Skip to content

mecfs_bio.assets.reference_data.hgnc.hgnc_complete_set

The HGNC complete set: one row per approved human gene, with its current symbol, name, locus type, Ensembl gene id and the UniProt accessions of its protein products. Pinned to a dated monthly snapshot.

HGNC_UNIPROT_GENES normalizes it into a UniProt-to-gene table: one row per (UniProt accession, gene), with the gene's symbol, name and Ensembl gene id. Genes without a UniProt accession are dropped, and so are readthrough genes (a transcript spanning two adjacent genes, such as PTGES3L-AARSD1, which HGNC lists under the accession of one of the genes it spans). An accession encoded by duplicated genes (for example P59665, by DEFA1 and DEFA1B) has one row per gene.

Source: https://www.genenames.org/download/archive/

Attributes:

HGNC_COMPLETE_SET module-attribute 

HGNC_COMPLETE_SET = DownloadFileTask(
    meta=ReferenceFileMeta(
        group="nomeclature",
        sub_group="hgnc",
        sub_folder=PurePath("raw"),
        filename="hgnc_complete_set_2026-10-06",
        extension=".tsv",
        id=AssetId("hgnc_complete_set"),
        read_spec=DataFrameReadSpec(
            DataFrameTextFormat(
                separator="\t",
                schema_overrides={
                    col: pl.String()
                    for col in (
                        HGNC_SYMBOL_COL,
                        HGNC_NAME_COL,
                        HGNC_LOCUS_TYPE_COL,
                        HGNC_ENSEMBL_ID_COL,
                        HGNC_UNIPROT_IDS_COL,
                    )
                },
            )
        ),
    ),
    url="https://storage.googleapis.com/public-download-files/hgnc/archive/archive/monthly/tsv/hgnc_complete_set_2026-10-06.tsv",
    md5_hash="af43fd56a64ad57c0fcc72d6ef65e516",
)

HGNC_UNIPROT_GENES module-attribute 

HGNC_UNIPROT_GENES = PipeDataFrameTask.create(
    source_task=HGNC_COMPLETE_SET,
    asset_id="hgnc_uniprot_genes",
    out_format=ParquetOutFormat(),
    backend="polars",
    pipes=[
        DropRowsByValue(
            target_column=HGNC_LOCUS_TYPE_COL,
            values_to_drop=[HGNC_READTHROUGH_LOCUS_TYPE],
        ),
        SelectColPipe(
            cols_to_select=[
                HGNC_SYMBOL_COL,
                HGNC_NAME_COL,
                HGNC_ENSEMBL_ID_COL,
                HGNC_UNIPROT_IDS_COL,
            ]
        ),
        SplitColPipe(
            col_to_split=HGNC_UNIPROT_IDS_COL,
            split_by=HGNC_LIST_SEPARATOR,
            new_col_name=HGNC_UNIPROT_ID_COL,
        ),
        UNNestPipe(col_to_unnest=HGNC_UNIPROT_ID_COL),
        SelectColPipe(
            cols_to_select=[
                HGNC_UNIPROT_ID_COL,
                HGNC_SYMBOL_COL,
                HGNC_NAME_COL,
                HGNC_ENSEMBL_ID_COL,
            ]
        ),
        DropNullsPipe(subset=[HGNC_UNIPROT_ID_COL]),
        UniquePipe(
            by=[HGNC_UNIPROT_ID_COL, HGNC_SYMBOL_COL],
            keep="any",
        ),
    ],
)