Fig. 1: Overview of scNiche. | Nature Communications

Fig. 1: Overview of scNiche.

From: Identification and characterization of cell niches in tissue from spatial omics data at single-cell resolution

Fig. 1: Overview of scNiche.

a Schematic workflow of scNiche. Given the spatial omics data, scNiche first extracts the multi-view features of cells within a pre-defined neighborhood range. Subsequently, scNiche combined features from different views into a joint representation (z). The combined loss function comprising the M-GAE reconstruction loss (\({L}_{{rec}}\)), graph reconstruction loss (\({L}_{{gre}}\)), and mutual information loss (\({L}_{{mim}}\)) is applied to guide the training process of the model. scNiche finally performs the unsupervised clustering step on the learned joint representation (z) to identify cell niches. b Downstream analytical framework of scNiche for the comprehensive characterization and interpretation of identified cell niches.

Back to article page