Introduction: The development of immune digital twins represents a critical step toward personalized medicine and in silico clinical trials. However, most current efforts remain at the level of conceptual frameworks or small-scale demonstrations. Here, we present a dynamic modeling framework to construct immune digital twins by integrating single-cell RNA-seq data with perturbation-informed analyses of disease phenotypes.

Methods: We employ Boolean discrete-state modeling to define network-based State Transition Graphs (STGs) that capture intracellular signaling dynamics and identify dominant attractors corresponding to stable cellular states. By integrating these dynamic models with pseudotime inference, we reconstruct disease-specific trajectories and estimate transition probabilities between cellular states. We further incorporate updated datasets and apply the scBONITA framework on gene regulatory networks generated using large language models LLM and the single-cell dataset was recently published here. (Wang et al., 2025).

To enable patient-specific characterization, we introduce clustering metrics that combine attractor similarity, proximity in reduced-dimensional space, and alignment with clinical phenotypes (e.g., AS+ vs AS−). These disease-network–specific stable states, and their combinations, define individualized immune digital twins. In silico perturbation experiments further identify key driver genes whose modulation significantly alters disease trajectories, revealing potential therapeutic targets and vulnerabilities.

This framework builds upon the scBONITA (single-cell Boolean Omics Network Invariant-Time Analysis) approach (Palshikar et al., 2022), extending it to generate patient-specific profiles that summarize attractor landscapes, regulatory drivers, and immune dynamics. Importantly, the proposed methods are extensible to other omics modalities and can interface with emerging agentic AI systems, enabling bidirectional and adaptive digital twins.

 Conclusion: Overall, this work advances a scalable and interpretable approach for modeling immune system dynamics.