We introduce cytoNet, a method to characterize multicellular topology from microscopy images. 7Accessible over the web, cytoNet quantifies the spatial relationships in cell communities using 8 principles of graph theory, and evaluates the effect of cell-cell interactions on individual cell 9 phenotypes. We demonstrate cytoNet's capabilities in two applications relevant to regenerative 10 medicine: quantifying the morphological response of endothelial cells to neurotrophic factors present 11 in the brain after injury, and characterizing cell cycle dynamics of differentiating neural progenitor cells. 12The framework introduced here can be used to study complex cell communities in a quantitative 13 manner, leading to a deeper understanding of environmental effects on cellular behavior. 14 A cell's place in its environment influences a large part of its behavior. Advances in the field of phenotypic 15 screening have yielded automated image analysis software that provide detailed phenotypic information 16 at the single-cell level (such as morphology, stain texture and stain intensity) from microscopy images in 17 a high-throughput manner 1,2 . However, current image analysis pipelines often do not account for spatial 18 and density-dependent effects on cell phenotype. Various types of cell-cell interactions including 19 juxtacrine and paracrine signaling are an integral part of biological processes that affect the behavior of 20 individual cells. The recent emergence of technologies for multiparametric mapping of protein and RNA 21 expression in individual cells while preserving the spatial structure of the tissue 3 has further highlighted 22 the need to study single-cell behavior in the context of cell communities. 23

We introduce Biowheel (https://biowheel.dibsvis.com/), a web--based award--winning data visualization tool, for exploring high--dimensional and heterogeneous biomedical data. Through interactive sorting and filtering of data, Biowheel enables researchers to quickly detect data outliers, evaluate data consistency, and discover mixed trends. Its interactive data presentation, visually--engaging design, and friendly user interface opens the door to easier, faster and better high--dimensional data interpretation for biomedical professionals with and without programming training. How to effectively visualize and explore high--dimensional data remains an active field of research in biomedicine. Recent years have witnessed a fast expansion of measuring dimensions (e.g., number of genes, samples, time points) brought by advances in high--throughput omics and sensor technologies 1-3 . Meanwhile, the degree of heterogeneity observed in biomedical data of the same type is rapidly increasing, thanks to improved resolution in measurements 4 , the awareness of tumor heterogeneity 5 , and a growing interest in personalized medicine 6 . This unprecedented scale, diversity, and heterogeneity of biomedical data calls for developments of unconventional visualization methods and novel software tools to drive data interpretation 7 . To that end, the HPN--DREAM breast cancer network inference crowd--sourced data challenge 8 in 2013 dedicated a sub--challenge to crowd--source visualization strategies for high--dimensional molecular time--course data sets in breast cancer. Here, we present Biowheel, a data visualization tool created from the winning design of the HPN--DREAM visualization sub--challenge. The idea of Biowheel was inspired in part by the aesthetics of circos 9 , the utility of heatmaps 10 , and the powerful interactivity of web--based visualization frameworks 11 . Circular heatmaps, enabling end--to--end comparison, serve as the core design in Biowheel to visualize both numeric and categorical data. Differentiating its design from other applications, Biowheel is fully interactive and drives data interpretation through interactive display, filtering and sorting of the raw data. In addition, Biowheel frees biomedical researchers from programming, and speeds up the scientific discovery process with its easy--to--learn graphical user interface. An example of visualizing high--dimensional molecular time--course data with Biowheel is shown in Figure 1A, using the main experimental breast cancer proteomics data set from the HPN--DREAM challenge. The original data set contains reverse phase protein array (RPPA) expression measurements of 45 phospho--proteins treated with 4 types of inhibitor and 8 types of stimulus at 7 post--stimulus