Accurate estimation of pathway activity in single cells for clustering and differential analysis

Table 1.

Summary of the number of differentially active pathways in the active infection group of the COVID monkey data set, by all four tested methods

Cell type Pathway status SiPSiC AUCell ssGSEA VAM Overlap
Alveolar Upregulated 31 30 34 28 24
Downregulated 12 9 8 10 4
Skipped 0 4 0 0 0
Missed 0 4 1 3
B cells Upregulated 1 0 0 0 0
Downregulated 0 0 0 0 0
Skipped 0 19 0 0 0
Missed 0 0 0 0
T cells Upregulated 1 0 1 3 0
Downregulated 0 0 0 0 0
Skipped 0 24 0 0 0
Missed 0 0 0 0
  • The overlap column shows the number of pathways found as differentially active by all methods. Skipped pathways are pathways for which AUCell could not calculate scores, and missed pathways are cases in which the relevant method failed to detect differential activity of a pathway detected by all other methods. For the analysis of all 50 hallmark pathways, see Supplemental Table S2.

This Article

  1. Genome Res. 34: 925-936

Preprint Server