Skip to content

lnp — lipid nanoparticle composition recommender

A rule-based shortlist of published or ML-discovered ionizable-lipid formulations for a given (target tissue × cargo × therapeutic intent).

Usage

# Cancer vaccine, lung delivery, saRNA cargo
mrnavax lnp --target lung --cargo saRNA --intent "cancer vaccine"

# Hepatic gene editing with Cas9 mRNA
mrnavax lnp --target liver --cargo Cas9 --intent "gene editing"

# Intratumoral injection (TRAIL-mRNA example)
mrnavax lnp --target tumor --cargo mRNA --intent "cancer vaccine"

Output schema

{
  "cargo": "sarna",
  "target": "lung",
  "intent": "cancer vaccine",
  "shortlist": [
    {
      "name": "FO-32 (pulmonary, ML-designed)",
      "ionizable_lipid": "FO-32",
      "helper_lipid": "DOPE",
      "cholesterol_pct": 24.0,
      "peg_lipid": "DMG-PEG2000",
      "peg_mol_pct": 1.0,
      "ionizable_mol_pct": 60.0,
      "helper_mol_pct": 10.0,
      "n_p_ratio": 8.0,
      "source": "Witten et al. Nat Biotech 2025",
      "notes": "Top hit from ML-guided screening (>1.6M candidates); ferret-lung delivery."
    }
  ],
  "notes": [
    "Pulmonary delivery benefits from ML-discovered biodegradable lipids (Witten 2025).",
    "saRNA prefers higher cholesterol and lower PEG for replicon stability."
  ]
}

Curated preset table

Preset Source Best for
SM-102 Moderna clinical vaccines (liver/spleen bias)
ALC-0315 Pfizer/BioNTech clinical vaccines (broader tropism)
C12-200 Love et al. PNAS 2010 hepatocyte gene editing
FO-32, FO-35 Witten et al. Nat Biotech 2025 lung delivery (ML-designed)
saRNA generic Arcturus disclosures self-amplifying mRNA
tumor-it Costa et al. IJN 2025 intratumoral TRAIL mRNA

The recommender is the human-facing shortlist layer above the trained models in Witten et al. 2025 (>9,000 measurements, 1.6M candidates screened in silico) and Li et al. 2024 (combinatorial chemistry + ML). Practitioners still pick from a shortlist, and this tool is the interface.