Fang Zhou, Krusen Kelsey, Priest Hannah, Wang Mingshuang, Kim Sungwoong, Sriram Anirudh, Yellanki Ashritha, Singh Ankur, Horwitz Edwin, Coskun Ahmet F
Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA, USA.
Machine Learning Graduate Program, Georgia Institute of Technology, Atlanta, GA, USA.
BME Front. 2025 Mar 13;6:0110. doi: 10.34133/bmef.0110. eCollection 2025.
We developed 3-dimensional spatially resolved gene neighborhood network embedding (3D-spaGNN-E) to find subcellular gene proximity relationships and identify key subcellular motifs in cell-cell communication (CCC). The pipeline combines 3D imaging-based spatial transcriptomics and graph-based deep learning to identify subcellular motifs. Advancements in imaging and experimental technology allow the study of 3D spatially resolved transcriptomics and capture better spatial context than approximating the samples as 2D. However, the third spatial dimension increases the data complexity and requires new analyses. 3D-spaGNN-E detects single transcripts in 3D cell culture samples and identifies subcellular gene proximity relationships. Then, a graph autoencoder projects the gene proximity relationships into a latent space. We then applied explainability analysis to identify subcellular CCC motifs. We first applied the pipeline to mesenchymal stem cells (MSCs) cultured in hydrogel. After clustering the cells based on the RNA count, we identified cells belonging to the same cluster as homotypic and those belonging to different clusters as heterotypic. We identified changes in local gene proximity near the border between homotypic and heterotypic cells. When applying the pipeline to the MSC-peripheral blood mononuclear cell (PBMC) coculture system, we identified CD4 and CD8 T cells. Local gene proximity and autoencoder embedding changes can distinguish strong and weak suppression of different immune cells. Lastly, we compared astrocyte-neuron CCC in mouse hypothalamus and cortex by analyzing 3D multiplexed-error-robust fluorescence in situ hybridization (MERFISH) data and identified regional gene proximity differences. 3D-spaGNN-E distinguished distinct CCCs in cell culture and tissue by examining subcellular motifs.
我们开发了三维空间分辨基因邻域网络嵌入（3D-spaGNN-E）来寻找亚细胞基因邻近关系，并识别细胞间通讯（CCC）中的关键亚细胞基序。该流程结合了基于3D成像的空间转录组学和基于图的深度学习来识别亚细胞基序。成像和实验技术的进步使得对三维空间分辨转录组学的研究成为可能，并且比将样本近似为二维更能捕捉到更好的空间背景。然而，第三维空间增加了数据复杂性，需要新的分析方法。3D-spaGNN-E在3D细胞培养样本中检测单个转录本，并识别亚细胞基因邻近关系。然后，图自动编码器将基因邻近关系投影到潜在空间中。接着，我们应用可解释性分析来识别亚细胞CCC基序。我们首先将该流程应用于在水凝胶中培养的间充质干细胞（MSC）。基于RNA计数对细胞进行聚类后，我们将属于同一聚类的细胞识别为同型细胞，将属于不同聚类的细胞识别为异型细胞。我们识别了同型和异型细胞边界附近局部基因邻近性的变化。当将该流程应用于MSC-外周血单个核细胞（PBMC）共培养系统时，我们识别出了CD4和CD8 T细胞。局部基因邻近性和自动编码器嵌入变化可以区分不同免疫细胞的强弱抑制作用。最后，我们通过分析三维多路复用误差稳健荧光原位杂交（MERFISH）数据，比较了小鼠下丘脑和皮质中的星形胶质细胞-神经元CCC，并识别出区域基因邻近性差异。3D-spaGNN-E通过检查亚细胞基序，区分了细胞培养和组织中不同的CCC。
