Biomedical Engineering Reference
In-Depth Information
Tabl e 3. 3
Mean and Q
3
classification precision of isoforms from the Cuff 513 verification set by
similarity classes (56 proteins with
CASTOR
and traditional procedures selected)
Name
CASTOR
Sim.class
PHD
DSC
PRED
MUL
NNSSP
Zpred
CONS
Length
1gal
2
0.907
7
0.591
0.569
0.489
0.462
0.639
0.500
0.612
186
1gal
3
0.981
7
0.698
0.698
0.689
0.689
0.741
0.637
0.732
116
1
scu
1
0.917
7
0.768
0.743
0.727
0.677
0.719
0.661
0.752
121
1
scu
2
0.938
7
0.802
0.777
0.777
0.691
0.827
0.654
0.814
81
1
scu
3
0.967
7
0.845
0.765
0.812
0.798
0.879
0.758
0.879
149
2dln
1
0.918
7
0.616
0.547
0.726
0.698
0.575
0.561
0.643
73
2dln
3
0.962
7
0.678
0.714
0.678
0.666
0.726
0.547
0.714
84
2adm
0.962
7
0.538
0.615
0.526
0.485
0.603
0.479
0.597
169
2adm
0.962
7
0.652
0.643
0.754
0.537
0.680
0.574
0.685
216
1dpg
1
0.961
0
0.875
0.711
0.779
0.717
0.830
0.694
0.875
177
1dpg
2
0.902
0
0.730
0.633
0.642
0.659
0.665
0.594
0.698
308
1rec
1
0.812
3
0.686
0.735
0.696
0.705
0.754
0.686
0.715
102
1rec
2
0.907
3
0.783
0.771
0.819
0.759
0.867
0.650
0.843
83
Tab le
3.3
shows the classification precision of
CASTOR
for six different families
of proteins and, again, the higher accuracy of this classification algorithm com-
pared with other procedures is noted. 1Gal is a flavoprotein oxydoreductase of the
glucose constituted by 581 amino acids involved in the respiratory chain of the ener-
getic pathways of the cells evidenced in the
PDB
database in two slightly different
forms: 1Scu is an ATP-binding protein tetramer, a ligase, also known as succinyl-
CoA synthetase. Its catalytic activity is involved in metabolic processes. 2Dln is a
ligase with a protein chain of 306 residues involved in the biosynthesis of peptido-
glycans of the cell wall in
Escherichia coli
. Further, 2Adm is a methyltransferase
that counts 385 amino acids that catalyzes methylations involved in nucleic acid
binding. 1Dpg is an oxydoreductase constituted by a dimer (485 amino acids) that
enters in the early stages of the pentose phosphate pathway. Finally, 1rec (or recov-
erin) is a calcium-binding protein with 185 amino acids that belongs to the EF hand
superfamily; it serves as a calcium sensor in vision.
In Table
3.4
, the small differences in the primary structure which give rise to
alternative sequences in the database can be considered as isoforms of the basic
protein, although their length may be quite different.
For this algorithm, the variation in the precision of the classification among given
classes of isoforms is very limited and the overall value is high, within experimental
variation. The secondary structure may differ among the isoforms of a protein and
precise structures are determined.
The folding characteristics of the protein and its isoforms are reported. The
database indications of the secondary structure were obtained from the database
specified. In many cases, the protein considered in the database is different from the
isoforms reported, obtained from the
PDB
database, hence all the estimated fold-
ing characteristics obtained with
CASTOR
algorithm should be compared with the
results indicated in the databases.
Search WWH ::
Custom Search