Skip to main content

Table 2 Fisher's z-scores representing the dependence between a given clustering and a set of keywords.

From: Quantitative global studies of reactomes and metabolomes using a vectorial representation of reactions and chemical compounds

EC level GO Interpro BRENDA Prosite
  #clusters reactome EC #clusters reactome EC #clusters reactome EC #clusters reactome EC
1st (X.-.-.-) 6 223.10 48.62 6 322.88 42.53 6 118.29 65.73 6 178.76 61.90
2nd (X.X.-.-) 36 1458.93 611.98 39 2163.93 905.54 39 334.23 329.99 33 455.45 316.99
3rd (X.X.X.-) 79 2325.21 1660.40 79 3438.31 1948.35 108 1236.48 1017.59 67 783.31 600.11
- 21 299.97 - 21 468.07 - 21 93.77 - 21 276.52 -
  1. The optimal reactome clustering with 21 groups, for which no equivalent exists in EC, is also included. Some EC groupings results in clusters without any enzyme associated to keywords in that particular dataset. These are excluded from the analysis and are responsible for the small variations in the number of clusters (e.g. 33, 36, 39 for the 2nd level). For comparison, the reactome is clustered with the same (reduced) number of clusters.