Skip to main content
. 2020 Oct 6;11:786. doi: 10.3389/fgene.2020.00786

Figure 4.

Figure 4

Embedding results on 1.3 million mouse brain cells dataset. (A) NeuralEE-SO with batch size as 5,000. (B) FIt-SNE. (C) UMAP. Labels represent different clusters of Louvain community detection algorithm (Blondel et al., 2008). As NeuralEE belongs to parametric methods, when applied on million-size data, it could return several outliers on the embedded space, we manually delete them to make layout tighter. The raw embedding is showed at Supplementary Figure 7.