Search

The advent of high throughput spatial transcriptomics (HST) has allowed for unprecedented characterization of spatially distinct cell communities within a tissue sample. While a wide range of computational tools exist for detecting cell communities in HST data, none allow for characterization of community connectivity, i.e., the relative similarity of cells within and between found communities – an analysis task that can elucidate cellular dynamics in important settings such as the tumor microenvironment. To address this gap, we introduce the concept of analysis of community connectivity (ACC), which entails not only labeling distinct cell communities within a tissue sample, but understanding the relative similarity of cells within and between communities. We develop a Bayesian multi-layer network model called BANYAN for integration of spatial and gene expression information to achieve ACC. We use BANYAN to implement ACC in invasive ductal carcinoma, and uncover distinct community structure relevant to the interaction of cell types within the tumor microenvironment. Next, we show how ACC can help clarify ambiguous annotations in a human white adipose tissue sample. Finally, we demonstrate BANYAN’s ability to recover community connectivity structure via a simulation study based on real sagittal mouse brain HST data.AvailabilityAn R package banyan is available at https://github.com/carter-allen/banyan.Contactchung.911@osu.eduSupplementary informationSupplementary data are available online.Author SummaryThe proliferation of spatial transcriptomics technologies have prompted the development of numerous statistical models for characterizing the makeup of a tissue sample in terms of distinct cell sub-populations. However, existing methods regard inferred sub-populations as static entities and do not offer any ability to discover the relative similarity of cells within and between communities, thereby obfuscating the true interactive nature of cells in a tissue sample. We develop BANYAN: a statistical model for implementing analysis of community connectivity (ACC), i.e., the process of inferring the similarity of cells within and between cell sub-populations. We demonstrate the utility of ACC through the analysis of a publicly available breast cancer data set, which revealed distinct community structure between tumor suppressive and invasive cancer cell sub-populations. We then showed how ACC may help elucidate ambiguous sub-population annotations in a publicly available human white adipose tissue data set. Finally, we implement a simulation study to validate BANYAN’s ability to recover true community connectivity structure in HST data.

High throughput spatial transcriptomics (HST) technologies have allowed for identification of distinct cell sub-populations in tissue samples, i.e., tissue architecture identification. However, existing methods do not allow for simultaneous analysis of multiple HST samples. Moreover, standard tissue architecture identification approaches do not provide uncertainty measures. Finally, no existing frameworks have integrated deep learning with Bayesian statistical models for HST data analyses. To address these gaps, we developed MAPLE: a hybrid deep learning and Bayesian modeling framework for detection of spatially informed cell spot sub-populations, uncertainty quantification, and inference of group effects in multi-sample HST experiments. MAPLE includes an embedded regression model to explain cell sub-population abundance in terms of available covariates such as treatment group, disease status, or tissue region. We demonstrate the capability of MAPLE to achieve accurate, comprehensive, and interpretable tissue architecture inference through four case studies that spanned a variety of organs in both humans and animal models.AvailabilityAn R package maple is available at https://github.com/carter-allen/maple.Contactchung.911@osu.eduSupplementary informationSupplementary data are available online.

Background Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) causes coronavirus disease 2019 (COVID-19) through direct lysis of infected lung epithelial cells, which releases damage-associated molecular patterns and induces a pro-inflammatory cytokine milieu causing systemic inflammation. Anti-viral and anti-inflammatory agents have shown limited therapeutic efficacy. Soluble CD24 (CD24Fc) blunts the broad inflammatory response induced by damage-associated molecular patterns via binding to extracellular high mobility group box 1 and heat shock proteins, as well as regulating the downstream Siglec10-Src homology 2 domain–containing phosphatase 1 pathway. A recent randomized phase III trial evaluating CD24Fc for patients with severe COVID-19 (SAC-COVID; NCT04317040) demonstrated encouraging clinical efficacy. Methods Using a systems analytical approach, we studied peripheral blood samples obtained from patients enrolled at a single institution in the SAC-COVID trial to discern the impact of CD24Fc treatment on immune homeostasis. We performed high dimensional spectral flow cytometry and measured the levels of a broad array of cytokines and chemokines to discern the impact of CD24Fc treatment on immune homeostasis in patients with COVID-19. Results Twenty-two patients were enrolled, and the clinical characteristics from the CD24Fc vs. placebo groups were matched. Using high-content spectral flow cytometry and network-level analysis, we found that patients with severe COVID-19 had systemic hyper-activation of multiple cellular compartments, including CD8+ T cells, CD4+ T cells, and CD56+ natural killer cells. Treatment with CD24Fc blunted this systemic inflammation, inducing a return to homeostasis in NK and T cells without compromising the anti-Spike protein antibody response. CD24Fc significantly attenuated the systemic cytokine response and diminished the cytokine coexpression and network connectivity linked with COVID-19 severity and pathogenesis. Conclusions Our data demonstrate that CD24Fc rapidly down-modulates systemic inflammation and restores immune homeostasis in SARS-CoV-2-infected individuals, supporting further development of CD24Fc as a novel therapeutic against severe COVID-19.

Spatially resolved transcriptomics provides a new way to define spatial contexts and understand biological functions in complex diseases. Although some computational frameworks can characterize spatial context via various clustering methods, the detailed spatial architectures and functional zonation often cannot be revealed and localized due to the limited capacities of associating spatial information. We present RESEPT, a deep-learning framework for characterizing and visualizing tissue architecture from spatially resolved transcriptomics. Given inputs as gene expression or RNA velocity, RESEPT learns a three-dimensional embedding with a spatial retained graph neural network from the spatial transcriptomics. The embedding is then visualized by mapping as color channels in an RGB image and segmented with a supervised convolutional neural network model. Based on a benchmark of sixteen 10x Genomics Visium spatial transcriptomics datasets on the human cortex, RESEPT infers and visualizes the tissue architecture accurately. It is noteworthy that, for the in-house AD samples, RESEPT can localize cortex layers and cell types based on a pre-defined region-or cell-type-specific genes and furthermore provide critical insights into the identification of amyloid-beta plaques in Alzheimer’s disease. Interestingly, in a glioblastoma sample analysis, RESEPT distinguishes tumor-enriched, non-tumor, and regions of neuropil with infiltrating tumor cells in support of clinical and prognostic cancer applications.

High throughput spatial transcriptomics (HST) is a rapidly emerging class of experimental technologies that allow for profiling gene expression in tissue samples at or near single-cell resolution while retaining the spatial location of each sequencing unit within the tissue sample. Through analyzing HST data, we seek to identify sub-populations within a tissue sample that reflect distinct cell types or states. Existing methods either ignore the spatial heterogeneity in gene expression profiles, fail to account for important statistical features such as skewness, or are heuristic network-based clustering methods that lack the inferential benefits of statistical modeling. To address this gap, we develop SPRUCE: a Bayesian spatial multivariate finite mixture model based on multivariate skew-normal distributions, which is capable of identifying distinct cellular sub-populations in HST data. We further implement a novel combination of Pólya–Gamma data augmentation and spatial random effects to infer spatially correlated mixture component membership probabilities without relying on approximate inference techniques. Via a simulation study, we demonstrate the detrimental inferential effects of ignoring skewness or spatial correlation in HST data. Using publicly available human brain HST data, SPRUCE outperforms existing methods in recovering expertly annotated brain layers. Finally, our application of SPRUCE to human breast cancer HST data indicates that SPRUCE can distinguish distinct cell populations within the tumor microenvironment.