Table 1 Zebrafish.

From: Unsupervised generative and graph representation learning for modelling cell differentiation

Cluster 1 (HSPCs)

Cluster 2 (Neutrophils)

Cluster 3 (Monocytes)

Cluster 4 (Erythrocytes)

Cluster 5 (Thrombocytes)

si:ch211-161c3.61, cad1,40, pcna40

illr41, ponzr6, npsn41, abcb942, lyz43

lgals2a1, c1qc, c1qa1, s100a10b, mafbb44,45

alas21, ba1l1, aqp1a.11, hbaa11, slc4a1a46,47, ba11, si:xx-by187g17.1

fn1b1, itga2b1,48, bmp6, thbs1b1, fhl1a, ctgfa, apln

  1. High weight genes computed using the high weight connections to the latent dimensions with the highest percentage for differentiating the corresponding cell type. Using references from scientific literature each cluster found using DiffVAE is mapped to a cell type.
Back to article page