Experiments of 5-fold Cross-Validation on five kingdoms using Artificial Neural Networks

*For ontology and other abbreviations, see here

Table 1: Confusion matrix for Animal group

Prd ⇒
Obs ⇓
nuc  mit  cyt  ext  gol  pex  end  lys  mem  N.P.  rowSUM  recall 
nuc 2487 12 148
10 9
2
6
1 46
125
2846 .874
mit 1
1158 15 1
0
0
0
0 18
5
1198 .967
cyt 4
2
1570 12
2
2
1
1
234 17
1845 .851
ext 3
0 5
3322
0
0 1
7
512
93
3943
.94.3
gol 0
0 6
0
58
0 2
0
100
1
167 .347
pex 0 1
0
0 0 66
0 0 36
0
103 .641
end 0 2
2
0
1 0 168
0
282
2
457 .368
lys 0 0
0
1
1
0 0 137 30
1
170 .806
mem 3
1 14 4
1
0
0
0
4762 35
4820 .988
colSUM 2498 1176 1760 3350
72
70
178
146 6020
279
15549 O.R. = .883
precision .996 .985 .892 .992 .806 .943 .944 .938 .791 S.C. = .982
  O.P. = .899
specificity
.999
.999
.986
.998
.999
1.00
.999
.999
.883


O.S. = .951

Table 2: Confusion matrix for Plant group

Prd ⇒
Obs ⇓
nuc  mit  cyt  ext  gol  chl  pex  end  vac  mem  N.P.  rowSUM   recall 
nuc 163 0 0
1
0
0 0 0
0 1
3
168 .970
mit 0
295 1
0
0 8
1 0 0 0
2
307 .961
cyt 0
0
440 0
0 4
0 0 1 0 2
447 .984
ext 0 0 1
113 0 0 0 0 1
0
12
127 .890
gol 0 0 0
0 34 0 0 0 0 0 1
35 .971
chl 2
4
8
0
0
1872 0
0
1
0
12
1899 .986
pex 0 0
0 0 0 2
27 0 0 0 0
29 .931
end 1
0 0
0 0 0 0 56 2
4
1
64 .875
vac 0 1
1
4
0
0 0 1
73
1
1
82 .890
mem 0 0
1
1
0
4
0 0
0
123
6
135 .911
colSUM 166 300
452 119 34 1890 28
57
78 129 40
3293 O.R. = .970
precision .982 .983 .973 .950 1.00
.990 .964 .982 .936 .953 S.C. = .988
  O.P. = .982
specificity
.999
.998
.996
.998
1.00
.987
1.00
1.00
.998
.998


O.S. = .992

Table: Confusion matrix for Fungi Group
 
Prd ⇒
Obs ⇓
nuc  mit  cyt  ext  gol  pex  end  mem  vac  N.P.  rowSUM   recall 
nuc 506 9
24
1
4
0
2
9
0 66
621 .815
mit 1
322 30
2
0 1
1
6
0 43
406 .793
cyt 2
9
365 2
1
3
1
2
1
9
395 .924
ext 0 0
2
156 0
0
0
4
1
8
171 .912
gol 0
0 0
0 43 0 0
5
0
4
52 .827
pex 1
2
1
0 0 57 0 2
0 1
64 .891
end 1
0 0
0
0
0 50
9
0 4
64 .781
mem 2
0
4
3
2
0
1
280 0
10
302 .927
vac 0 0
3 1
1
0
1 0
11 0
17 .579
colSUM 513 342 429
165 51
61 56 317
13
145
2094  O.R. = .856
precision .986 .942 .851 .945 .843 .934 .893 .883 .846 S.C = .931
  O.P. = .919
specificity
.995
.988
.962
.995
.996
.998
.997
.979
.999


O.S. = .984

Table: Confusion matrix for Gram-positive Bacteria group

Prd ⇒
Obs ⇓
cyt  wal  ext  mem  N.P.  rowSUM  recall 
cyt 914
0
2
5
9
930 .983
wal
1
16 0
0
2
19
.842
ext 1
0
226 7
18
252 .897
mem 5
0
5
307
23
340 .903
colSUM 921
16
233
319 52
1541  O.R. = .949
precision .992 1.00
.970
.962
S.C. = .966
  O.P. = .983
specificity
.989
1.00
.995
.990


O.S. = .990

Table: Confusion matrix for Gram-negative Bacteria group
 
Prd ⇒
Obs ⇓
cyt  ext  per  inn  wal  out  N.P.  rowSUM   recall 
cyt 1815 10
13
3
0 0
20
1861 .975
ext 1
225 3
0 0
0
24
253 .889
per 6 10
350 0
0 1
18
385
.909
inn 0
2
2
417 0 0 11
432 .965
wal
0 0
0 0 44 1 1
46 .957
out 1
3
2 0
0 183 8
197 .929
colSUM 1823 250
370 420 44 185 82
3174  O.R. = .956
precision .996 .900 .946 .993 1.00
.989 S.C. = .974
  O.P. = .981
specificity
.994
.991
.993
.999
1.00
.999


O.S. = .995

Table: Confusion matrix for Nair & Rost (1161 proteins in Swiss-Prot)

Prd ⇒
Obs ⇓
mit  ext  nuc  chl  cyt  end  lys  gol  pex  vac  N.P.  rowSUM  recall 
mit 188 1
0 0 0
0 0 0 0 0 1
190 .989
ext 1 319 1 0 2
0 0
0 0 0 11
334 .955
nuc 0 0
337 0 2
0 0 0 0 0 13
352 .957
chl 0
0
0 90
0
0 0 0 0 0 4
94 .957
cyt 1
0
5
1 129 0 0 0 0 0 0
136 .947
end 2
3
1
2 3
0
0 2
1
0 0
14 0.00
lys 0 0
2
0 0 0 3
2
0 0 0
7 .429
gol 0 0
0 0 0
0 0 22 0 0 0
22 1.00
pex 0 0 1 0 0
0 0 0 7
0 0
8 .875
vac 1 1 2
0 0
0 0 0 0 0
0
4 0.00
colSUM 193 324 349 93
136 0
3
26
8
0
29
1161 O.R. = .943
precision .974 .985 .966 .968 .949 -
1.00
.846 .875
-
S.C. = .975
  O.P. = .967
specificity
.995
.994
.985
.997
.993
1.00
1.00
.996
.999
1.00


O.S. = .992

Table: Confusion matrix for PSORT-B data

Prd ⇒
Obs ⇓
cyt  inn  per  out  ext  N.P.  rowSUM  recall 
cyt 236 2
3
0 1
10
252 .937
inn 0
286 3
1 0
18
308 .929
per 3
3
234 2
8
14
264 .886
out 0
1
1 365 2
9
378 .966
ext 0
1
0
0
217 23
241 .900
colSUM 239 293 241 368 228
74
1443 O.R. = .927
precision .947 .965 .915 .986 .876 S.C. = .949
  O.P. = .977
specificity
.997
.994
.994
.997
.991


O.S. = .995