Spatially visualized single-cell pathology of highly multiplexed protein profiles in health and disease.
Mayar AllamThomas HuShuangyi CaiLaxminarayanan KrishnanRobert B HughleyAhmet F CoskunPublished in: Communications biology (2021)
Deep molecular profiling of biological tissues is an indicator of health and disease. We used imaging mass cytometry (IMC) to acquire spatially resolved 20-plex protein data in tissue sections from normal and chronic tonsillitis cases. We present SpatialViz, a suite of algorithms to explore spatial relationships in multiplexed tissue images by visualizing and quantifying single-cell granularity and anatomical complexity in diverse multiplexed tissue imaging data. Single-cell and spatial maps confirmed that CD68+ cells were correlated with the enhanced Granzyme B expression and CD3+ cells exhibited enrichment of CD4+ phenotype in chronic tonsillitis. SpatialViz revealed morphological distributions of cellular organizations in distinct anatomical areas, spatially resolved single-cell associations across anatomical categories, and distance maps between the markers. Spatial topographic maps showed the unique organization of different tissue layers. The spatial reference framework generated network-based comparisons of multiplex data from healthy and diseased tonsils. SpatialViz is broadly applicable to multiplexed tissue biology.
Keyphrases
- single cell
- rna seq
- high throughput
- induced apoptosis
- healthcare
- public health
- electronic health record
- high resolution
- machine learning
- deep learning
- big data
- mental health
- gene expression
- oxidative stress
- poor prognosis
- binding protein
- cell cycle arrest
- signaling pathway
- cell proliferation
- photodynamic therapy
- data analysis
- human health
- drug induced
- mass spectrometry
- health information
- convolutional neural network
- nk cells
- social media
- living cells
- network analysis