Table 1 Top ranked cluster-specific features detected by the analysis of the latent space using COMET software.
From: Sparsely-connected autoencoder (SCA) for single cell RNAseq data mining
Latent space | Cluster | Feature 1 | Feature 2 | Feature 3 | Feature 4 | COMETsc statistics | TP | TN |
|---|---|---|---|---|---|---|---|---|
SCA TF | 1 | PAX5 | NFAT5 | RFXANK | CHD4 | 1.45E–49 | 0.589 | 0.997 |
SCA TF | 2 | CEBPA | KHSRP negation | CEBPB | CREBBP | 1.19E–46 | 0.561 | 0.997 |
SCA IS | 4 | NK signature | – | – | – | 5.75E–54 | 1.0 | 0.81 |
SCA IS | 5 | ASTHMA KEGG negation | – | – | – | 5.35E–84 | 0.970 | 0.972 |
SCA miRNA CLR | 3 | miR-191 | – | – | – | 1.01E–49 | 0.714 | 0.98 |
SCA miRNA RLE | 3 | miR-191 | – | – | – | 1.01E–49 | 0.714 | 0.98 |
SCA miRNA TMM | 3 | miR-132-3p | – | – | – | 1.08E–49 | 0.714 | 0.98 |
SCA miRNA FQ | 2 | miR-187-3p Rank 1 | – | – | – | 2.85E–60 | 0.67 | 0.953 |
SCA miRNA SUM | 2 | miR-187-3p Rank 4 | – | – | – | 3.45E–58 | 0.925 | 0.918 |
SSCA | 3 | miR-129-2-3P | – | – | – | 1.18E–49 | 0.742 | 0.98 |
SSCA | 4 | NK signature | – | – | – | 6.49E–103 | 1.0 | 0.99 |
SSCA | 5 | POU2F2 negation | – | – | – | 4.50E–41 | 0.851 | 0.832 |
vSCA TF | 1 | CHD4 | – | – | – | 9.35E–76 | 0.775 | 0.989 |
vSCA TF | 2 | CEBPA | – | – | – | 6.63E–62 | 0.829 | 0.977 |
