Table 2 Annotation performance of native FGGA-CC and GO-BP enriched FGGA-CC+ classifiers when predicting GO-CC terms for protein sequences in five model organisms.

From: Consistent prediction of GO protein localization

Organism

HP

HR

HF

FGGA-CC

FGGA-CC+

FGGA-CC

FGGA-CC+

FGGA-CC

FGGA-CC+

D. rario

72.877

72.914

71.848

72.348

68.313

68.507

A. thaliana

73.587

75.919

68.471

76.442

69.616

71.190

S. cerevisiae

66.473

67.248

83.426

83.896

70.931

71.935

D. melanogaster

70.690

72.015

73.563

74.635

69.531

71.067

M. musculus

66.592

67.043

77.840

79.002

69.952

70.943

  1. Protein sequences are characterized with the Physicochemical+ method. The average 5-fold hierarchical precision (HP), recall (HR) and F-score (HF) measures are reported. For each model organism, the best performing method according to the HP, HR and HF measures (p < 0.01; Wilcoxon test) is shown in bold.