整合受体调控基因表达信息构建细胞通信网络

目的 构建细胞通信网络有助于揭示细胞间协同工作机制、生物学过程和疾病发病机理。目前基于配体-受体相互作用构建细胞通信网络的方法大多只考虑配体和受体的表达信息,忽略了受体对其调控基因的信号传递影响,导致构建的细胞通信网络可靠性较低。鉴于此,本文提出IRRG算法,旨在构建更为准确的细胞通信网络,并挖掘具有生物学意义的细胞通信模式。方法 本文提出了一种整合受体调控基因表达信息构建细胞通信网络的方法（命名为IRRG）。该方法通过随机游走方式计算受体对下游基因的影响得分,进而与配体-受体共表达量结合构建细胞通信网络。结果 使用IRRG构建了小鼠滤泡间表皮（IFE）细胞通信网络并分析了配体-受体对的生物学意义,验证了IRRG计算受体影响得分的稳定性和细胞通信网络构建的可靠性。此外,使用IRRG构建了透明细胞肾细胞癌（ccRCC）的细胞通信网络,挖掘并分析其肿瘤微环境细胞通信模式。结论 IRRG可以构建富有生物学意义并且可靠的细胞通信网络,帮助人们从细胞通信的角度更深入地了解多种生物过程。IRRG算法代码可从GitHub获取：https://github.com/NWPU-903PR/IRRG。

Inference of Cell-cell Communication Networks by Integrating Receptor-regulated Gene Expression Information

Objective In multicellular organisms, cell communication allows multiple cells to coordinate with each other and involved in various important biological processes, and the abnormal cell communication pattern have also been demonstrated to be associated with many diseases. In the past decades, the rapid development of single-cell RNA sequencing technology makes it possible to study cell communication pattern through constructing cell communication network based on inferring ligand-receptor interactions, and many methods have been proposed. However, most of them only consider the expression of ligands and receptors to quantify the cell-cell interaction strength, but ignore the effect of specific receptor to its downstream gene regulatory network which reflects whether that receptor participates in particular biological process or not. To fill this gap, we propose a novel method (named IRRG) to construct cell communication networks at cell type level.Methods Our IRRG algorithm consists of the following three main steps. (1) IRRG uses the signaling pathway database to construct the gene regulatory network for each receptor in each cell type, and then calculates receptor impact score (RIS) for each receptor based on a random walk algorithm. RIS represents the degree of receptor impacts on downstream genes during cell communication. (2) IRRG uses permutation test to identify ligand-receptor pairs with significant specificity across cell types in order to discover more biologically significant cellular communication phenomena. (3) IRRG combined ligand-receptor co-expression with RIS to calculate the strength of all ligand-receptor interactions between cell types, and then constructed cellular communication networks.Results To validate the effectiveness of our IRRG, we construct cell communication networks of mouse epidermal tissue and human kidney cancer tissue separately. For mouse interfollicular epidermis (IFE) dataset, IRRG constructs biologically meaningful cellular communication networks, reasonably infers the different roles played by cells at different levels of epidermal tissue in cooperation, and discovers ligand-receptor pairs that are closely related to epidermal physiological processes. We also verified the robustness of the IRRG calculated receptor impact score, indicating that the IRRG calculated receptor impact score can reasonably reflect the signaling of cellular communication processes. By counting the supporting literature possessed of the top ranked ligand-receptor pairs, we also indirectly verified that the cell communication network constructed by IRRG is more reliable. In addition, we identify the cellular communication patterns of tumor tissues in clear cell renal cells (ccRCC) using IRRG and analyze how tumor cells influence normal cells to complete their own growth and development or migration process through cellular communication, further understanding the tumor microenvironment.Conclusion In this paper, we propose a novel method for constructing cellular communication networks named IRRG which integrating receptor-regulated gene expression information through random walk. Case studies on two datasets show that IRRG can construct biologically significant cellular communication networks, which can help us understand the mechanism of biological processes from the perspective of cellular communication. The source code and associated datasets used in this work are publicly available at https://github.com/NWPU-903PR/IRRG.