Background: Noise from Dissociated Cholangiocyte Structures and the Long-Term Gut-Liver Axis Multi-Omics Data Bottleneck in Cholangitis R&D

Primary sclerosing cholangitis (PSC) is a rare disease characterized by an unclear pathogenesis, progressive fibrosis, and biliary obstruction. Existing single-cell transcriptomics and bulk sequencing guidelines have limitations due to the inability to control noise arising from dissociated cellular structures, resulting in a lack of spatial resolution. In particular, the gut-liver axis-mediated metabolic flux between epithelial and immune cells within the microenvironment has formed a resistant feedback loop, leading to failure in maintaining effective drug concentrations. The American Association for the Study of Liver Diseases (AASLD) guidelines' suspension of oral vancomycin recommendations, as demonstrated in the recent Lancet Seminar (DOI: 10.1016/S0140-6736(26)01258-4), highlights the clinical bottleneck caused by concerns about antibiotic resistance and the lack of sufficient evidence from clinical trials. Existing R&D approaches, which lack precise computational prediction and patient-specific genetic gradient correction, have failed to demonstrate the efficacy of candidate substances, resulting in high clinical failure rates.

Discovery: Implementation of Spatial Barcoding and In Silico Dynamic Modeling, Demonstrating Single-Cell Resolution Gut-Liver Axis Tensor Synchronization

To address this bottleneck, we implemented an integrated omics architecture that combines spatial single-cell transcriptomic barcoding technology with machine learning-based dynamic modeling. We performed single-cell-level gut-liver metabolic tensor synchronization to proactively calculate, in silico, the free energy of ligand-receptor binding between microbial metabolites and receptors within cholangiocytes. Heterogeneous batch effects arising during multi-omics integration were removed using computational autoencoders and matrix decomposition algorithms, revealing the topological variation curves of biliary acid metabolism pathways and transcriptomic networks. This provides a basis for scanning the microbial composition changes and transcriptional activity of immune pathways that occur during oral vancomycin administration. This analytical architecture surpasses existing simple cell line analysis models and fully demonstrates molecular biological integrity in silico.

Coordination of Cholangiocyte Immune-Epithelial Cell Interactions and Establishment of a Precise Layered Model for Reversible Cholangiocyte Homeostasis

This platform establishes a precise layered model that stratifies patients based on a multi-dimensional omics matrix, thereby delineating their molecular phenotypes. By visualizing cell toxicity based on patient-specific gut microbiota and biliary acid concentration gradients, we derived rate constants for reversible restoration of biological homeostasis. We designed an up-clamping technique to simulate the activation of bicarbonate-defending gene expression, which protects the cholangiocytes, and a down-clamping scenario to autonomously regulate the inhibition of harmful signaling cascades within the computational network. This establishes a dynamic ecological homeostasis backbone that allows patient-derived cells to maintain homeostasis even under aberrant biliary stasis stress conditions. This goes beyond simply evaluating the pros and cons of simple antibiotic administration and provides a computational biological solution that predicts clinical response rates in real-time based on the molecular signature of individual patients and induces reversible preservation of cholangiocyte function.

Prospects: Establishing a Standard for Programmable Cholangiopathy Immunotherapy and Implementing Next-Generation IND Digital Governance

This computational biology architecture represents a milestone in transforming R&D governance from a post-hoc symptomatic treatment focus to a programmable infrastructure based on AI-driven multi-dimensional tensors. In the cholangiopathy market, where global pharmaceutical companies such as Gilead (Cilofexor, Phase III) and Intercept (Obeticholic acid) are competing, this platform provides a computational moat by perfectly controlling the genetic gradient correction coefficient at the screening stage, thereby eliminating batch-to-batch variation. By ensuring the validity of the companion diagnostic (CDx) platform, a key standard in digital healthcare, it is expected to disruptively shorten the timeline for the regulatory approval framework for Investigational New Drug (IND) applications by global regulatory agencies. The establishment of digital governance that supports consistent data correction up to the cGMP commercial production stage will be a key asset in building the exclusive value chain of future precision biotech businesses.