Data Analysis Core

Data Analyses Core: Abstract
The Data Analysis Core (DAC) will provide the expertise to manage, model, and analyze data generated by the
Duke Tissue Mapping Center (TMC), so as to deliver senescent cell signatures and tissue maps of senescent
cells to the CODCC. This will be achieved by pragmatic and innovative execution of the mandated aims – Data
Processing, Data Analysis, Map Construction and Consortium Coordination. The Data Processing team will be
responsible for the implementation of a cloud native platform on Microsoft Azure that will process data
according to FAIR (Findable, Accessible, Interoperable and Reusable) guidelines. The team will coordinate
with the Biospecimen Core to document potential confounding variables such as race, sex, live or cadaveric
tissue origin; with the Biological Analysis Core for their expertise in optimal pipelines for processing specific
assay data, and with the Data Analysis team to ensure the data is collected in a format that is interoperable
with downstream analysis. The Data Analysis team will be responsible for the characterization of senescent
cell signatures that takes into account the heterogeneity of senescent cells and the dynamics of transitioning to
the senescent state. The team will use an iterative strategy to identify senescent cells, identify and expand
associated markers, and characterize the functional signature conditional on the biological context of the
senescent cell. The team will make use of organoids for initial characterization of the dynamic signature, using
these putative signatures to identify rare senescent cells in normal tissue (including biofluids), and refine the
putative signature by re-weighting signature elements based on the extent to which they occur in senescent
cells in normal tissue. The Map Construction team will be responsible for the development of spatial maps of
senescent cells in normal tissue using advanced computational biology methods, innovative tensor analysis
approaches and modern deep learning architectures. The team will integrate data from spatial assays
(multiplexed immunohistochemistry images, Visium spatial transcriptomics, and Cartana in-situ sequencing)
and single cell assays (combined scRNA-seq and scATAC-seq) to build spatial maps predictive of the
transcriptome, epigenome and secretome of senescent cells in normal tissue from lung, heart, muscle and
skin. The team will also develop a dashboard tool that interfaces with Azure for map visualization, and evaluate
the accuracy of these maps using cross-validation, data sets from public repositories, and maps constructed by
other TMCs. The Consortium Coordination team will be responsible for annotation of all data sets using terms
from NIH Common Data Elements Repository and OBO Foundry ontologies, creation of policies for data and
metadata capture, definition of practices for reproducible analysis including use of containers and workflow
orchestration scripts, and conversion of data, models, pipelines and tissue maps to interoperable formats for
uploading to the CODCC. The team will also lead the collaborative development, with other interested parties
from the SenNet consortium, of a Senescent Cell Ontology.

Grant Number: 5U54AG075936-05
NIH Institute/Center: NIH
Principal Investigator: Cliburn Chan