Fig. 1: Overview of the Spot2vector method. | Communications Biology

Fig. 1: Overview of the Spot2vector method.

From: Flexible integration of spatial and expression information for precise spot embedding via ZINB-based graph-enhanced autoencoder

Fig. 1

a Spot2vector first constructs a spatial graph and an expression graph based on the relative spatial location of spots and the gene expression matrix, respectively. b Spot2vector employs ZINB-based graph-enhanced autoencoder to extract low-dimensional representations of spots, using GCN and GAT as encoders and three-layer MLP as decoders. c The linear combination parameter can be further adjusted during the inference process: setting it to 0 focuses on spatial information, while setting it to 1 focuses on expression information. d The output of Spot2vector can be used for various downstream analyses.

Back to article page