sc perturbation models (Perspective)

	A	B	C	D	E	F	G	H	I	J
1					modeling objectives/ prediction types
2	Type	Name	Data	Description	1.i	1.ii	2	3	4	Code

3	Shallow classification & regression	SCATTome	targeted (qRT-PCR), cell line sensitivity	semi-ensemble of linear predictors		x				R, https://github.com/bvnlab/SCATTome
4		Augur	scRNA, scATAC, STARmap, MERFISH	measure response as the sampled cross-val AUROC from random forest trained to predict perturbation label		x				R, https://github.com/neurorestore/Augur
5		MIMOSCA	Perturb-seq	linear regression on combinations with interaction terms for covariates	x	possible	x	x		python, https://github.com/asncd/MIMOSCA
6		Beyondcell	scRNA	Comparison and matching with L1000 signatures		x	x			R
7		CellDrift								python
8	Factor modeling	DRUG-NEM	CyTOF	Nested Effect Model using drug-subpopulation effect matrix, maximize combined effect probabilities with fewest compounds			possible	x		not available
9		MUSIC	CRISPR	topic modeling decomposition of perturbed conditions		x			N/A	R, https://github.com/bm2-lab/MUSIC
10		scMAGeCK	CROP-seq	two separate models - LR for perturbations and RRA for ranking genes from the perturbed condition		possible	x			R, https://github.com/weililab/scMAGeCK
11		GSFA	CRISPR	Bayesian factorization model which uses gRNA target as prior for factor weights	possible		x	possible	N/A	R
12		scINSIGHT	scRNA	decompose expression into both condition specific and general factor matrices		x	possible			R
13	Non-linear distribution modeling	scGen	scRNA	VAE and perturbational effect modeled with latent space arithmetic	x	x				python, https://github.com/theislab/scgen
14		trVAE	scRNA	CVAE + maximum mean discrepancy to handle multiple perturbations	x	x				python, https://github.com/theislab/trVAE
15		PhEMD	CyTOF, scRNA	earth mover's distance on cluster proportions as embedded by PHATE	x	x		x		R, https://github.com/KrishnaswamyLab/phemd
16		MELD	scRNA	signal processing on the neighbors graph to measure a response per single cell		possible	x**			python, https://github.com/KrishnaswamyLab/MELD
17		PopAlign	scRNA	alignment of subpopulations represented as Gaussian probability distributions in oNMF latent space		x	x**			python, https://github.com/thomsonlab/popalign
18		contrastiveVI	scRNA	multi-encoder AE to prioritize target source of variation	x		possible			python
19		CellOT	4i	optimal transport	x	possible				python
20	Network models	CellOracle*	scRNA, scATAC	signal propagation through inferred gene regulatory networks	x	possible	x		N/A	python, https://github.com/morris-lab/CellOracle
21	* for genetic perturbations
22	** relies on classical differential expression methods
23			Note that this field indicates data modalities used in the paper, and that most tools are not technically limited to these modalities.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100