Table 3

Predictive results obtained by using various similarity definitions for proteins

Similarity Definition
MAX
AVG
SUM
AVG_BestPairs

Compartment
MCC (Accuracy %)

PML BODY
0.189 (28.9)
0.153 (34.2)
0.129 (76.3)
-0.031 (0.0)
Nuclear Lamina
0.344 (45.5)
0.535 (63.6)
0.455 (45.5)
0.315 (61.8)
Nuclear Splicing Speckles
0.377 (35.7)
0.251 (71.4)
0.289 (33.9)
0.013 (12.5)
Chromatin
0.236 (19.7)
0.218 (16.4)
0.112 (4.9)
0.142 (8.2)
Nucleoplasm
0.272 (29.3)
0.039 (9.3)
-0.079 (4.0)
0.118 (6.7)
Nucleolus
0.367 (75.8)
0.431 (44.7)
0.214 (26.0)
0.289 (75.3)
Overall for Single-localization
0.298 (50.8)
0.271 (40.3)
0.187 (27.0)
0.141 (42.9)

Similarity Definition
SUM_BestPairs
AVG_Match
SUM_Match
MAX_Match

Compartment
MCC (Accuracy %)

PML BODY
0.242 (44.7)
0.187 (34.2)
0.253 (34.2)
0.211 (31.6)
Nuclear Lamina
0.53 (67.3)
0.586 (60.0)
0.578 (63.6)
0.344 (45.5)
Nuclear Splicing Speckles
0.438 (46.4)
0.397 (66.1)
0.607 (62.5)
0.487 (46.4)
Chromatin
0.325 (36.1)
0.467 (45.9)
0.518 (60.7)
0.263 (21.3)
Nucleoplasm
0.284 (36.0)
0.332 (32.0)
0.504 (56.0)
0.298 (32.0)
Nucleolus
0.512 (66.7)
0.615 (72.6)
0.656 (79.0)
0.407 (76.7)
Overall for Single-localization
0.388 (54.6)
0.431 (58.3)
0.519 (66.5)
0.335 (53.2)

(Based on SimLP: The GO term similarity is defined on the longest path shared by two GO terms [22].)

Lei and Dai BMC Bioinformatics 2006 7:491   doi:10.1186/1471-2105-7-491