Every cell starts identical and pluripotent — able to become almost any cell type. What it actually becomes is decided by which combination of extracellular signaling factors (growth factors and morphogens) it receives, and for how long. Each target lineage here needs its own exact factor pair; the model scores each cell's commitment probability by how closely the applied cocktail matches that pair, then rolls the outcome per cell so the population settles stochastically, exactly as a real bioreactor batch would.
overlap = matched / required
commit(L) = overlap² · e^(−0.35·extras) · min(1, days/7)
if Σcommit > 1 → normalize (splits into a mix)
remaining probability → stays pluripotent
- Exact pair, full duration — high, tight commitment down one branch: a pure product.
- Missing a factor or too short a culture — most cells never commit and stay pluripotent: undifferentiated, therapeutically useless.
- Extra unrelated factors, or two full pairs at once — commitment probability spreads across branches: an uncontrolled mixed population, the exact contamination risk that manufacturing QC exists to catch.
Real-world relevance: this factor/timing sensitivity is why clinical stem-cell-derived products (dopaminergic neurons for Parkinson's, cardiomyocyte patches, islet-replacement beta cells) are manufactured under tightly controlled differentiation protocols and released only after a purity assay confirms the batch converged on the intended cell type.