Celcomen: spatial causal disentanglement for single-cell and tissue perturbation modeling.
Megas S, Chen DG, Polanski K · Nature communications · 2026 · Article
10.1038/s41467-026-69856-5
Verified evidence
Synthetic-data benchmark
20764cd4-aecb-518e-b632-04f189b6a001Prediction source Version of record · PDF page(s) 3
Study 6eaf5035-d345-5df1-8b55-5529b2f2effb · Prediction record (JSON)
Prediction fields
- Comparator
- experimental_or_observational_reference · PDF page(s) 3
- Endpoint
- inferred_interaction_matrix__qualitative_agreement · PDF page(s) 3
- Family
- agreement · PDF page(s) 3
- Measured variable
- inferred_interaction_matrix · PDF page(s) 3
- Measurement role
- reference_agreement · PDF page(s) 3
- Metric
- qualitative_agreement · PDF page(s) 3
- Population
- simulated_system · PDF page(s) 1, 3
- Reference type
- synthetic_reference · PDF page(s) 3
- Role
- qualitative_model_fit · PDF page(s) 3
- Scenario
- synthetic_reference · PDF page(s) 3
Validation Direct comparison · Same source · Sample overlap · Supports · Matched conditions · Not applicable
Cross-sample stability
eb8ef0ae-9731-520f-bfe3-e0b9137ee23fPrediction source Version of record · PDF page(s) 3
Study b10df589-9152-5e95-a116-fa729afb3455 · Prediction record (JSON)
Prediction fields
- Measured variable
- spatial_transcriptomics_interaction_matrix_association · PDF page(s) 3
- Reference type
- alternative_computational_methods · PDF page(s) 3
- Scenario
- nonoverlapping_training_sample_comparison · PDF page(s) 3
Validation Related evidence · Same source · No sample overlap · Partial support · Partially comparable · Not applicable