Skip to main content

Table 2 Clustering performance of OCAT on four individual scRNA-seq datasets: Romanov, Zeisel, Retina and PBMC 68k, benchmarked with scVI, Seurat v3 and SIMLR. The clustering performance is measured by the Normalized Mutual Information (NMI), where NMI=1 implies correctly clustering all the cells with the same cell types while NMI=0 indicates random guessing. The bolded entries are the best performance metrics for each scenario. See Additional file 2: Table S5 for additional evaluation metrics

From: One Cell At a Time (OCAT): a unified framework to integrate and analyze single-cell RNA-seq data

Dataset

ncells

ngenes

m

d

OCAT

scVI

Seurat

SIMLR

Romanov

2,881

24,341

20

30

0.6443

0.5436

0.6343

0.4234

Zeisel

3,005

19,972

50

30

0.7884

0.7126

0.6724

0.7373

Retina

19,829

13,166

60

60

0.8742

0.7572

0.7865

0.6337

PBMC 68k

68,579

1000

40

40

0.5750

0.4638

0.4899

0.5344