Embedded Files
Understanding Disease Biology Language Using LCM and CAG by Replacing Human Language
Understanding disease biology language involves replacing human language in multimodal oncology knowledge abstraction and integration models with Large Concept Models (LCMs) instead of traditional Large Language Models (LLMs). With over 200 human languages, the focus shifts to the language of disease biology. The recent release of LCM by Meta emphasizes knowledge abstraction based on Concept, Content, and Language (CCL). Unlike LLMs, LCMs move from token-based processing to reasoning at the sentence level by embedding entire sentences, eliminating the need for next-token prediction.
Additionally, Cache-Augmented Generation (CAG) replaces Retrieval-Augmented Generation (RAG) models. While RAG suffers from retrieval delays and errors, CAG preloads knowledge directly into large language models, ensuring lightning-fast and accurate responses. CAG thus offers a superior alternative to RAG for building knowledge bases and graph models to understand cellular-level interactions in disease versus normal states.
The year 2025 is predicted to be a transformative year for GenAI in healthcare, according to industry experts. Let's harness the power of GenAI, integrate multimodal data, and deepen our understanding of disease biology through collaboration and team building.
AI Therapeutic Strategy for TAM Macrophage Reprogramming in CRC TME via Pathway-Guided Spatial Omics and LLM Integration
A novel approach has been designed and demonstrated here as a framework for reprogramming tumor-associated macrophages (TAMs) from the pro-tumor M2 phenotype to the tumoricidal M1 phenotype, emphasizing the integration of signaling pathways, TAM marker databases, and advanced Large Language Model (LLM) agents trained on spatial omics data such as MERFISH, CODEX, and XENIUM within the colorectal cancer (CRC) tumor microenvironment (TME). The approach highlights critical pathways such as cGAS-STING, CD47-SIRPα, IL-6/STAT3, CCL2/CCR2, and TGF-β, which serve as central targets for therapeutic strategies aimed at restoring immune balance and enhancing anti-tumor responses.
The workflow depicts how LLM agents, trained on spatially resolved data, identify and analyze gene expression patterns, cellular interactions, and co-expression trends within the TME. These agents extract specific pathway features, enabling a comprehensive understanding of signaling dynamics. The TAM marker database is integral to this workflow, classifying markers into M1 (tumoricidal) and M2 (pro-tumor) categories. For instance, markers like CD86, CD80, and iNOS identify M1 macrophages, while CD163, ARG1, and TGF-β are indicative of M2 macrophages. Using these markers, LLM agents localize and analyze TAM populations within the TME, correlating their phenotype with pathway activity.
This data-driven methodology leverages the strengths of LLMs to build a digital library of spatial omics information, offering an unprecedented resolution of pathway analysis. By mapping the activation and suppression of pathways like CD47-SIRPα, which mediates the “don’t eat me” signal, or cGAS-STING, which triggers innate immune responses, the system identifies regions of therapeutic potential. The strategy underscores how this integration of computational tools, TAM marker data, and spatial transcriptomics facilitates the rational design of therapies targeting specific TAM phenotypes, ultimately paving the way for precision oncology in CRC TME.
Architecting LLM Embeddings for Knowledge Graph Construction Using Multimodal Spatial Omics Data
The architecture for integrating multimodal spatial omics data, including imaging and single-cell data, with PubMed database text involves several steps of data integration, feature extraction, and embedding in a high-dimensional space using large language models (LLMs). Initially, tools like SpaTalk, Seurat, and Scanpy are employed to integrate diverse data types such as CODEX, TILs, scRNA, nuclei, and text. Following this, feature extraction tools like Morph and STUtility recognize patterns and align features spatially and temporally. These features are then embedded using LLMs and transformers into vectors that capture semantic relationships and interactions among nodes, such as genes and cells. Knowledge graph embedding techniques like TransE and RotatE, along with BioBERT and PyKEEN, are utilized to refine these embeddings further. The process culminates in constructing knowledge graphs and graph neural networks (GNNs) with tools like Neo4j and RDF, enabling visualization and analysis of complex data interactions.
This approach is pivotal in drug discovery and targeted immunotherapy, especially in understanding the tumor microenvironment (TME). By integrating and analyzing multimodal data, researchers can uncover intricate patterns and correlations that are otherwise overlooked. Utilizing LLMs and foundation models, the embeddings capture the latent space of these interactions, providing a comprehensive view of biological processes. Knowledge graphs and GNNs built from these embeddings allow for detailed mapping of interactions within the TME, identifying potential drug targets and biomarkers. The no-code/less-code paradigm, facilitated by APIs and libraries, simplifies the construction of this pipeline, enabling efficient data processing and integration. Ultimately, this method enhances our understanding of complex biological systems and improves patient outcomes through personalized medicine and precision therapies.
MultiOmics RAG/GraphRAG Architecture and Data Embedding Feature Space
The MultiOmics RAG/GraphRAG architecture is a sophisticated framework designed to process and analyze spatial omics data. It comprises several stages: Data Collection & Integration, utilizing tools like MISO, VISTA-2D, Mcadet, Pubget, and NER; Feature Extraction (LLMs) with tools such as LangChain, SpaCy, and NLTK; Chunking & Summarizing; Embedding using Milvus, Weaviate, and Pinecone; and Vector Store/RAG Query with vector databases like PostgreSQL, MariaDB, SQLite, KD.AI, Qdrant, and Vectorize. This structured approach facilitates the efficient management and retrieval of complex spatial omics data.
In the realm of drug discovery, this architecture enables the integration and analysis of diverse multimodal data, leading to the identification of novel drug targets and therapeutic pathways. By leveraging spatial omics data, researchers can uncover predictive biomarkers for targeted immunotherapy, enabling more precise and effective treatments for patients. The ability to process and analyze large-scale data with the MultiOmics RAG/GraphRAG architecture accelerates the development of new drugs and personalized therapies, ultimately improving patient outcomes.
Let LLMs Pick the Right Tool and Data: Unlocking Multimodal Spatial Omics for Targeted Immunotherapy and Drug Discovery
Model Context Protocol (MCP) is a standardized interface enabling large language models (LLMs) to connect with multimodal data sources. Its significance lies in accelerating drug discovery and targeted immunotherapy by integrating and extracting insights from spatial omics data, fostering innovations through structured knowledge graphs like GraphRAG.
MCP can serve as an interface for LLMs to connect with multimodal data sources, such as spatial transcriptomics, single-cell RNA sequencing (scRNA), PubMed articles, tabular patient data, and electronic health records (EHR). MCP enables LLMs to extract features from these diverse datasets and deploy AI models effectively for tasks like building knowledge graphs or GraphRAG (Graph Retrieval-Augmented Generation).
MCP acts as a bridge, providing a standardized protocol for integrating LLMs with external data sources and tools. This allows LLMs to retrieve, process, and analyze complex datasets, making it possible to generate insights, establish relationships, and create structured representations like knowledge graphs. By leveraging MCP, we can enhance the capabilities of LLMs to work with multimodal data in a seamless and efficient manner for designing personalized genomic medicine, predictive biomarkers and understanding biological pathways.
AI Agents in immunotherapy: Enhancing Biomarkers, Therapies, and Outcomes
AI agents serve a critical role in addressing the challenges and limitations of immunotherapy by training on diverse multimodal datasets, such as spatial transcriptomics, scRNAseq, PubMed literature, and clinical records. These agents can identify hidden patterns and relationships within the data, offering insights into complex mechanisms like therapy resistance and immune escape. By analyzing biomarkers across various patient populations, AI agents enable data-driven refinement of treatment strategies, including patient stratification based on molecular profiles and identifying subsets of patients most likely to benefit from specific therapies like PD-1/PD-L1 inhibitors, CAR-T therapy, or monoclonal antibodies. This training empowers AI agents to become a robust knowledge hub, synthesizing data and highlighting areas for improvement, such as designing combination therapies or targeting resistant tumor subtypes.
Clinically, AI agents are pivotal in immunotherapy research by uncovering predictive and resistance biomarkers, optimizing therapeutic combinations, and enhancing our understanding of the tumor microenvironment in CRC cancer. Through spatial transcriptomics, they map immune cell infiltration, while scRNAseq enables detailed profiling of cellular responses. AI agents also suggest novel approaches to overcome therapy resistance by integrating multimodal insights to identify alternate pathways or combination strategies tailored to patient-specific needs. Predictor such as PD-L1, TMB, microsatellite instability, etc. do not sufficiently address these issues. These efforts are fundamental to driving personalized immunotherapy, understanding central role of Dendritic cells (DCs) and macrophages in various immunotherapy approaches (model is under development for CRC cancer), ensuring that treatments are more effective, precise, and adaptable to diverse patient populations, ultimately advancing the clinical impact of immunotherapy in cancer care.
Unraveling Tumor-Immune Dynamics: Macrophage-Based Targeted Immunotherapy in CRC Using LLMs and Knowledge Graphs
In this analysis, spatial mapping of CD163, CD68, and MUC1 markers in CODEX data for the small intestine provides a powerful approach to understand their combined role in the tumor microenvironment (TME) of colorectal cancer (CRC). By visualizing the spatial relationships between immune cells and tumor cells, we aim to decode interactions within the TME to inform macrophage-based targeted immunotherapies.
M2 macrophages, marked by CD163 and CD68, are known for their immunosuppressive roles and their ability to promote tumor growth through immune evasion and angiogenesis. When analyzed alongside MUC1, which is aberrantly overexpressed on CRC tumor cells, regions of marker co-expression can reveal critical macrophage-tumor cell dynamics in the TME. Such regions may indicate macrophages actively supporting tumor progression, suggesting targets for therapies that reprogram M2 macrophages to a tumor-suppressive phenotype while attacking MUC1-positive tumor cells.
The figure illustrates the spatial patterns of these markers and integrates them into a knowledge graph, capturing both their interactions and their influence on neighboring cells. By visualizing molecular relationships within the TME, the approach highlights pathways and cell communication networks critical for tumor growth, potentially identifying intervention points for combined therapies.
These integrated methodologies, including spatial analysis and knowledge graph construction, allow researchers to untangle the complexities of the TME. Such insights enable the design of precise, macrophage-focused immunotherapies, offering a promising avenue for improving CRC treatment outcomes.
Agentic AI-Orchestrated Knowledge Graph: Unlocking DC-Macrophage Therapeutic Insights for CRC Immunotherapy
Transforming the landscape of targeted immunotherapy, a novel approach reveals how advanced large language model (LLM) agents harness multimodal spatial omics data to construct enriched knowledge graphs (KGs), advancing our understanding of the colorectal cancer (CRC) tumor microenvironment (TME). This allows LLM agents to uncover intricate connections and crosstalk between genes, proteins, and signaling pathways like STING, SMAD2, and TREM2, facilitating therapeutic strategy development.
As illustrated in figure (lower right), the KG nodes and connections highlighted in the context of CRC TME, dendritic cells (DCs), macrophages, cytokines, and signaling pathways present a detailed narrative. TREM2 emerges as a key regulator of tumor-associated macrophages (TAMs), favoring their M2-like immunosuppressive state and facilitating CRC progression. SMAD2, tied to TGF-β signaling, amplifies immune suppression by inhibiting DC maturation and fostering TAM activity. On the other hand, the STING pathway, central to innate immunity, counters these suppressive forces by producing type I interferons (IFN-I), enhancing DC antigen presentation, and reprogramming TAMs into a pro-inflammatory M1 phenotype, actively aiding antitumor responses. This reveals the complex interplay of these entities within the TME, exposing how cancer evasion mechanisms and immune activation pathways continuously oppose each other.
In the context of DCs and macrophages, the highlighted nodes explain their functional states and therapeutic reprogramming potential. DCs, primarily activated through the cGAS-STING pathway, contribute to T-cell priming and the tumor immune cycle via enhanced antigen presentation. Macrophages, polarized into TAMs, use TREM2 and PI3K/AKT/mTOR pathways to suppress immune responses. However, targeted strategies like TREM2 inhibitors aim to reprogram macrophages, shifting their phenotypes from M2 to M1, while STING agonists stimulate DC activity and cytokine production, synergizing innate and adaptive immunity. The connections reveal how disrupting suppressive signaling pathways restores immune balance and initiates robust antitumor immunity.
Therapeutically, the knowledge graph highlights strategies that integrate these pathways for immunotherapy. STING agonists stand out as pivotal tools to enhance both DC and TAM functionalities, while TGF-β/SMAD2 inhibitors and TREM2-targeted therapies strategically dismantle immunosuppressive barriers in CRC. The interplay of cytokines such as IL-12, IFN-α, and TNF-α amplifies immune activation, while signaling pathways like NF-κB and cGAS-STING drive inflammation and immunity. Together, these therapeutic approaches, guided by the network of nodes and connections, offer new insights for overcoming immune evasion in CRC and improving treatment outcomes through a multi-faceted immunotherapy approach. This approach highlights the transformative role of AI-driven multimodal data analysis in advancing precision oncology. The resulting Agentic AI built KG provides a robust framework for targeted immunotherapy by identifying actionable patterns in immune regulation, reprogramming DCs and TAMs, and optimizing cytokine interventions to counter CRC progression.
LLM Agents Transforming CRC TME Analysis for Precision Immunotherapy
A novel approach is designed and tested as POC here to demonstrate how large language model (LLM) agents analyze multimodal (MM) data, including spatial transcriptomics and single-cell RNA sequencing, to identify critical pathways in the colorectal cancer (CRC) tumor microenvironment (TME). LLM agents ingest and integrate data to extract key molecular patterns, such as immune checkpoints (PD-1/PD-L1, CTLA-4) and signaling pathways (TGF-β, CD47-SIRPα, VEGF), from spatially resolved multiomics datasets. They construct knowledge graphs to map the interplay of immune cells and pathways and correlate these insights with targeted immunotherapies like checkpoint inhibitors, macrophage-modulating strategies, and cytokine therapies. This integrative process enables the design of advanced therapeutic strategies by uncovering the molecular underpinnings of CRC TME.
These efforts are pivotal in driving personalized medicine and targeted immunotherapy in clinical settings. By identifying pathways and extracting immune-regulatory patterns from complex MM data, LLM agents facilitate precise therapeutic interventions tailored to individual patients. Their ability to integrate spatial omics data in real-time clinical workflows ensures better identification of biomarkers, improved treatment efficacy, and the development of innovative therapeutic strategies tailored to unique TME characteristics, highlighting AI's transformative role in oncology.
Development of a Targeted Immunotherapy Pipeline Using LLM Agents
Building on previous efforts, this work advances the pipeline for biomarker discovery and immunotherapy development. It focuses on understanding immune pathways and aligning them with targeted therapies using large language model (LLM) agents, pushing beyond conventional approaches.
The upper-left section of the figure outlines the training process for immune-activating and immune-suppressive LLM agents in relation to Tumor-Infiltrating Lymphocytes (TILs). Immune-activating pathways (illustrated with green circles), such as Type I interferon signaling (IFN-α/β), CXCR3/CXCL9/CXCL10 chemokines, TCR signaling, IFN-γ, IL-12, and apoptosis mechanisms like Fas-FasL and Granzyme B, are used for training. Conversely, immune-suppressive factors (shown in red circles), including TGF-β, VEGF, IDO pathways, hypoxia-induced HIF-1α, CXCL12-CXCR4 axis, and PD-1/PD-L1 signaling, form the basis for immune-suppressive training. This distinction emphasizes the diverse roles of TIL-associated pathways, classified as either activating or suppressing immune responses.
The figure (bottom) also highlights the training pipeline for multimodal (MM) CODEX data and marker databases. It demonstrates how CODEX data is leveraged to train LLM agents in identifying TIL-related markers within the colorectal cancer (CRC) tumor microenvironment (TME). These markers are pivotal for uncovering immuno-activating and immunosuppressive pathways, offering insights into targeted immunotherapy.
This work showcases a systematic approach to biomarker training for LLM agents. It incorporates therapeutic pairing, data integration, feature extraction, model refinement, and pathway prioritization. By analyzing high-dimensional data to extract patterns and identify key pathways, these agents enable targeted therapy recommendations. This effort marks the evolution from proof-of-concept to proof-of-value applications of Agentic AI in drug discovery, demonstrating its transformative potential in advancing targeted cancer adaptive cell immunotherapy.
Therapeutic AI Strategy for Targeting the CD47-SIRPα Pathway: Converting the 'Don't Eat Me' Signal into an 'Eat Me' Signal
The effort focuses on a therapeutic strategy to convert the "don't eat me" signal into an "eat me" signal by targeting the CD47-SIRPα pathway. This approach aims to design macrophage and dendritic cell (DC)-based cell therapies, leveraging multimodal spatial transcriptomics data such as MERFISH, Codex, Xenium, and scRNA-seq to understand the tumor microenvironment (TME) in colorectal cancer (CRC), a highly complex system influenced by interactions among immune cells, cytokines, and extracellular matrix components.
MERFISH analysis reveals markers such as IL1B, CXCL8, and S100A9, which shape the inflammatory milieu by driving immune cell recruitment and macrophage polarization toward a tumor-supportive phenotype. Factors like SOCS3 and SPP1 specifically regulate macrophage activity, promoting either pro-inflammatory or immune-suppressive states. Proteins such as COL4A1 and DES highlight structural remodeling within the TME, enabling tumor progression and metastasis. Key cytokines like IL6 reinforce immune evasion and angiogenesis, while enzymes such as PTGS2 contribute to chronic inflammation.
The analysis highlights the expression of immune-modulating genes such as IL1B, CXCL8, and SOCS3, which are mapped to specific tissue regions to identify immune-active "hotspots." These hotspots represent areas where CD47-targeting therapies could be most effective. By targeting specific markers, it becomes possible to disrupt immune suppression, prevent extracellular matrix remodeling, and modulate macrophage polarization, paving the way for improved CRC treatment outcomes.
Training LLM agents, along with predictions and feedback loops, refines therapeutic strategies by identifying key immune-modulating genes and regions within the TME that are most responsive to intervention. This integrated approach offers a promising pathway for enhancing anti-tumor immunity through macrophage and DC-based therapies.
Moreover, LLM agents play a pivotal role in uncovering the complexities of the CD47-SIRPα pathway and recommending strategies to effectively convert the "don't eat me" signal into an "eat me" signal, advancing the development of targeted immunotherapies. Their ability to synthesize large-scale multimodal data ensures precise and actionable insights for improved therapeutic outcomes.
Decoding CD47-SIRPα Axis Interactions: Computational Insights into Tamoxifen's Role in CRC Immunotherapy
Efforts to understand the interplay between the CD47-SIRPα axis and SHP1 signaling were advanced through the construction of LLM agent-powered knowledge graphs (KG1 and KG2), alongside integrated multi-modal (MM) data analysis. These analyses included insights from single-cell RNA (scRNA) sequencing, which revealed overexpressed phenotypes in colorectal cancer (CRC) related to immune evasion mechanisms. Tamoxifen's potential role in modulating immune responses was predicted within these networks, suggesting it could alter macrophage polarization, inhibit SHP1 activity, and synergize with CD47-targeting therapies to enhance macrophage and dendritic cell-mediated tumor clearance.
The figure showcases the structured research efforts—spanning immuno-signaling pathways, scRNA analysis, PubMed literature mining, pathway enrichment studies, and knowledge graph integration. It highlights cross-pathway interactions, tamoxifen's influence on SHP1 signaling, and its repurposing potential, blending computational insights with biological discovery and therapeutic innovation.
Decoding Immune Pathways by Training LLM Agents: Toward Targeted Therapy Development
The effort presented here focuses on harnessing large language models (LLMs) agents training to analyze multimodal (MM) unstructured data, such as imaging, spatial omics, single-cell RNA sequencing (scRNA Seq), PubMed literature, and clinical patients (tabular) data in understanding complex immune-suppressive and immune-supportive signaling pathways. This approach enables us to address critical questions, such as why patients respond to specific therapies and develop resistance to others, providing a foundation for developing effective strategies in adaptive cell therapy and other immunotherapies.
Training LLM agents involves integrating spatial omics features and patterns to identify immune-suppressive and immune-supportive signaling pathways, such as those involving dendritic cells (DCs), macrophages, chimeric antigen receptors (CAR), and T-cell receptors (TCR). Training LLM agent for scRNA Seq data analysis of the intestine reveals a detailed UMAP visualization showcasing cellular diversity and heterogeneity. It identifies key immune cell types, including tumor-associated macrophages (TAMs) and DCs, marked by CD68. Further cell-subtype analysis by LLM agent uncovers distinct populations such as LYVE1 macrophages, SPP1 macrophages, conventional dendritic cells (cDC1 and cDC2), plasmacytoid dendritic cells (pDCs),CD14 and CD16 monocytes(illustrated here). This analysis highlights the intricate cellular landscape and provides insights into the functional roles of these subtypes in immune regulation and response mechanisms.
Integrating this methodology into immunotherapy drug development pipelines through artificial intelligence creates a fully automated framework for analyzing therapeutic responses and resistance. By training LLM agents on each step of this process, researchers can achieve a comprehensive understanding of the underlying molecular mechanisms, making it a cornerstone for creating innovative and effective therapeutic strategies.
Mapping Contributing Features to Spatial Distribution of Cell Localization
The knowledge graph (KG) identifies clusters of cells based on their features, such as gene expression profiles, cell types, and spatial relationships. These clusters are then mapped back to the spatial transcriptomics images, allowing researchers to visualize the spatial distribution of cells and their contributing features. This mapping helps in understanding the localization and interaction of different cell types within the tumor microenvironment, providing insights into the underlying biological processes and potential therapeutic targets.
The image presents above a detailed workflow integrating multi-modality imaging data (histopathology and spatial omics) with large language models (LLMs) and knowledge graphs to analyze tumor-infiltrating lymphocytes (TILs) in the tumor microenvironment (TME) of colorectal cancer (CRC). The process involves identifying clusters of cells in the knowledge graph, then mapping these clusters back to spatial transcriptomics images to understand the contributing attributes (features) and their spatial localization (distribution).
The red and green bars in the image represent the relative importance of different features in the analysis. The red bars indicate features that have a negative contribution, while the green bars indicate features that have a positive contribution to the model's prediction. This helps in understanding which features are driving the model's decision-making process
LLMs in Biomedicine (Clinical AI)
Meditron: A large language model for biomedical applications. Used for medical text analysis and understanding.
BioBERT: A biomedical version of BERT. Applied in biomedical text mining and information retrieval.
BioGPT: A biomedical adaptation of GPT. Utilized for generating and understanding biomedical text.
BioMedGPT: A specialized GPT model for biomedical contexts. Used for generating biomedical research summaries and insights.
BioMegatron: A large-scale transformer model for biomedical data. Applied in large-scale biomedical data processing and analysis.
PubMedBERT: A BERT model trained on PubMed articles. Used for extracting and understanding information from biomedical literature.
PubMedGPT: A GPT model trained on PubMed data. Utilized for generating and summarizing biomedical research articles.
Med-PaLM: A medical version of the PaLM model. Applied in medical diagnostics and patient care recommendations.
BLOOM: A large-scale language model for various domains, including biomedicine. Used for cross-domain biomedical research and applications.
AntGLM-Med: A biomedical version of the AntGLM model. Applied in predictive modeling and biomedical research.
GatorTron: A biomedical language model developed by the University of Florida. Used for clinical text analysis and patient data processing.
Hi-BEHRT: A hierarchical biomedical version of BEHRT. Applied in hierarchical biomedical data analysis.
ClinicalBERT: A BERT model adapted for clinical text. Used for clinical text mining and patient record analysis.
These algorithms enhance various aspects of clinical trials, from data analysis and patient recruitment to generating insights and improving diagnostics. They streamline processes, improve accuracy, and ultimately contribute to more efficient and effective clinical research.
Advanced Computational Approaches in GenAI for MASH Liver Diagnostics: Integrating In-Context Learning (ICL) and Large Language Models (LLMs)
In-Context Learning (ICL), a recent model developed by Meta, offers significant advantages in the realm of machine learning and natural language processing. Unlike traditional fine-tuning methods, ICL allows models to adapt to new tasks quickly by conditioning on provided examples without the need for parameter updates. This agility and efficiency make ICL a powerful tool for prompt engineering, where the quality of prompts directly impacts the model's performance. By combining ICL with fine-tuning, one can leverage the flexibility of ICL and the specialized expertise of fine-tuned models, resulting in a robust and versatile approach to various tasks.
Metabolic dysfunction-associated steatohepatitis (MASH) is a liver disease characterized by inflammation and fibrosis. Understanding liver zonation, particularly the mid-lobular zone, is crucial for studying the onset of MASH, as this area is key to liver regeneration and fibrosis triggering. This use case of ICL is particularly relevant for testing in MASH, where Liver Sinusoidal Endothelial Cells (LSECs) play a vital role in liver zonation, regulating hepatic vascular pressure and exhibiting anti-inflammatory and anti-fibrotic functions. Hepatic Stellate Cells (HSCs), on the other hand, are central to fibrosis and inflammation, producing extracellular matrix components that contribute to fibrosis. The interplay between LSECs and HSCs is essential for understanding the progression of MASH. According to the study "Spatial Computational Hepatic Molecular Biomarker Reveals LSECs Role in Mid Lobular Liver Zonation Fibrosis in DILI and NASH Induced Liver Injury," LSECs in the mid-lobular zone are crucial for early fibrosis detection and liver regeneration.
Integrating spatial transcriptomics data, single-cell RNA-seq data, and histopathology H&E liver images using ICL and large language models (LLMs) in prompt engineering can significantly enhance our understanding of MASH. By building a knowledge graph that incorporates these diverse data types, researchers can develop computational diagnostic biomarkers that provide insights into the spatial distribution and heterogeneity of cell types and subtypes in MASH liver. This approach, powered by Generative AI (GenAI, is under testing on more data sets and at validation stages), can improve diagnostic accuracy and offer new perspectives on disease mechanisms, ultimately aiding in the development of targeted therapies.
Therpeutic targets: MASH (LLMs models multimodal data integeration analysis)
Page updated
Google Sites
Report abuse