Vascular cells are present throughout the human body and contribute to risk of multiple diseases.
Vascular dysfunction directly affects risk for arterial diseases (e.g. coronary artery disease and stroke) as well
as manifestations of other diseases such as dementia, cancer, and diabetes. Single cell analysis of the human
vasculature has already begun to identify the basic mechanisms of vascular dysfunction in the large number of
associated diseases. Our group, and several other labs, have used single cell RNA-sequencing (scRNA-seq) to
identify vascular cell heterogeneity. We performed scRNA-seq of the aorta to identify functionally distinct
endothelial cell (EC) subpopulations, and multiple groups have identified activated myofibroblasts in diseased
mouse and human vascular tissue. These studies prove heterogenous cell populations exist in the arterial wall,
but it remains undetermined which populations play a causal role in early vascular dysfunction and disease risk.
The Human BioMolecular Atlas Program (HuBMAP) provides a rich source of data to begin to establish
a causal link for specific vascular cell subpopulations with disease. In HuBMAP data, ECs and vascular smooth
muscle cells (VSMCs) comprise a large portion of the single cells identified from each organ. However, to
establish the cell types and transcriptional pathways associated with disease it will be necessary to incorporate
the new datasets and computational methods we propose in this application. We aim to use new computational
methods to integrate data from diseased vascular tissue with normal HuBMAP data, to identify the disease-
relevant features of vascular cells. New methods to integrate disease associated genes from GWAS will also
help investigators prioritize causal cells for multiple common diseases. To achieve this, we will: 1) Use new
software to identify organotypic features of vascular cells in HUBMAP reference data; 2) Identify disease-specific
vascular cell signature by comparing HUBMAP reference data with samples from vascular disease; and 3) Build
and share a computational program to identify disease-relevant cell populations and gene modules through
integration with genetic association data. These analyses make use of existing vascular disease snRNA-seq
data from a rich collection of diseased subjects we can share with the HuBMAP. All data from vascular disease
subjects is available for open data sharing, and has been collected to include a diverse collection of subjects
with respect to sex and ancestry. Our methods and statistical software to perform this integration of multiple
single cell datasets with genetic associations will establish a generalizable methodology to rapidly discover the
disease-relevant cells and processes of the vasculature, and all other cell-types, for any diseases with genetic
risk and available GWAS. Our team is immediately ready to undertake the proposed studies and share the
software with the HuBMAP community. We have a track record of rapidly sharing single cell RNA-seq data, and
have a diverse team with expertise in vascular biology, statistical genetics, and computational biology.
