Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Jun 24:12:693735.
doi: 10.3389/fphys.2021.693735. eCollection 2021.

The SPARC DRC: Building a Resource for the Autonomic Nervous System Community

Mahyar Osanlouy  1 , Anita Bandrowski  2 , Bernard de Bono  1 , David Brooks  1 , Antonino M Cassarà  3 , Richard Christie  1 , Nazanin Ebrahimi  1 , Tom Gillespie  2 , Jeffrey S Grethe  2 , Leonardo A Guercio  4 , Maci Heal  5 , Mabelle Lin  1 , Niels Kuster  3   6 , Maryann E Martone  2 , Esra Neufeld  3   6 , David P Nickerson  1 , Elias G Soltani  1 , Susan Tappan  5 , Joost B Wagenaar  4 , Katie Zhuang  3 , Peter J Hunter  1
Affiliations

The SPARC DRC: Building a Resource for the Autonomic Nervous System Community

Mahyar Osanlouy et al. Front Physiol. .

Abstract

The Data and Resource Center (DRC) of the NIH-funded SPARC program is developing databases, connectivity maps, and simulation tools for the mammalian autonomic nervous system. The experimental data and mathematical models supplied to the DRC by the SPARC consortium are curated, annotated and semantically linked via a single knowledgebase. A data portal has been developed that allows discovery of data and models both via semantic search and via an interface that includes Google Map-like 2D flatmaps for displaying connectivity, and 3D anatomical organ scaffolds that provide a common coordinate framework for cross-species comparisons. We discuss examples that illustrate the data pipeline, which includes data upload, curation, segmentation (for image data), registration against the flatmaps and scaffolds, and finally display via the web portal, including the link to freely available online computational facilities that will enable neuromodulation hypotheses to be investigated by the autonomic neuroscience community and device manufacturers.

Keywords: FAIR; SPARC; autonomic nervous system; computational life sciences; data annotation; data curation; knowledge management; neural mapping.

PubMed Disclaimer

Conflict of interest statement

AB, MM, and JG have equity interest in SciCrunch.com, a tech startup out of UCSD that develops tools and services for reproducible science, including support for RRIDs. AB is the CEO of SciCrunch.com. ST and MH are company employees of MBF Bioscience, a commercial entity. LG and JW were employed by Blackfynn Inc. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
The data workflow: experimental data and mathematical models created by the SPARC community are uploaded to a data repository managed by DAT-Core, along with metadata that is verified by K-Core and used to create a semantic knowledge base for the ANS community. Image segmentation and other data processing steps, including mapping data and models into appropriate anatomical locations using common coordinate systems, is undertaken by MAP-Core. The simulation environment for using the data and models to understand ANS function, and to assist in the development of neuromodulation therapies and devices that modulate that function, is provided by SIM-Core. All cores contribute to the https://sparc.science portal that provides a resource for the ANS biomedical science and bioengineering communities, and for the manufacturers of neuromodulation devices.
Figure 2
Figure 2
Schematic overview of the relationship between the DMP, Discover, SPARC portal, and o2S2PARC applications. Data are uploaded to the DMP and curated through the SPARC curation workflow. Once a dataset is curated, a snapshot of the dataset is persisted in the Discover platform and a DOI is assigned. Initially, the data is released as an embargoed dataset, with a final release date on which the entire dataset is made publicly available. Users have the ability to request early access to embargoed data through the platform after signing a SPARC Data Use Agreement.
Figure 3
Figure 3
Detail of an ApiNATOMY schematic diagram representing neural connectivity between the rodent spinal cord and the lower urinary tract (Surles-Zeigler et al., 2020). Note that the pathway information in diagrams such as this is stored in SPARC Connectivity Knowledge Base and used to render the pathways in the more anatomically oriented flatmaps (see Figure 4).
Figure 4
Figure 4
Flatmaps of individual species available through the SPARC portal at sparc.science/maps. Zooming in on these mapping diagrams provides additional levels of detail and clickable links that retrieve relevant SPARC data resources from Discover via the SKG (SPARC Knowledge Graph).
Figure 5
Figure 5
An Example of mapping from material space to physical space. (A,B) The heart scaffold defined with 282 tri-cubic Hermite elements. (C) A single element defined with (ξ1, ξ2, ξ3) coordinates. (D) The element in (C) mapped into 3D physical space where the colors illustrate a continuous scalar field defined over that space. In addition, (E) an example of a domain refinement of the deformed cube into subdomains or “elements” is illustrated.
Figure 6
Figure 6
Scaffolds for the colon in the human (dataset available from https://sparc.science/datasets/95), pig (dataset available from https://sparc.science/datasets/98), and mouse (dataset available from https://sparc.science/datasets/76). The scaffolds for each species are built in three steps: (A–C) define a sub-scaffold that captures the cross-sectional anatomy of the colon for the three species, (D–F) define the centerline of the colon, (G–I) attach these sub-scaffolds sequentially to the centerline to form the final scaffold.
Figure 7
Figure 7
Front (left) and back (right) views of the (A) human (dataset available from https://sparc.science/datasets/100), (B) pig (dataset available from https://sparc.science/datasets/102), and (C) rat (dataset available from https://sparc.science/datasets/99) scaffolds (not to scale). The topology of the scaffold is different for each species as each has a different number of pulmonary veins entering the left atrium (4, 2, and 3, respectively).
Figure 8
Figure 8
Generating a whole rat body anatomical scaffold from data for integration with other organ scaffolds. (A) The 3D body coordinate system (dataset available from https://sparc.science/datasets/112) shows distinct anatomical regions with different colors. (B) The rat model (NeuroRat, IT'IS Foundation. 10.13099/VIP91106-04-0) in the vtk format with spinal cord and diaphragm visible. This model contains 179 segmented tissues with neuro-functionalized nerve trajectories (i.e., associated electrophysiological fiber models). (C) All tissue tissues from the rat body were resampled and converted into a data cloud to provide the means for generating the whole body scaffold. The data cloud is a 3D spatial representation of the tissue surface consisting of a certain density of points in the 3D Euclidean space. These points are used to define the objective function for the scaffold fitting procedure. The data cloud of the skin, inner core, spinal cord, and diaphragm are shown in (C) with the limbs and tail excluded. (D) The 3D body scaffold (A) was fitted to the rat data to generate the anatomical scaffold. (E) A generic rat heart scaffold was projected into the body scaffold using three corresponding fiducial landmarks (green arrows in heart and green spheres in the body). This transformation allows embedding of the organ scaffolds into their appropriate locations in the body scaffold, providing the required physical environment required for simulation and modeling. SCV, superior vena cava; RPV, right pulmonary vein; MPV, middle pulmonary vein; LPV, left pulmonary vein; RAA, right atrial appendage; RA, right atrium; LA, left atrium; LAA, left atrial appendage; PTo, pulmonary trunk outlet; RV, right ventricle; LV, left ventricle.
Figure 9
Figure 9
Mapping individual rat ICNs and cardiac anatomy onto a generalized 3D scaffold for comparison across animals through an automated pipeline. (A) Segmentation contours from the imaged heart is used to (B) customize and generate the atrial scaffold; (C) individual data points from the contours are projected on the nearest scaffold surface to compute the minimum distance for the fitting process to (D) fit and morph the scaffold to match the data; (E) the atrial contour data with ICN cells are processed to (F) map and register the spatial distribution of individual cells on the scaffold as material points. This material point mapping provides a powerful way to (G) capture information from multiple subjects (e.g., orange vs. yellow spheres showing individual ICNs from two different rats) on one “generic” scaffold. Here the heart is visualized from a superior angle to appreciate how the neuron locations in the original data are projected onto the generic scaffold. Figure adapted and modified from Leung et al. (2020).
Figure 10
Figure 10
Biophysical modeling and analysis workflow implemented at the end of the first o2S2PARC development year as a demonstrator of the feasibility of an open, readily extensible, online-based platform for collaborative computational neurosciences. (purple: published on the portal, blue: o2S2PARC services, yellow: 3rd party viewers embedded in o2S2PARC) Electromagnetic exposure of the vagus nerve by a bioelectronic implant is simulated, along with the resulting vagus nerve stimulation and its impact on cardiovascular activity: A “NEUROCOUPLE” service injects the male NEUROMAN anatomical model, along with a multi-fascicular cervical vagus nerve model layer and integrated nerve fiber trajectories, into a “Modeler” service, where interactive constructive geometry is used to integrate an implant electrode geometry. The different NEUROMAN anatomical regions are already tagged with tissue names and the fiber trajectories with electrophysiological information (fiber type and diameter). The implant-enhanced anatomical model serves as input to an “EM Simulator” service, which obtains dielectric properties from a “Tissue Properties” service and assigns them to regions based on their tags. After defining boundary conditions, discretization, and solver parameters, the high-performance computing enabled electro-quasistatic solver is called. The resulting electric potential is injected into a “NEURON Simulator” service, which also receives the tagged fiber trajectories from the “NEUROCOUPLE” service as input and discretizes these trajectories into compartment with associated ion channel distributions (according to predefined, diameter-parameterized fiber models), produces input files for the neuronal dynamics simulator from NEURON, and executes the simulations using NEURON's “extracellular potential” mechanism to consider the electric field exposure. A “Python Runner” service implements the computation of stimulation selectivity indices, which feed a “Kember Model” service that contains an implementation of the multi-scale cardiac regulation model from Kember et al. (2017). A “Paraview 3D Viewer” service and a “Rawgraphs Viewer” service provide result visualization. Finally, a “Kember Viewer” service performs predefined post-processing analysis on the output of the “Kember Model” service and visualizes the results. The “Kember Viewer” service is a specialized instance of the “Jupyter Notebook” service, which permits to present a (Python) script along with documentation in an interactively explorable and editable form. Selected components of this study (anatomical model, 3D visualization of EM field and neural activity, Kember cardiac regulation model, example JupyterLab service) can be found on https://sparc.science/data?type=simulation.
See this image and copyright information in PMC

References

    1. Bandrowski A., Grethe J. S., Pilko A., Gillespie T. H., Pine G., Patel B., et al. (2021). SPARC data structure: rationale and design of a fair standard for biomedical research data. bioRxiv [Preprint]. 10.1101/2021.02.10.430563 - DOI
    1. Bradley C., Pullan A., Hunter P. (1997). Geometric modeling of the human torso using cubic hermite elements. Ann. Biomed. Eng. 25, 96–111. 10.1007/BF02738542 - DOI - PubMed
    1. Bug W. J., Ascoli G. A., Grethe J. S., Gupta A., Fennema-Notestine C., Laird A. R., et al. (2008). The nifstd and birnlex vocabularies: building comprehensive ontologies for neuroscience. Neuroinformatics 6, 175–194. 10.1007/s12021-008-9032-z - DOI - PMC - PubMed
    1. Chan A.-W., Song F., Vickers A., Jefferson T., Dickersin K., Gótzsche P. C., et al. (2014). Increasing value and reducing waste: addressing inaccessible research. Lancet 383, 257–266. 10.1016/S0140-6736(13)62296-5 - DOI - PMC - PubMed
    1. Cuellar A. A., Lloyd C. M., Nielsen P. F., Bullivant D. P., Nickerson D. P., Hunter P. J. (2003). An overview of cellml 1.1, a biological model description language. Simulation 79, 740–747. 10.1177/0037549703040939 - DOI