Interferon response digital twin for treatment… (F31LM015310) – Project Grant to University Of Rochester

Abstract: Interferon (IFN) signaling is a key component of immune-mediated inflammatory diseases (imids), functioning in a disease- and cell-specific manner. Ifns modulate immune responses across diverse conditions and play a large role in the overall pathogenesis of many diseases. However, current IFN network models do not account for diseases context or cell type specificity, limiting their ability to uncover novel effectors of imid pathogenesis and translation into clinical interventions. An applicable case for context-dependent modeling is systemic lupus erythematosus (SLE), a heterogeneous autoimmune disease of which lupus nephritis of the kidney is a severe manifestation. Yet, the mechanisms of IFN signaling and intercellular crosstalk within the kidney remain poorly understood. Modeling this context-dependence using in silico dynamic tools- such as the biologically interpretable scbonita algorithm from thakar lab- and scrna-SEQ data from patients, will enable accurate representation of personalized IFN response across cell types through cell steady state analyses. This project will leverage scbonita modeling with two innovative machine learning (AI/ML) applications. Aim i will employ deep learning neural networks to infer robust and novel logic rules governing regulatory behavior within the IFN network. Aim II will integrate a bespoke ML model that identifies features contributing to disease- and cell-specificity within individual patients. These efforts will result in the first patient-specific deep learning model that learns logic gates directly from scrna-SEQ data for a specific immune-mediating signaling pathway. Scbonita allows for generation of personalized, virtual representations of patients- immunological digital twins (idts)- that reflect intra- and inter-patient heterogeneity. These distinct idts can be characterized, analyzed for key regulators and cell steady states, and mapped to associated clinical metadata for translational relevance through perturbation analyses. Evaluation and interpretation of idts for clinical use require both rigorous methodology and interdisciplinary collaboration, which this proposal is well-positioned to address. The goal of this project is to construct a novel immunological digital twin for immune-mediated inflammatory diseases that model the context-dependent IFN response, through systems immunology and AI/ML techniques, carving the path towards personalized medicine. This proposal addresses the need to train future physician-investigators in the biomedical workforce who can bridge medical knowledge with computational innovation by elevating the powers of AI/ML to advance discovery. The models, tools, and insights developed through this proposal will serve as foundational resources for next-generation intervention strategies conducted in silico.

Interferon response digital twin for treatment in systemic lupus erythematosus

Award facts

Description

Quick answers

Who received award F31LM015310?

How much is award F31LM015310 worth?

See this award's full record