Skip to main content

Table 1 Runtimes for different normalization pipelines

From: Analytic Pearson residuals for normalization of single-cell RNA-seq UMI data


Sqrt(CPMedian), PCA

Pearson residuals, PCA


33k PBMC

31 s

35 s

15 h

33k PBMC, 1000 HVGs

3 s

4 s

4 h

2M organogenesis, 2000 HVGs

166 s

224 s

52 h

  1. The datasets are the full 33k PBMC dataset, the PBMC dataset after selecting 1, 000 HVGs, and the organogenesis dataset [39] after selecting 2, 000 HVGs. Genes with largest Pearson residual variances were selected, which took 9 s (PBMC) and 15 min (organogenesis), respectively. See “Methods” for details. All runtimes measured on a machine with 256 Gb RAM and 30 CPU threads at 2.1 GHz