SCMAG: A Semisupervised Single-Cell Clustering Method Based on Matrix Aggregation Graph Convolutional Neural Network

<div>The workflow of the SCMAG. The input is a gene expression matrix; the algorithm includes four steps: (1) the similarity matrix is calculated by the cosine similarity formula; (2) the incidence matrix is judged by the threshold; (3) the consensus matrix is constructed by the matrix aggregation method; (4) the consensus matrix is saved as a graph; (5) lastly, the graph is used as input to the GCN classifier for training.</div>

Computational and Mathematical Methods in Medicine

fig1

Figure 1

Figure 1: SCMAG: A Semisupervised Single-Cell Clustering Method Based on Matrix Aggregation Graph Convolutional Neural Network 