scripts/_hallmark_descriptions.py

"""Official MSigDB Hallmark (h.all) collection brief descriptions — all 50 gene sets.

Source: MSigDB Hallmark collection (Liberzon et al. 2015, Cell Systems), brief descriptions
as published by the Broad Institute. MSigDB is free under CC-BY 4.0 — attribute the Broad Institute.
Gene-set names verified against the official GMT:
    https://data.broadinstitute.org/gsea-msigdb/msigdb/release/2023.2.Hs/h.all.v2023.2.Hs.symbols.gmt

This module exists so the asset build (`scripts/build_embeddings.py`) is fully offline and
reproducible without R/msigdbr. `scripts/build_descriptions.R` documents the msigdbr regeneration path.
"""

HALLMARK_DESCRIPTIONS: dict[str, str] = {
    "HALLMARK_ADIPOGENESIS": "Genes up-regulated during adipocyte differentiation (adipogenesis).",
    "HALLMARK_ALLOGRAFT_REJECTION": "Genes up-regulated during transplant rejection.",
    "HALLMARK_ANDROGEN_RESPONSE": "Genes defining response to androgens.",
    "HALLMARK_ANGIOGENESIS": "Genes up-regulated during formation of blood vessels (angiogenesis).",
    "HALLMARK_APICAL_JUNCTION": "Genes encoding components of apical junction complex.",
    "HALLMARK_APICAL_SURFACE": "Genes encoding proteins over-represented on the apical surface of epithelial cells, e.g., important for cell polarity (apical area).",
    "HALLMARK_APOPTOSIS": "Genes mediating programmed cell death (apoptosis) by activation of caspases.",
    "HALLMARK_BILE_ACID_METABOLISM": "Genes involve in metabolism of bile acids and salts.",
    "HALLMARK_CHOLESTEROL_HOMEOSTASIS": "Genes involved in cholesterol homeostasis.",
    "HALLMARK_COAGULATION": "Genes encoding components of blood coagulation system; also up-regulated in platelets.",
    "HALLMARK_COMPLEMENT": "Genes encoding components of the complement system, which is part of the innate immune system.",
    "HALLMARK_DNA_REPAIR": "Genes involved in DNA repair.",
    "HALLMARK_E2F_TARGETS": "Genes encoding cell cycle related targets of E2F transcription factors.",
    "HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION": "Genes defining epithelial-mesenchymal transition, as in wound healing, fibrosis and metastasis.",
    "HALLMARK_ESTROGEN_RESPONSE_EARLY": "Genes defining early response to estrogen.",
    "HALLMARK_ESTROGEN_RESPONSE_LATE": "Genes defining late response to estrogen.",
    "HALLMARK_FATTY_ACID_METABOLISM": "Genes encoding proteins involved in metabolism of fatty acids.",
    "HALLMARK_G2M_CHECKPOINT": "Genes involved in the G2/M checkpoint, as in progression through the cell division cycle.",
    "HALLMARK_GLYCOLYSIS": "Genes encoding proteins involved in glycolysis and gluconeogenesis.",
    "HALLMARK_HEDGEHOG_SIGNALING": "Genes up-regulated by activation of hedgehog signaling.",
    "HALLMARK_HEME_METABOLISM": "Genes involved in metabolism of heme (a cofactor consisting of iron and porphyrin) and erythroblast differentiation.",
    "HALLMARK_HYPOXIA": "Genes up-regulated in response to low oxygen levels (hypoxia).",
    "HALLMARK_IL2_STAT5_SIGNALING": "Genes up-regulated by STAT5 in response to IL2 stimulation.",
    "HALLMARK_IL6_JAK_STAT3_SIGNALING": "Genes up-regulated by IL6 via STAT3, e.g., during acute phase response.",
    "HALLMARK_INFLAMMATORY_RESPONSE": "Genes defining inflammatory response.",
    "HALLMARK_INTERFERON_ALPHA_RESPONSE": "Genes up-regulated in response to alpha interferon proteins.",
    "HALLMARK_INTERFERON_GAMMA_RESPONSE": "Genes up-regulated in response to IFNG.",
    "HALLMARK_KRAS_SIGNALING_DN": "Genes down-regulated by KRAS activation.",
    "HALLMARK_KRAS_SIGNALING_UP": "Genes up-regulated by KRAS activation.",
    "HALLMARK_MITOTIC_SPINDLE": "Genes important for mitotic spindle assembly.",
    "HALLMARK_MTORC1_SIGNALING": "Genes up-regulated through activation of mTORC1 complex.",
    "HALLMARK_MYC_TARGETS_V1": "A subgroup of genes regulated by MYC - version 1 (v1).",
    "HALLMARK_MYC_TARGETS_V2": "A subgroup of genes regulated by MYC - version 2 (v2).",
    "HALLMARK_MYOGENESIS": "Genes involved in development of skeletal muscle (myogenesis).",
    "HALLMARK_NOTCH_SIGNALING": "Genes up-regulated by activation of Notch signaling.",
    "HALLMARK_OXIDATIVE_PHOSPHORYLATION": "Genes encoding proteins involved in oxidative phosphorylation.",
    "HALLMARK_P53_PATHWAY": "Genes involved in p53 pathways and networks.",
    "HALLMARK_PANCREAS_BETA_CELLS": "Genes specifically up-regulated in pancreatic beta cells.",
    "HALLMARK_PEROXISOME": "Genes encoding components of peroxisome, e.g., for the import, biogenesis, and metabolic functions of the organelle.",
    "HALLMARK_PI3K_AKT_MTOR_SIGNALING": "Genes up-regulated by activation of the PI3K/AKT/mTOR pathway.",
    "HALLMARK_PROTEIN_SECRETION": "Genes involved in protein secretion pathway.",
    "HALLMARK_REACTIVE_OXYGEN_SPECIES_PATHWAY": "Genes up-regulated by reactive oxygen species (ROS).",
    "HALLMARK_SPERMATOGENESIS": "Genes up-regulated during production of male gametes (sperm), as in spermatogenesis.",
    "HALLMARK_TGF_BETA_SIGNALING": "Genes up-regulated in response to TGFB1.",
    "HALLMARK_TNFA_SIGNALING_VIA_NFKB": "Genes regulated by NF-kB in response to TNF.",
    "HALLMARK_UNFOLDED_PROTEIN_RESPONSE": "Genes up-regulated during unfolded protein response, a cellular stress response related to the endoplasmic reticulum.",
    "HALLMARK_UV_RESPONSE_DN": "Genes down-regulated in response to ultraviolet (UV) radiation.",
    "HALLMARK_UV_RESPONSE_UP": "Genes up-regulated in response to ultraviolet (UV) radiation.",
    "HALLMARK_WNT_BETA_CATENIN_SIGNALING": "Genes up-regulated by activation of WNT signaling through accumulation of beta catenin CTNNB1.",
    "HALLMARK_XENOBIOTIC_METABOLISM": "Genes encoding proteins involved in processing of drugs and other xenobiotics.",
}

assert len(HALLMARK_DESCRIPTIONS) == 50, f"expected 50 hallmarks, got {len(HALLMARK_DESCRIPTIONS)}"