Deciphering the Spatial Modular Patterns of Tissues by Integrating Spatial and Single-Cell Transcriptomic Data

Progress of single-cell RNA sequencing combined with spatial transcriptomics in tumour microenvironment and treatment of pancreatic cancer

In recent years, single-cell analyses have revealed the heterogeneity of the tumour microenvironment (TME) at the genomic, transcriptomic, and proteomic levels, further improving our understanding of the mechanisms of tumour development. Single-cell RNA sequencing (scRNA-seq) technology allow analysis of the transcriptome at the single-cell level and have unprecedented potential for exploration of the characteristics involved in tumour development and progression. These techniques allow analysis of transcript sequences at higher resolution, thereby increasing our understanding of the diversity of cells found in the tumour microenvironment and how these cells interact in complex tumour tissue. Although scRNA-seq has emerged as an important tool for studying the tumour microenvironment in recent years, it cannot be used to analyse spatial information for cells. In this regard, spatial transcriptomics (ST) approaches allow researchers to understand the functions of individual cells in complex multicellular organisms by understanding their physical location in tissue sections. In particular, in related research on tumour heterogeneity, ST is an excellent complementary approach to scRNA-seq, constituting a new method for further exploration of tumour heterogeneity, and this approach can also provide unprecedented insight into the development of treatments for pancreatic cancer (PC). In this review, based on the methods of scRNA-seq and ST analyses, research progress on the tumour microenvironment and treatment of pancreatic cancer is further explained.

A war on many fronts: cross disciplinary approaches for novel cancer treatment strategies

Cancer is a disease characterized by uncontrolled cellular growth where cancer cells take advantage of surrounding cellular populations to obtain resources and promote invasion. Carcinomas are the most common type of cancer accounting for almost 90% of cancer cases. One of the major subtypes of carcinomas are adenocarcinomas, which originate from glandular cells that line certain internal organs. Cancers such as breast, prostate, lung, pancreas, colon, esophageal, kidney are often adenocarcinomas. Current treatment strategies include surgery, chemotherapy, radiation, targeted therapy, and more recently immunotherapy. However, patients with adenocarcinomas often develop resistance or recur after the first line of treatment. Understanding how networks of tumor cells interact with each other and the tumor microenvironment is crucial to avoid recurrence, resistance, and high-dose therapy toxicities. In this review, we explore how mathematical modeling tools from different disciplines can aid in the development of effective and personalized cancer treatment strategies. Here, we describe how concepts from the disciplines of ecology and evolution, economics, and control engineering have been applied to mathematically model cancer dynamics and enhance treatment strategies.

STAMarker: determining spatial domain-specific variable genes with saliency maps in deep learning

Spatial transcriptomics characterizes gene expression profiles while retaining the information of the spatial context, providing an unprecedented opportunity to understand cellular systems. One of the essential tasks in such data analysis is to determine spatially variable genes (SVGs), which demonstrate spatial expression patterns. Existing methods only consider genes individually and fail to model the inter-dependence of genes. To this end, we present an analytic tool STAMarker for robustly determining spatial domain-specific SVGs with saliency maps in deep learning. STAMarker is a three-stage ensemble framework consisting of graph-attention autoencoders, multilayer perceptron (MLP) classifiers, and saliency map computation by the backpropagated gradient. We illustrate the effectiveness of STAMarker and compare it with serveral commonly used competing methods on various spatial transcriptomic data generated by different platforms. STAMarker considers all genes at once and is more robust when the dataset is very sparse. STAMarker could identify spatial domain-specific SVGs for characterizing spatial domains and enable in-depth analysis of the region of interest in the tissue section.

Identifying spatial domain by adapting transcriptomics with histology through contrastive learning

Recent advances in spatial transcriptomics have enabled measurements of gene expression at cell/spot resolution meanwhile retaining both the spatial information and the histology images of the tissues. Accurately identifying the spatial domains of spots is a vital step for various downstream tasks in spatial transcriptomics analysis. To remove noises in gene expression, several methods have been developed to combine histopathological images for data analysis of spatial transcriptomics. However, these methods either use the image only for the spatial relations for spots, or individually learn the embeddings of the gene expression and image without fully coupling the information. Here, we propose a novel method ConGI to accurately exploit spatial domains by adapting gene expression with histopathological images through contrastive learning. Specifically, we designed three contrastive loss functions within and between two modalities (the gene expression and image data) to learn the common representations. The learned representations are then used to cluster the spatial domains on both tumor and normal spatial transcriptomics datasets. ConGI was shown to outperform existing methods for the spatial domain identification. In addition, the learned representations have also been shown powerful for various downstream tasks, including trajectory inference, clustering, and visualization.