ImmPrint immune gene set collection · v0.2.0-alpha

Reliable gene sets for immune signalling.

ImmPrint is a focused collection of immune signalling gene sets, built so you can score them with confidence. Each gene is specific to the process, kept apart from the ligand that triggers it, placed in a Receptor → Transducer → Effector hierarchy and tied to a paper for that gene in that pathway.

Browse the gene sets Use it in R

data = FileAttachment("data/immprint_0.2.0-alpha.csv").csv()

heroSet = "IMMPRINT_IFNG_SIGNALLING"
heroTiers = [
  { tier: "r", label: "Receptor", genes: ["IFNGR1", "IFNGR2"] },
  { tier: "t", label: "Transducer", genes: ["JAK1", "JAK2", "STAT1"] },
  { tier: "e", label: "Effector", genes: ["CXCL9", "CXCL10", "GBP1"] }
]
tierColour = ({ r: "var(--tier-receptor)", t: "var(--tier-transducer)", e: "var(--tier-effector)" })
heroPapers = ({
  IFNGR1: 'Aguet, Dembić & Merlin (1988), "Molecular cloning and expression of the human interferon-gamma receptor"',
  IFNGR2: 'Hemmi et al. (1994), "A novel member of the interferon receptor family complements functionality of the murine interferon gamma receptor in human cells"',
  JAK1: 'Müller et al. (1993), "The protein tyrosine kinase JAK1 complements defects in interferon-alpha/beta and -gamma signal transduction"',
  JAK2: 'Watling et al. (1993), "Complementation by the protein tyrosine kinase JAK2 of a mutant cell line defective in the interferon-gamma signal transduction pathway"',
  STAT1: 'Shuai et al. (1992), "Activation of transcription by IFN-gamma: tyrosine phosphorylation of a 91-kD DNA binding protein"',
  CXCL9: 'Liao et al. (1995), "Human Mig chemokine: biochemical and functional characterization"',
  CXCL10: 'Luster, Unkeless & Ravetch (1985), "γ-Interferon transcriptionally regulates an early-response gene containing homology to platelet proteins"',
  GBP1: 'Prakash et al. (2000), "Triphosphate structure of guanylate-binding protein 1 and implications for nucleotide binding and GTPase mechanism"'
})
mutable picked = "STAT1"

function pmidLine(gene) {
  const row = data.find(d => d.gene_set === heroSet && d.gene_symbol === gene);
  const ids = row ? String(row.pmids).split(";").map(s => s.trim()).filter(Boolean) : [];
  const col = heroTiers.find(c => c.genes.includes(gene));
  const colour = col ? tierColour[col.tier] : "var(--ink)";
  return html`<div class="pmid">
    <div class="pmid-line"><strong style=${"color:" + colour}>${gene}</strong> &nbsp;·&nbsp; ${ids.length
      ? ids.map((id, i) => html`${i ? ", " : ""}<a style=${"color:" + colour} href=${"https://pubmed.ncbi.nlm.nih.gov/" + id + "/"} target="_blank" rel="noopener">PMID ${id}</a>`)
      : "no recorded citation"}</div>
    <div class="paper-title">${heroPapers[gene] || ""}</div>
  </div>`;
}

{
  const el = html`<div class="trace">
    <div class="trace-head">one set, three tiers: <span class="setname">${heroSet}</span>*</div>
    <div class="tiers">
      ${heroTiers.map(col => html`<div class="tier ${col.tier}">
        <div class="tlabel">${col.label}</div>
        ${col.genes.map(g => html`<span class="gene" data-gene="${g}"
            style=${picked === g ? ("border-bottom-color:" + tierColour[col.tier]) : ""}>${g}</span>`)}
      </div>`)}
    </div>
    <div class="click-note">Click any gene to see its recorded citation</div>
    ${pmidLine(picked)}
    <div class="trace-foot">*These are a few example genes from this set rather than the full list.</div>
  </div>`;
  el.addEventListener("click", (e) => {
    const span = e.target.closest(".gene");
    if (span && span.dataset.gene) mutable picked = span.dataset.gene;
  });
  return el;
}

html`<div class="statstrip">
  <div class="stat"><div class="num">${new Set(data.map(d => d.gene_set)).size}</div><div class="cap">gene sets</div></div>
  <div class="stat"><div class="num">${data.length}</div><div class="cap">gene memberships</div></div>
  <div class="stat"><div class="num">${new Set(data.flatMap(d => d.cell_type.split(";").map(s => s.trim())).filter(s => s && s !== "pan")).size}</div><div class="cap">cell type variants</div></div>
  <div class="stat"><div class="num">100%</div><div class="cap">PMID-backed</div></div>
</div>`

What sets it apart

Broad collections (Hallmark, Reactome, GO) are built for coverage. ImmPrint is built to be a focused immune signalling resource you can score and defend.

Mechanistic specificity

Each set is built around a single process. Ubiquitous machinery (ribosomal, proteasomal, ubiquitin) and non-specific stress or activation families (MAPK, ATF, JUN, FOS, HSP) are excluded by policy.

Inducer and responder separation

A pathway’s own ligand is excluded from its signalling set but kept where it is a genuine output, so IL2 is absent from IMMPRINT_IL2_SIGNALLING, yet IFNG is present in IMMPRINT_TYPE1_RESPONSE. This matters most in single-cell data, where the sending and receiving cells are not the same cell.

A signalling hierarchy

Every gene is placed in a three-tier Receptor → Transducer → Effector hierarchy, so a set can be read, or subset, by where each gene sits in the cascade.

Citations you can follow

Every gene set membership carries a PubMed ID for a paper linking that gene to that named pathway. Where MSigDB corroborates a membership, that reference is recorded too.

Read the curation rationale →