Skip to main content

Table 1 Comparison of direct application of k-means and SIMLR clustering on raw gene expression data with autoencoder-based k-means and SIMLR ensemble

From: Autoencoder-based cluster ensembles for single-cell RNA-seq data analysis

 RawAutoencoder
 ARINMIFMJaccardARINMIFMJaccard
k-means     
Mouse neurons0.22 ±0.020.36 ±0.030.39 ±0.020.22 ±0.010.38 ±0.020.56 ±0.010.53 ±0.020.34 ±0.02
Mouse striatum0.36 ±0.070.69 ±0.060.51 ±0.060.31 ±0.050.45 ±0.010.75 ±0.010.58 ±0.010.37 ±0.01
Human archived brain0.29 ±0.010.49 ±0.010.35 ±0.010.21 ±0.010.37 ±0.010.56 ±00.43 ±0.010.27 ±0.01
Mouse archived brain0.32 ±0.010.49 ±00.35 ±0.010.21 ±0.010.43 ±0.010.58 ±00.46 ±0.010.3 ±0.01
SIMLR     
Mouse neurons0.44 ±00.65 ±00.58 ±00.39 ±00.71 ±00.7 ±00.81 ±00.67 ±0
Mouse striatum0.55 ±00.81 ±00.67 ±00.34 ±00.8 ±0.020.87 ±0.010.85 ±0.020.74 ±0.03
  1. Cell type identification accuracy were quantified by the four evaluation metrics