Domain
Fidelity
Sort
Fidelity vs Complexity
72 trials. Each dot one RCT. X = trial phenotype complexity, Y = match to published result.
A searchable record of published outcomes compared with oNeura simulations
What is shown: Each card compares a published clinical trial endpoint with the corresponding oNeura simulation. The fidelity score shows how closely the simulated endpoint matched the published result.
Pipeline: Drug properties such as LogP, TPSA and MW are computed from SMILES molecular structure using atom-contribution methods. Binding kinetics use diffusion-limited on-rates and transition-state off-rates. Multi-organ pharmacodynamics are generated by the Human Biology Loop simulation with 40+ coupled organ modules.
Limitations: Atom-contribution LogP/PSA methods have typical errors of ±0.5 and ±8 Ų respectively. Simulated trial outcomes reflect a modeling pipeline, not patient-level data. Fidelity scores measure endpoint-level agreement, not causal mechanism validation. These results should be interpreted as research benchmarks, not clinical evidence.
Trial data: Published endpoints are sourced from original trial publications. All trials are real, registered clinical trials. The "simulated" column shows modeled outcomes from the oNeura pipeline.