hitlist 0.8.0

Curated mass spectrometry evidence for MHC ligand data from IEDB and CEDAR

hitlist

Curated mass spectrometry evidence for MHC ligand data from IEDB and CEDAR.

hitlist scans IEDB and CEDAR MHC ligand exports, classifies each observation by biological source context (cancer tissue, healthy tissue, cell line, tumor-adjacent, etc.), maps peptides to source proteins with flanking sequences, and produces data quality reports. PMID-level curation overrides and tissue classifications are stored as YAML data files, not hardcoded Python.

Install

pip install hitlist

Quick start

# Register your IEDB/CEDAR downloads
hitlist data register iedb /path/to/mhc_ligand_full.csv
hitlist data register cedar /path/to/cedar-mhc-ligand-full.csv

# Generate a data quality report
hitlist report
hitlist report --class I --output report.txt

from hitlist.scanner import scan
from hitlist.curation import classify_ms_row, is_cancer_specific
from hitlist.aggregate import aggregate_per_peptide

# Scan for specific peptides
hits = scan(
    peptides={"SLYNTVATL", "GILGFVFTL"},
    iedb_path="mhc_ligand_full.csv",
    mhc_class="I",
)

# Or profile the entire dataset
full = scan(peptides=None, iedb_path="mhc_ligand_full.csv")

# Per-peptide summary with cancer-specific classification
summary = aggregate_per_peptide(hits)

Source classification

Every IEDB/CEDAR mass spec observation is classified into one of these categories:

Category	Flag	Rule
Cancer	src_cancer	Tumor tissue, cancer patient biofluids, or non-EBV cell lines
Adjacent to tumor	src_adjacent_to_tumor	Surgically resected "normal" tissue (per-PMID override)
Activated APC	src_activated_apc	Monocyte-derived DCs/macrophages with pharmacological activation
Healthy somatic	src_healthy_tissue	Direct ex vivo, healthy donor, non-reproductive, non-thymic
Healthy thymus	src_healthy_thymus	Direct ex vivo thymus (expected for CTAs, AIRE-mediated)
Healthy reproductive	src_healthy_reproductive	Direct ex vivo testis, ovary, etc. (expected for CTAs)
EBV-LCL	src_ebv_lcl	EBV-transformed B-cell lines
Cell line	src_cell_line	Any cultured cell line

Key rule: all non-EBV cell lines are classified as cancer-derived, even when IEDB labels them "No immunization". This catches HeLa, THP-1, A549, and other cancer lines used in non-cancer studies.

Cancer-specific = found in cancer AND NOT found in healthy somatic tissue. Thymus, reproductive tissue, adjacent tissue, EBV-LCLs, and activated APCs do NOT disqualify.

PMID curation overrides

Expert per-study overrides are stored in hitlist/data/pmid_overrides.yaml:

- pmid: 29557506
  label: "Neidert 2018 — Tübingen/Zurich biobank"
  override: healthy
  tissue_overrides:
    Blood: healthy           # blood bank donors
    Bone Marrow: healthy     # hip arthroplasty
    Colon: adjacent          # Visceral Surgery dept → likely CRC margin
    Kidney: adjacent         # Urology → likely nephrectomy
    Liver: adjacent          # likely HCC/met adjacent

Tissue-level overrides take priority over study-level overrides.

Proteome mapping

Map peptides to source proteins with flanking context:

from hitlist.proteome import ProteomeIndex

# Human proteome (from pyensembl)
idx = ProteomeIndex.from_ensembl(release=112)

# Or combined human + viral
idx = ProteomeIndex.from_ensembl_plus_fastas(
    fasta_paths=["hpv16.fasta", "ebv.fasta"],
)

# Map peptides with 5-residue flanks
df = idx.map_peptides(["SLLMWITQC"], flank=5)
# → protein_id, gene_name, gene_id, position, n_flank, c_flank, n_sources, unique_n_flank, unique_c_flank

Per-sample peptidome context

The full peptidome context for each sample is critical for interpreting whether a peptide's presence is meaningful:

from hitlist.scanner import scan
from hitlist.samples import sample_peptidomes, overlay_targets

# Full scan (ALL peptides, not just targets)
full = scan(peptides=None, iedb_path="mhc_ligand_full.csv", mhc_class="I")

# Per-sample stats
samples = sample_peptidomes(full)

# Overlay CTA peptides for context fractions
# "1 CTA out of 762 peptides = 0.13% = stochastic noise"
context = overlay_targets(full, target_peptides=my_cta_set, label="cta")

Data management

hitlist data available          # show all 14 known datasets
hitlist data fetch hpv16        # auto-download viral proteome from UniProt
hitlist data register iedb /path/to/file  # register manual download
hitlist data list               # show registered datasets with size/date
hitlist data info iedb          # detailed JSON metadata
hitlist data path iedb          # resolve to file path
hitlist data refresh hpv16      # re-download
hitlist data remove iedb        # unregister

Storage: ~/.hitlist/ (override with HITLIST_DATA_DIR env var).

Development

./develop.sh    # install in dev mode
./format.sh     # ruff format
./lint.sh       # ruff check + format check
./test.sh       # pytest with coverage
./deploy.sh     # lint + test + build + upload to PyPI