Summary
Predicting how cells develop their specialized roles from gene activity data is challenging because many different explanations can fit the data equally well but suggest different cell fates. The authors created FateMultiplicity, a way to identify all plausible fate assignments without needing prior labels, revealing that uncertainty in these predictions is large and depends on the variety of possible models considered. They found that trying to certify single cells as having reliable fate predictions is not better than trusting the main model’s own confidence. Some new measures can separate real developmental decisions from false signals, improving interpretation of these complex predictions.
What this means in practice
- •For biomedical data analysts: Assess uncertainty in cell developmental paths when analyzing single-cell transcriptomics without relying on known lineage labels.
- •For bioinformatics software developers: Incorporate methods to quantify multiple consistent cell fate models and identify unreliable fate assignments in trajectory inference tools.
Tested on simulated data.
Abstract
Single-cell trajectory inference maps transcriptomic measurements onto developmental continua, yet configurations that fit the data equally well can assign conflicting cell fates. FateMultiplicity is a label-free framework that constructs a statistically admissible model set, or Rashomon set, without lineage labels, by evaluating model discrepancy on cross-fitted held-out genes under non-inferiority testing calibrated against random-seed variation. Multiplicity is large and depends more on the diversity of the model space than its size: twelve configurations of a second algorithm expose 20.0% of cells where twenty-four of the first expose 3.8%. Whether the per-cell certified fate margin FM yields more reliable assignments than the fitted model already provides is then tested, and it does not. On simulation ground truth, on the same cells, FM discriminates misassignment at AUC 0.682, against 0.965 for the baseline configuration's own decision margin (p = 0.003) and 0.854 for a seed-dispersion baseline. Informativeness is governed by the breadth of the admitted set, not its cardinality: at cardinality four, seed refits give 0.933 and hyperparameter-perturbed sets 0.701. Relaxing the infimum to a q-quantile recovers discrimination but converges toward the single model's own confidence; the supremum reaches 0.973 because theta*'s membership bounds it from below, while the infimum is unanchored. Multiplicity in trajectory inference is worth measuring and reporting, but per-cell certification over a label-free Rashomon set is not a route to more reliable fate calls. Two constructions survive: a margin-erosion ratio separates real from spurious branch points in simulation (AUC 0.890, untested on real data), and against clonally observed fate, uncertified cells disagree with their clone's outcome 16.4 percentage points more often than certified cells (p < 0.001).