar=True)
return {
"genes": genes,
"expr": expr,
"pgen": pgen,
"basenji2_targets": basenji2_targets,
"samples": samples,
}
def _1kgp():
zenodo = pooch.Pooch(
pooch.os_cache("genvarloader/1kgp"),
"https://zenodo.org/records/15596480/files/",
registry={
"1kGP.snp_indel.split_multiallelics.pgen.xz": "md5:3e07b63b48e5a205a2b2ae5022e01871",
"1kGP.snp_indel.split_multiallelics.psam": "md5:cb8ea444a41cceb27483df26bb76ad1b",
"1kGP.snp_indel.split_multiallelics.pvar.zst": "md5:cb00b271504713dbbdbbf17ee0b1825e",
},
retry_if_failed=N_RETRIES,
)
pgen = Path(
zenodo.fetch(
"1kGP.snp_indel.split_multiallelics.pgen.xz",
pooch.Decompress(),
progressbar=True,
)
)
_ = zenodo.fetch("1kGP.snp_indel.split_multiallelics.psam", progressbar=True)
_ = zenodo.fetch("1kGP.snp_indel.split_multiallelics.pvar.zst", progressbar=True)
return {"pgen": pgen}
