TLDR: ApiNATOMY is a framework with knowledge representation and management tools that helps researchers create, visualize, and integrate multiscale models of the peripheral nervous system and other physiological systems. It simplifies complex anatomical connectivity data, supports the SPARC program’s goals, and integrates with existing knowledge bases like SCKAN and SciGraph, making biological data more understandable and usable for developing new therapies.
Understanding the intricate networks of the human body, especially the nervous system, is a monumental task. Researchers are constantly seeking better ways to map and interpret complex biological data. A new framework called ApiNATOMY is emerging as a crucial tool in this endeavor, providing a sophisticated way to represent and manage multiscale physiological circuit maps.
ApiNATOMY is designed to support researchers in mapping data related to the peripheral nervous system and other physiological systems, focusing on their relevance to specific organs. It combines a Knowledge Representation (KR) model with a suite of Knowledge Management (KM) tools. The KR model allows physiology experts to easily capture interactions between anatomical entities, while the KM tools help convert high-level ideas into detailed models of physiological processes. These models can then be integrated with existing ontologies and knowledge graphs, ensuring a standardized and shared understanding of biological data.
Ontologies are essentially standardized vocabularies that define relationships, making it possible for different systems to exchange data with consistent meaning. They are vital in biological and biomedical research for providing standard identifiers, including metadata, offering machine-readable definitions, and standardizing vocabulary across diverse data sources. In neuroscience, ontology-based data integration is particularly important for synthesizing knowledge across various fields like physiology, anatomy, and molecular biology.
A significant application of ApiNATOMY is within the Stimulating Peripheral Activity to Relieve Conditions (SPARC) program, an NIH-funded initiative. SPARC aims to understand how the Autonomic Nervous System (ANS) interacts with end organs and the Central Nervous System (CNS) to develop new neuromodulation therapies. ApiNATOMY helps in mapping the pathways of ANS neuron populations, identifying anatomical structures they traverse or terminate in. This mapping forms the organizational backbone for SPARC’s data repository and computational simulations.
The SPARC Connectivity Knowledge Base of the Autonomic Nervous System (SCKAN) is a network of conduits representing connectivity. SCKAN, which incorporates ApiNATOMY models, encapsulates detailed knowledge about CNS-ANS-end organ circuitry, derived from SPARC data and scientific literature. It is structured to support computational reasoning and uses standard reference vocabularies for data integration. Natural Language Processing (NLP) is also employed to extract connectivity relationships from scientific texts, further enriching this knowledge base.
Visualizing and Managing Knowledge
A central element of the ApiNATOMY approach is its visual toolkit, which supports data integration in a multiscale connectivity model of cells, tissues, and organs. This toolset includes both KR and KM tools that enable topological and semantic modeling of physiological process routes and associated anatomical compartments. Domain experts can provide model specifications in a semi-structured format, which ApiNATOMY tools then expand into a visual graph that can be manipulated, saved, and integrated with other knowledge bases like the Neuroscience Information Framework (NIF) ontology and the SPARC Knowledge Graph.
The ApiNATOMY model viewer application is a key component, featuring a dynamic WebGL graph that can be rendered in 2D or 3D. It includes a control panel for adjusting viewer parameters and selecting parts of the model to display. Resource editors, such as the Material editor, Lyph editor, Chain editor, and Coalescence editor, provide graphical user interfaces (GUIs) for defining and modifying various elements of the models. These editors simplify complex tasks, allowing users to create, edit, and delete materials, lyphs (layered compartments representing organs or systems), chains (sequences of connections), and coalescences (overlapping layers for material exchange).
For instance, the Material editor displays a Directed Acyclic Graph (DAG) of material composition, allowing users to visualize and edit relationships between materials and lyphs. The Lyph editor provides hierarchical tree views to manage lyphs and their layers or internal components. These tools are equipped with validation and error handling services to ensure the creation of consistent and accurate models. They also integrate with the SciGraph querying API, enabling users to annotate models with terms from other biomedical ontologies, which is crucial for merging ApiNATOMY models with other SCKAN resources.
To help arrange connectivity specifications, ApiNATOMY uses “scaffolds.” These are models that define reusable layout templates with fixed or constrained coordinates, providing a wireframe schematic of the body. An example is the TOO map (T-shaped Cerebro-Spinal Fluid, inner O-shape for blood circulation, outer O-shape for surface materials), which offers a subway-style whole-body map for overviewing and quality-checking connectivity. Nodes and chains in the connectivity models can be bound to these scaffold anchors or wires, guiding their visual placement.
The visualization system automatically generates layouts, placing nodes and links based on various properties and algorithms. It also handles multi-scale modeling, allowing large body parts and tiny neuron chains to appear in the same model. Default scaling rules ensure that elements fit appropriately within their hosting structures, making complex anatomical systems easier to understand visually.
Also Read:
- CLIMB: A New Framework for Data-Driven Occupation Taxonomies
- EHR-MCP: Bridging Large Language Models with Electronic Health Records
Integration and Future Directions
Once created and validated, ApiNATOMY models are exported as JSON-LD files, which are then translated into RDF/OWL and Neo4J formats to become part of SciGraph within the SCKAN infrastructure. This integration makes the knowledge accessible and searchable. While the automatically generated layouts are helpful for experts, the SPARC Portal also offers “Flatmaps” – 2D graphical representations of anatomy and connectivity, similar to Google Maps. These flatmaps, which incorporate ApiNATOMY modeling data, provide an intuitive way for users to explore neural connectivity in different species.
The development of ApiNATOMY highlights the extensive effort involved in collaboratively assembling specialized data and the need for suitable editing and quality validation tools. This framework represents a significant step forward in modeling multi-scale physiological systems, offering user-friendly and transparent tools for biomedical experts. The ongoing work aims to further evolve these tools to tackle challenges in physiological system modeling and enable quantitative multi-scale analysis, ultimately advancing our understanding of health and disease. For more detailed information, you can refer to the full research paper here.
