Based on the trajectory from 1.9, the next major version (2.0, ~2027–2028) will likely include:
| Feature | QSP 1.9 | QSP 2.0 (forecast) | |---------|---------|--------------------| | Hybrid modeling | Mechanistic ODEs only | ODEs + neural ODEs for unknown submodules | | Real-time calibration | Offline (batch) | Online Bayesian filtering (e.g., particle MCMC) | | Patient digital twins | Virtual populations | Individualized via clinical covariates + short pre-dose data | | Interoperability | Proprietary APIs | HL7/FHIR integration for EHR data | | Causality inference | Manual network construction | Automated from perturbation screens + causal discovery (e.g., DAGs) |
A critical enabler will be federated QSP learning – training models across pharma companies without sharing proprietary data. qsp 1.9
Traditional pharmacokinetic/pharmacodynamic (PK/PD) models struggle with complex, multi-pathway diseases (cancer, autoimmune disorders, fibrosis). QSP addresses this by coupling drug action with dynamic, mechanistic networks of signaling, metabolism, and immune response.
Version 1.9 (a representative label for contemporary QSP software) introduces three breakthroughs: Based on the trajectory from 1
Unlike earlier versions (1.0–1.5), which required extensive coding and manual fitting, 1.9 offers graphical user interfaces (GUIs) and scripting APIs (Python/MATLAB) for industrial use.
QSP 1.9 is not a revolution but a necessary maturation — it transforms mechanistic biology from an academic exercise into a predictive, pragmatic tool for drug development. Its ability to integrate with PBPK, handle virtual populations, and undergo automated reduction makes it the first version suitable for routine industrial use. However, until parameter identifiability is solved and regulatory pathways are formalized, QSP will remain a powerful complement — not a replacement — for clinical trials. Unlike earlier versions (1
For practitioners: adopt QSP 1.9 if your question involves ≥2 interacting pathways or combination therapies. For simple small-molecule PK/PD, stick with standard models. For the rest of us, QSP 1.9 is the closest we have to a digital patient.
Unlike its predecessors that required local high-performance computing clusters, QSP 1.9 is cloud-native. It supports parallel simulations on AWS, Azure, and Google Cloud, making it accessible to small biotechs and academic labs.