Table 1 Automatic cell identification methods included in this study
From: A comparison of automatic cell identification methods for single-cell RNA sequencing data
Name | Version | Language | Underlying classifier | Prior knowledge | Rejection option | Reference |
|---|---|---|---|---|---|---|
Garnett | 0.1.4 | R | Generalized linear model | Yes | Yes | [14] |
Moana | 0.1.1 | Python | SVM with linear kernel | Yes | No | [15] |
DigitalCellSorter | GitHub version: e369a34 | Python | Voting based on cell type markers | Yes | No | [16] |
SCINA | 1.1.0 | R | Bimodal distribution fitting for marker genes | Yes | No | [17] |
scVI | 0.3.0 | Python | Neural network | No | No | [18] |
Cell-BLAST | 0.1.2 | Python | Cell-to-cell similarity | No | Yes | [19] |
ACTINN | GitHub version: 563bcc1 | Python | Neural network | No | No | [20] |
LAmbDA | GitHub version: 3891d72 | Python | Random forest | No | No | [21] |
scmapcluster | 1.5.1 | R | Nearest median classifier | No | Yes | [22] |
scmapcell | 1.5.1 | R | kNN | No | Yes | [22] |
scPred | 0.0.0.9000 | R | SVM with radial kernel | No | Yes | [23] |
CHETAH | 0.99.5 | R | Correlation to training set | No | Yes | [24] |
CaSTLe | GitHub version: 258b278 | R | Random forest | No | No | [25] |
SingleR | 0.2.2 | R | Correlation to training set | No | No | [26] |
scID | 0.0.0.9000 | R | LDA | No | Yes | [27] |
singleCellNet | 0.1.0 | R | Random forest | No | No | [28] |
LDA | 0.19.2 | Python | LDA | No | No | [29] |
NMC | 0.19.2 | Python | NMC | No | No | [29] |
RF | 0.19.2 | Python | RF (50 trees) | No | No | [29] |
SVM | 0.19.2 | Python | SVM (linear kernel) | No | No | [29] |
SVMrejection | 0.19.2 | Python | SVM (linear kernel) | No | Yes | [29] |
kNN | 0.19.2 | Python | kNN (k = 9) | No | No | [29] |