Vol.1, No.2, 63-92 (2009) Natural Science
http://dx.doi.org/10.4236/ns.2009.12011
Copyright © 2009 SciRes. OPEN ACCESS
REVIEW
Recent advances in developing web-servers for
predicting protein attributes*
Kuo-Chen Chou1,2, Hong-Bin Shen1,2
1Gordon Life Science Institute, San Diego, California 92130, USA; kcchou@gordonlifescience.org
2Institute of Image Process & Pattern Recognition, Shanghai Jiaotong University, Shanghai, China
Received 7 August 2009; revised 25 August 2009; accepted 28 August 2009.
ABSTRACT
Recent advance in large-scale genome se-
quencing has generated a huge volume of pro-
tein sequences. In order to timely utilize the in-
formation hidden in these newly discovered
sequences, it is highly desired to develop com-
putational methods for efficiently identifying
their various attributes because the information
thus obtained will be very useful for both basic
research and drug development. Particularly, it
would be even more useful and welcome if a
user-friendly web-server could be provided for
each of these methods. In this minireview, a sy-
stematic introduction is presented to highlight
the development of these web-servers by our
group during the last three years.
Keywords: Cell-PLoc; Signal-CF; Signal-3L;
MemType-2L; EzyPred; HIVcleave; GPCR-CA;
ProtIdent; QuatIdent; FoldRate
1. INTRODUCTION
Proteomics, or “protein-based genomics”, is the large-
scale study of proteins. It was born due to the explosion
of protein sequences generated in the post genomic era
[1] as well as the necessity to understand the biological
process at the cellular or system level.
To effectively conduct studies in proteomics, it is
highly desired to develop high throughput tools by
which one can timely identify various attributes of pro-
teins in a large-scale manner.
For instance, given an uncharacterized protein se-
quence, how can we identify which subcellular location
site it resides at? Does the protein stay in a single sub-
cellular location or can it simultaneously exist in or
move between two and more subcellular locations?
Which part of the protein is its signal sequence? Is it a
membrane protein or non-membrane protein? If it is the
former, to which membrane protein type does it belong?
Is it an enzyme or non-enzyme? If the former, to which
main functional class and sub-functional class does it
belongs to? Is it a protease on non-protease? If it is the
former, to which protease type does it belong? Which
sites of the protein can be cleaved by proteases such as
HIV protease and SARS enzyme? Is it a GPCR (G-pro-
tein coupled receptor) or non-GPCR? If it is the former,
to which type of GPCR does it belongs to? What kind of
quaternary structure does it belong to? What kind of fold
pattern does it assume? How can we estimate its folding
rate? The list of questions is vast.
Although the answers to these questions can be deter-
mined by conducting various biochemical experiments,
the approach of purely doing experiments is both time-
consuming and costly. Consequently, the gap between
the number of newly discovered protein sequences and
the knowledge of their attributes is becoming increas-
ingly wide.
For instance, in 1986 the Swiss-Prot databank contained
merely 3,939 protein sequence entries (Table 1), but the num-
ber has since jumped to 428,650 according to version 57.0 of
24-Mar-2009 (www.ebi.ac.uk/swiss-prot), meaning that the
number of protein sequence entries now is more than
108 times the number from about 23 years ago. The
rapid increase in protein sequence entries is also shown
by the Figure 1, where a statistical illustration to show
the growth of the UniProtKB/ TrEMBL Protein Database
(http://www.ebi.ac.uk/unipro t/TrEMBLstats/) is given.
In order to use these newly found proteins for basic
research and drug discovery in a timely manner, it is
highly desired to bridge such a gap by developing effec-
tive computational methods to predict their 3D (three-
dimensional) structures [2,3] as well as various func-
tion-related attributes based on their sequence informa-
tion alone.
* Part of the contents in this article was presented in Shanghai Univer-
sit
y
in June of 2009.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
64
In this mini-review, we are to systematically introduce
the recent progresses in addressing the aforementioned
problems, particularly, for those prediction methods with
web-servers available.
Table 1. The growth of protein sequences in SWISS-PROT data banka.
Release Date
Number of sequence
entries
Number of amino
acids
Average length per
sequenceb
2.0
5.0
9.0
12.0
16.0
20.0
24.0
27.0
30.0
32.0
34.0
35.0
37.0
38.0
39.0
40.0
42.0
45.0
48.0
51.0
56.0
57.0
09/86
09/87
11/88
10/89
11/90
11/91
12/92
10/93
10/94
11/95
10/96
11/97
12/98
07/99
05/00
10/01
10/03
10/04
09/05
10/06
07/08
03/09
3,939
5,205
8,702
12,305
18,364
22,654
28,154
33,329
40,292
49,340
59,021
69,113
77,977
80,000
86,593
101,602
135,850
163,235
194,317
241,242
392,667
428,650
900,163
1,327,683
2,498,140
3,797,482
5,986,949
7,500,130
9,545,427
11,484,420
14,147,368
17,385,503
21,210,389
25,083,768
28,268,293
29,085,965
31,411,114
37,315,215
50,046,799
59,631,787
70,391,852
88,541,632
141,217,034
154,416,236
229
236
287
309
326
331
339
345
351
352
359
363
363
364
363
367
368
365
362
367
360
360
a. From http://www.ebi.ac.uk/swissprot/.
b. The average length per sequence is defined as the total number of amino acids divided by the total number of sequences. The
quotient is rounded to an integer.
Figure 1. A statistical illustration to show the growth of the UniProtKB/TrEMBL Protein Database
(http://www.ebi.ac.uk/uniprot/TrEMBLstats/).
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
65
2. WEB-SERVERS
Recently, a series of web-servers have been developed in
our group, as described below.
2.1. Cell-PLoc
Thought by many as the most basic structural and func-
tional unit of all living organisms, a cell is constituted by
many different components, compartments or organelles
(Figure 2), and they are specialized to perform different
tasks. For instance: cytoplasm, a jelly-like material,
takes up most of the cell volume, filling the cell and
serving as a “molecular soup” in which all of the cell’s
organelles are suspended; cell membrane functions as a
boundary layer to contain the cytoplasm, while cell wall
provides protection from physical injury; the cell nu-
cleus contains the genetic material (DNA) governing
all functions of the cell; the cytoskeleton functions as a
cell’s scaffold, organizing and maintaining the cell’s
shape, as well as anchoring organelles in place; mito-
chondrion is the “power generator” playing a critical role
in generating energy in the eukaryotic cell; and so forth.
However, most of these functions, which are critical to
the cell’s survival, are performed by the proteins in a cell
[4,5]. Divided by many different compartments or or-
ganelles usually termed as “subcellular locations” (Fig-
ure 2), a cell typically contains approximately one bil-
lion or 9
10 protein molecules each having its own lo-
cation (for a single-location protein) or locations (for a
multiple-location or multiplex protein). Therefore, one
of the fundamental goals in proteomics and cell biology
is to identify the subcellular localization of proteins and
their functions.
During the past 18 years, varieties of predictors have
been developed to address this problem (see, e.g., [6-48]
and the relevant references cited in a recent review paper
[49].
Developed recently, the Cell-PLoc [50] package con-
tains a set of six web-servers for predicting subcellular
localization of proteins in six different organisms. The
six web servers and their coverage scopes can be sum-
marized by the following formulation
, for eukaryotic proteins covering 22 sites
, for human proteins covering 14 sites
, for plant proteins covering 11 sites
, f
Euk- mPLoc
Hum-mPLoc
Plant - PLoc
Cell - PLocGpos - PLocor Gram positive proteins covering 5 sites
, for Gram negative proteins covering 8 sites
, for virus proteins covering 7 sites
Gneg - PLoc
Virus - PLoc
(1)
Nucleus
Plasma
membrane
Cytoplasm
ChloroplastCell Wall
Mitochondrion
Endoplasmic
reticulum
Cytoskeleton
Peroxisome
Lysosome
Golgi
apparatus
Centriole
Ext rac ell
Vacuole
Cyanelle
Hydrogenosome
Spindle pole
body
Endosome
Synapse
Microsome
Acrosome
Melanosome
NucleusNucleus
Plasma
membrane
Cytoplasm
ChloroplastCell Wall
Mitochondrion
Endoplasmic
reticulum
Cytoskeleton
Peroxisome
Lysosome
Golgi
apparatus
Centriole
Ext rac ell
Vacuole
Cyanelle
Hydrogenosome
Spindle pole
body
Endosome
Synapse
Microsome
Acrosome
Melanosome
Figure 2. Schematic illustration to show many different compo-
nents or organelles in a eukaryotic cell. Reproduced from [51] with
permission.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
66
where the character “m” in front of “PLoc” stands for
“multiple”, meaning that the corresponding predictor can
be used to deal with both single-location and multiple-
location proteins.
To use the web-server package, just do the following
procedures. (1) Open the webpage
http://chou.med.ha rvard.edu/b ioinf/Cell-PLoc/, and you
will see the top page of the Cell-PLoc package [50] on
your computer screen, as shown in Figure 3. (2) To pre-
dict the subcellular localization of eukaryotic proteins,
click the “Euk-mPLo c” button; to predict the subcellular
localization of human proteins, click the “Hum-mP L o c
button; to predict the subcellular localization of plant
proteins, click the “Plant-PLoc” button; and so forth. (3)
Now, you can follow the procedures (3) – (11) as de-
scribed in [50] to get the desired results for the query
proteins in the six different organisms.
To maximize the convenience for the people working
in the relevant areas, each of the six predictors in the
Cell-PLoc package has been used to identify all the pro-
tein entries in the corresponding organism (except those
annotated with “fragment” or those with less than 50
amino acids) in the Swiss-Prot database that do not have
subcellular location annotations or are annotated with
uncertain terms such as “probable”, “potential”, “likely”,
or “by similarity”. These large-scale predicted results
can be directly downloaded by clicking the Download
button after getting on the top page of each of the six
web-servers. These results can serve two purposes: one
is that they can be directly used by those who need the
information immediately; the other is to set a preceding
mark to examine the accuracy of these web-server pre-
dictors by the future experimental results.
For example, listed in Appendix A are 334 eukaryotic
proteins. Their experimental annotated subcellular loca-
tions were not available before Swiss-Prot 53.2 was re-
leased on 26-June-2007. However, according to the large-
scale predicted results by Euk-mPLoc that were submitted
for publication on November-12-2006 as Supporting Infor-
mation B in [51] and were also at the same time placed in
the downloadable file called Tab_Euk-mPLoc at
http://chou.med.harvard.edu/bioinf/euk-multi/ [50] or
http://202.12 0.37.186/bio inf/euk-multi/ [51], the pre-
dicted subcellular locations of the 334 eukaryotic pro-
teins are given in column 4 of Appendix A, where for
facilitating comparison the corresponding experimental
results available about seven months later are also listed
in column 5. From the table we can see the following: of
the 334 eukaryotic proteins, 309 are with single location
site and 25 with multiple location sites. Of the 309 single
location proteins, only 22 were incorrectly predicted; of
the 25 multiple location proteins, 2 (i.e., No.104 and
No.322) were incorrectly predicted. It is interesting to
see that the predicted result for No.104 was “Centriole;
Nucleus” while the experimental observation “Cyto-
plasm; Nucleus”, meaning only one of its two location
sites was incorrectly predicted; and that the predicted
result for No.322 was “Centriole; Cytoplasm; Nucleus”
while the experimental observation “Nucleus; Cyto-
plasm”, meaning both of its observed location sites were
correctly predicted although the site of “Centriole” was
over-predicted. Accordingly, the overall success rate for
the 334 proteins is over 93% as proved later by experi-
ments.
Cell-PLoc: A package of web-servers for predicting subcellular
localization of proteins in different organisms
Euk-mPLoc
Plant-PLoc
Gneg-PLoc
Hum- mPL o c
Gpos-PLoc
Virus-PLoc
Figure 3. A semi-screenshot to show the Cell-PLoc web-page at
(http://chou.med.harvard.edu/bioinf/Cell-PLoc/).
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
67
Although the predictors in the Cell-PLoc package [50]
are very powerful, they have the following shortcomings.
(1) In order for taking the advantage of Gene Ontology
(GO) [52] approach [49], the input for a query protein
must include its accession number. However, many pro-
teins, such as synthetic and hypothetical proteins, as well
as those newly-discovered proteins that have not been
deposited into databanks yet, do not have accession
numbers, and hence their subcellular locations cannot be
predicted via the GO approach. (2) Since the current GO
database is far from complete yet, many proteins cannot
be meaningfully formulated in a GO space even if their
accession numbers are available. (3) Although the
PseAA (pseudo amino acid) composition [18,53] or
PseAAC approach, a complement to the GO approach in
Cell-PLoc, can take into account some partial sequence
order effects, the original PseAAC [18] missed the func-
tional domain (FunD) [23] and sequential evolution
(SeqE) information [54,55]. To improve the aforemen-
tioned shortcomings, the Cell-PLoc package is currently
under developing to be a new version, the Cell-PLoc 2.0.
At this stage, some of the predictors therein, such as
Hum-mPLoc2.0[56], Plant-mPLoc [56], Gpos-mPLoc
[57], and Gneg-mPLoc [58], have been completed, as
will be briefed below.
To show the difference of Hum-mPLoc 2.0 with the
original Hum-mPLoc [44] in the Cell-PLoc package
[55], let us see the following demonstration steps.
Step 1. Open the webpage
http://www.csbio .sjtu.edu.cn /bioinf/hum-multi-2/, and
you will see its top page on your computer screen [50],
as shown in Figure 4a.
Step 2. Either type or copy and past the query protein
sequence into the input box (depicted by the box at the
center of Figure 4a). The input sequence should be in
FASTA format (http://en.wikipedia.org/wiki/Fasta_format),
as shown by clicking on the Example button right above
the input box. For example, if you use the 1st query pro-
tein sequence in the Example window, the input screen
should look like the illustration in Figure 4b.
Step 3. After clicking the Submit button, you will see
Cell membrane; Cytoplasm; Nucleus” shown on the
screen (Figure 4c) after 15 seconds or so, indicating that
the query protein is a multiplex protein that may simul-
taneously exist in the three subcellular location sites,
fully in agreement with experimental observations.
Step 4. If using the 2nd query protein sequence in the
Example window as an input, after clicking the Submit
Figure 4. A semi-screenshot to show (a) the top page of the web-server Hum-mPLoc 2.0 at http://www.csbio.sjtu.edu.cn/bioinf/hum-multi-2/,
(b) the input in FASTA format taken from the 1st query protein sequence in the Example window, (c) the output generated by clicking
the Submit button in panel b, and (d) the output generated through the similar procedure but using the input taken from the 2nd query
protein sequence in the Example window.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
68
button, you will see “Cytoplasm” shown on the screen
(Figure 4d), indicating the query protein is a sin-
gle-location protein residing at the cytoplasm compart-
ment or organelle, also fully in agreement with experi-
mental observations.
As we can see from the above steps, no accession
numbers whatsoever are needed for the input data. This
is quite different with the cases when using the original
Hum-mPLoc in [55] to conduct prediction. Furthermore,
the success rate expectancy has also been enhanced ow-
ing to taking into account the FunD and SeqE informa-
tion.
Besides the improvements mentioned above, the de-
velopments from Plant-PLoc [43] in the Cell-PLoc
package [50] to Plant-mPLoc [59], from Gpos-PLoc
[60] to Gpos-mPLoc [57], and from Gneg-PLoc [61] to
Gneg-mPLoc [58], have made it possible to deal with
the multiple-location problem for plant proteins, Gram-
positive bacterial proteins, and Gram-negative bacterial
proteins, respectively, as well.
2.2. Nuc-PLoc
The nucleus exists only in eukaryotic cells. Located at
the center of a cell like its kernel, the nucleus is the most
prominent and largest cellular organelle [5], with the
diameter from 11 to 22 micrometers (μm) and occupying
about 10% of the total volume of a typical animal cell
[62]. The life processes of a eukaryotic cell are guided
by its nucleus. In addition to the genetic material, the
cellular nucleus contains many proteins located at its
different compartments, called subnuclear locations.
Therefore, the information of protein subnuclear local-
ization is not only equally important to that of protein
subcellular localization but also possesses the sense at a
deeper level.
By fusing the SeqE approach and PseAAC approach
[63], a web-server called Nuc-PLoc was developed that
is accessible to the public via the website
http://chou. med.harvard.edu /bioinf/Nuc -PLoc/. It can be
used to identify nuclear proteins among the following
nine subnuclear locations: (1) chromatin, (2) hetero-
chromatin, (3) nuclear envelope, (4) nuclear matrix, (5)
nuclear pore complex, (6) nuclear speckle, (7) nucleolus,
(8) nucleoplasm, (9) nuclear PML body (Figure 5).
2.3. Signal-CF
Functioning as a “zip code” or “address tag” in guiding
proteins to the cellular locations where they are sup-
posed to be (Figure 6), signal peptides control the entry
of virtually all secretory proteins to the pathway, both in
eukaryotes and prokaryotes [64-66]. If the signal peptide
for a nascent protein was changed, the protein could end
in a wrong cellular location causing a variety of strange
diseases. Accordingly, knowledge of signal peptides can
be utilized to reprogram cells in a desired way for future
cell and gene therapy. However, to realize this, an indis-
pensable thing is to identify the signal peptide for a
Figure 5. Schematic drawing to show the nine subnuclear locations: (1)
chromatin, (2) heterochromatin, (3) nuclear envelope, (4) nuclear matrix, (5)
nuclear pore complex, (6) nuclear speckle, (7) nucleolus, (8) nucleoplasm, (9)
nuclear PML body. Adapted from [252] with permission.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
69
Nucleus
Plasma membrane
Cytoplasm
Mitochondria
Endoplasmic
reticulum
Cytoskeleton
Peroxisome
Lysosome
Golgi
apparatus
Centriole
Extracell
Microsome
Signal protein
NucleusNucleus
Plasma membrane
Cytoplasm
Mitochondria
Endoplasmic
reticulum
Cytoskeleton
Peroxisome
Lysosome
Golgi
apparatus
Centriole
Extracell
Microsome
Signal protein
Figure 6. A schematic drawing to show: how the signal peptides of secretory proteins function
as an “address tag” in directing the proteins to their proper cellular and extracellular locations.
The signal peptide sequence is colored in puple, and the mature protein sequence in blue.
Signal Peptidase
-L
s
-4 -2
-1 +1
+2
+L
m
-3
Signal Peptidase
-L
s
-4 -2
-1 +1
+2
+L
m
-3
Figure 7. A schematic drawing to show the signal sequence of a protein and how it is cleaved by the signal
peptidase. An amino acid in the signal part is depicted as a red circle with a white number to indicate its
sequential position, while that in the mature protein depicted as an open circle with a blue number. The
signal sequence contains Ls residues and the mature protein Lm residues. The cleavage site is at the position
(-1, +1), i.e., between the last residue of the signal sequence and the first residue of the mature protein.
nascent protein. Many efforts have been made in this
regards (see, e.g., [67-76] as well as the relevant refer-
ences listed in a review article [77]).
The signal peptide of a secretory protein is usually
located at its N-terminal, and it will be cleaved off by a
signal peptidase once the protein is translocated through
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
70
a membrane (Figure 7), where the cleavage site is
commonly symbolized by (1, +1)
, namely the posi-
tion between the last residue of the signal peptide and
the first residue of the mature protein. It can also be seen
from Figure 7 that once the cleavage site is identified,
the corresponding signal peptide is automatically known;
and vice versa.
The difficulty in predicting signal peptides is that for
different secretory proteins, their signal peptides are
quite different not only in sequence components and
sequence orders but also in sequence lengths. Also,
many previous methods were lacking of considering the
coupling effects of the subsites around the cleavage sites,
as analyzed in [78].
To address the above two problems, the web-server
predictor called Signal-CF [79] was developed recently.
Its features are reflected by its name, where “C” stands
for “Coupling” and “F” for “Fusion”, meaning that Sig-
nal-CF is formed by incorporating the subsite coupling
effects along a protein sequence and by fusing the results
derived from many width-different scaled windows
through a voting system.
Signal-CF is a 2-layer predictor: the 1st-layer predic-
tion engine is to identify a query protein as secretory or
non-secretory; if it is secretory, the process will be auto-
matically continued with the 2nd-layer prediction engine
to further identify the cleavage site of its signal peptide.
The predictor is also featured by high success prediction
rates with short computational time, and hence is par-
ticularly useful for the analysis of large-scale datasets.
Signal-CF is freely accessible at
http://chou.med.h arvard.edu/b ioinf/Signal-CF/.
2.4. Signal-3L
This is a 3-layer predictor developed for identifying the
signal peptides of human, plant, animal, eukaryotic,
Gram-positive, and Gram-negative proteins. The target
of the 1st-layer is to identify a query protein as secretory
or non-secretory. If the protein is identified as secretory,
the process will be automatically continued by the 2nd-
layer prediction engine to identify the potential cleavage
sites (Figure 7) along its sequence. The 3rd-layer is to
finally determine the unique cleavage site through a
global sequence alignment operation. Signal-3L is ac-
cessible to the public as a web-server at
http://chou.med.harvard.edu/bioinf/Signal-3L/. Compared
with Signal- CF, it might take a little longer computa-
tional time but yield a little higher accuracy.
Table 2. List of examples showing that signal peptides miss-predicted by SignalP-NN and/or SignalP-HMM are corrected by Sig-
nal-3L.
Proteina Experimentally verified signal peptidea SignalP 3.0-NN SignalP 3.0-HMM Signal-3L
AAF91396.1 1-40 1-37 1-37
1-40
DKK1_HUMAN 1-31 1-22 1-28
1-31
MIME_HUMAN 1-20 1-19 1-19
1-20
NP_057466.1 1-21 1-19 1-19
1-21
NP_057663.1 1-35 1-30 1-46
1-35
NP_443122.2 1-21 1-22 1-22
1-21
NP_443164.1 1-26 1-33 1-33
1-26
Q6UXL0 1-28 1-29 1-29
1-28
STC1_HUMAN 1-17 1-21 1-18
1-17
TRLT_HUMAN 1-25 1-24 1-27
1-25
CD5L_HUMAN 1-19 1-18 1-19 1-19
EDAR_HUMAN 1-26 1-28 1-26 1-26
FZD3_HUMAN 1-22 1-17 1-22 1-22
IBP7_HUMAN 1-26 1-26
1-29 1-26
KLK3_HUMAN 1-17 1-17
1-23 1-17
NMA_HUMAN 1-20 1-20
1-26 1-20
NP_064510.1 1-22 1-22
1-23 1-22
NP_068742.1 1-24 1-24
1-25 1-24
NTRI_HUMAN 1-33 1-30 1-33
1-33
SY01_HUMAN 1-23 1-23
1-18 1-23
TIE1_HUMAN 1-21 1-21
1-22 1-21
TL19_HUMAN 1-26 1-23 1-26 1-26
TR14_HUMAN 1-38 1-36 1-38 1-38
TR19_HUMAN 1-29 1-29
1-25 1-29
XP_166856 1-17 1-17
1-20 1-17
XP_209141 1-22 1-23 1-22 1-22
a Data taken from [251]. The signal peptides experimentally verified and correctly predicted are in bold-face type colored in blue; those incorrectly
predicted in red. (For interpretation of the references to color in this table caption, the reader is referred to the web version of this paper.)
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
71
Both Signal-CF and Signal-3L can be used to refine
the results by other predictors in this area. For instance,
listed in Table 2 are the signal peptides that were miss-
predicted by SignalP-NN and/or SignalP-HMM in the
SignalP package [75] but corrected by Signal-3L.
Also, according to a recent report (see Table 1 of [80])
Signal-CF performed the best in predicting the long
signal peptides, among the following eight web-server
predictors: SignalP-NN [75], SignalP-HMM [75], Sig-
nalP-NN or SignalP-HMM [75], Phobius [81], PrediSi
[76], Signal-CF [79], Signal-3L [82], and Philius [83].
2.5. MemType-2L
Given a protein sequence, how can one identify whether
it is a membrane protein or not? If it is, which membrane
protein type it belongs to? It is important to address
these problems because they are closely relevant to the
biological function of the protein concerned and to its
interaction process with other molecules in a biological
system. Most functional units or organelles in a cell are
“enveloped” by one or more membranes, which are the
structural basis for many important biological functions.
Although the basic structure of membranes is lipid bi-
layer, many specific functions of the cell membrane are
performed by the membrane proteins (see, e.g., [4,5]).
For example, it is through membrane proteins that vari-
ous chemical messages such as nerve impulses and hor-
mone activity can be passed between cells (see, e.g.,
[84]); that cells can be attached to an extracellular matrix
in grouping cells together to form tissues; that parts of
the cytoskeleton can be attached to the cell membrane in
order to provide shape; that the metabolism process and
body’s defense mechanisms can be completed; as well as
that molecules can be transported into and out of cells by
such methods as proton pumps (see, e.g., [85-87]) and
ion pumps (see, e.g., [88,89]), channel proteins [90-92]
and carrier proteins (see, e.g., [93]).
Membrane proteins possess different types, which are
closely correlated with their functions. For instance, the
transmembrane proteins can transport molecules across
the membrane or function on both its sides, whereas
proteins functioning on only one side of the lipid bilayer
are often associated exclusively with either the lipid
monolayer or a protein domain on that side. Therefore,
information about membrane protein type can provide
useful hints for determining the function of an unchar-
acterized membrane protein. Furthermore, because of the
fluid nature of their infrastructure, membrane proteins
can move around the cell membrane so as to reach where
their function is required. Therefore, it will certainly
expedite the pace in determining the function and action
process of uncharacterized membrane proteins if we can
timely acquire the knowledge of their type. With the
avalanche of protein sequences generated in the post
genomic age and the fact that membrane proteins are
encoded by 20-35% of genes [94], it is self-evident why
it is so important to develop a sequence-based automated
method for fast and effectively addressing the two prob-
lems posed at the beginning of this Section.
Stimulated by the encouraging results in predicting
the structural classification of proteins based on their
amino acid (AA) composition or AAC [95-103], the co-
variant discriminant algorithm was introduced to identify
the types of membrane proteins also based on their AA
composition in 1999 [104]. However, the AA composi-
tion does not contain any sequence order information. To
avoid completely losing the sequence order information,
the PseAA composition or PseAAC was introduced [18].
Since then, various prediction methods have been pro-
posed in this area [53,105-118].
Recently, a user-friendly web-server predictor called
MemType-2L” was developed [54]. Compared with
the other predictors which only cover 5-6 membrane
types, MemType-2L can cover 8 membrane types (Fig-
ure 8). MemType-2L is a 2-layer predictor: the 1st layer
prediction engine is to identify a query protein as mem-
brane or non-membrane; if it is membrane, the process
will be automatically continued with the 2nd-layer pre-
diction engine to further identify its type among the fol-
lowing eight categories (Figure 8): (1) type I, (2) type II,
(3) type III, (4) type IV, (5) multipass, (6) lipid-chain-
anchored, (7) GPI-anchored, and (8) peripheral.
MemType-2L is accessible to the public via the web-
site at http://chou.med.harvard.ed u/bioinf/MemType/.
2.6. EzyPred
Nearly all known enzymes are proteins that catalyze
chemical reactions and are vitally important in the me-
tabolic process. Given a protein sequence, how can we
identify whether it is an enzyme or non-enzyme? If it is,
which main functional class it belongs to? What about
its sub functional class? These problems are closely
correlated with the biological function of an uncharac-
terized protein and its acting object and process [119].
Although their answers can be found by conducting
various biochemical experiments, it is both time-con-
suming and costly to do so solely by experimental ap-
proaches. During the last six years, a number of predic-
tors have been developed to address these problems
[53,120-125].
Recently, a top-down automated method called “Ezy-
Pred” was developed [126]. It not only covers all the six
enzyme main-functional classes [127], but also many of
their sub-functional classes (see Figure 9). EzyPred is a
3-layer predictor: the 1st layer prediction engine is for
identifying a query protein as enzyme or non-enzyme;
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
72
C
N
(5) (6) (7)
GPI
(8)
N
N'
C
C
N
C
C'
N
(1) (2) (3) (4)
C
N
(5) (6)(6) (7)
GPI
(7)
GPI
(8)
N
N'
C
C
N
C
C'
N
(1) (2) (3) (4)
Figure 8. Schematic illustration showing the 8 types of membrane proteins: (1) type I transmembrane, (2) type II,
(3) type III, (4) type IV, (5) multipass transmembrane, (6) lipid-chain-anchored membrane, (7) GPI-anchored mem-
brane, and (8) peripheral membrane. As shown in the figure, types I, II, III, and IV are all of single-pass transmem-
brane proteins; see [253] for a detailed description about their difference. Reproduced from [54] with permission.
the 2nd layer for the main functional class; and the 3rd
layer for the sub functional class. Within 90 seconds of
submitting the sequence of a query protein into its input
box, EzyPred will identify whether the query protein is
enzyme or non-enzyme and, if it is an enzyme, to which
main-functional class and sub-functional class it be-
longs.
EzyPred is accessible to the public as a web-server at
http://chou.med.ha rvard.edu/b ioinf/EzyPred/.
2.7. ProtIdent
Called by many as the biology’s version of Swiss army
knives, proteases cut long sequences of amino acids into
fragments and regulate most physiological processes.
They are vitally important in life cycle and have become
a main target for drug design (see, e.g., [2,128-134]).
The actions of proteases are exquisitely selective (see,
e.g. [135-139]), with each protease being responsible for
splitting very specific sequences of amino acids under a
preferred set of environmental conditions. According to
their catalytic mechanisms, proteases are classified the
following six types: (1) aspartic, (2) cysteine, (3) glu-
tamic, (4) metallo, (5) serine, and (6) threonine [140].
Different types of proteases have different action
mechanisms and biological processes.
Therefore, it is important for both basic research and
drug discovery to consider the following two problems.
Given the sequence of a protein, can we identify whether
it is a protease or non-protease? If it is, what protease
type does it belong to?
During the last three years, some efforts have been
made in this regard [141,142]. However, none of these
methods provided a web-server that can be easily used
by the majority of experimental and pharmaceutical sci-
entists to obtain the desired data.
Very recently, a web-server called “ProtIdent” was
developed [55] by fusing the FunD (functional domain)
and SeqE (sequential evolution) information (Figure
10a). ProtIdent is a 2-layer predictor: the 1st layer is for
identifying a query protein as protease or non-protease;
if it is a protease, the process will automatically go to the
second layer to further identify it among the six different
mechanistic types (Figure 10b).
Furthermore, a step-by-step protocol guide [143] was
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
73
provided for demonstrating how to use the ProtIdent
web-server, by which one can get the desired 2-level re-
sults for a query protein sequence in around 25 seconds.
ProtIdent is freely accessible to the public via the site
at http://www.csb io.sjtu.edu.cn/b ioinf/Protease.
2.8. GPCR-CA
One of the largest families in the human genome is the
one encoding the G-protein-coupled receptors (GPCRs),
which are cell surface receptors. Owing to their charac-
teristic transmembrane topology, GPCRs are also known
as 7-transmembrane receptors, 7TM receptors, hepta-
helical receptors, and serpentine receptors that “snake”
Figure 9. A schematic drawing to use tree branches to classify enzyme and non-enzyme as well as the six main functional classes of
enzymes and their subclasses.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
74
Fusion
RPS-Blast
CDD
……
Generate FunD
descriptor for P
Generate PsePSSM
descriptor for P
……
1(1, μ)1(1, μ)1(2, μ)1(2, μ)1(10, μ)1(10, μ)2(1,0,μ)2(1,1,μ)2(1,1,μ)2(10,49,μ)2(10,49,μ)
Final protease
type
Input protein
sequence P
OET-KNN
Swiss-
Prot
OET-KNN
PSI-
Blast
a
Base
Input a protein
sequence
ProtIdent
ensemble classifier
( = 2)
Protease
ProtIdent
ensemble classifier
( = 6)
Non-protease
Stop
Final
protease type
Training
dataset I
Training
dataset II
(a) (b)
Figure 10. A flowchart to show (a) how to fuse the FunD approach and PsePSSM approach, and (b) how the two-layer Prot-Ident
ensemble classifier works in identifying proteases and their functional types. See [55] for further explanation.
Lipid
bilayer
22 33 44
55
11 66
77
Intrace lluar
E xt r ac ellular
N
C
Figure 11. Schematic representation of a GPCR with a trademark of seven-transmembrane helices, depicted
as cylinders and connected by alternating cytoplasmic and extracellular hydrophilic loops. The 7-helix bundle
thus formed has a central pore on its extracellular surface. The red entity located in the central pore represents
a ligand messenger.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
75
across a cell membrane seven times (Figure 11). The
major role of GPCRs is to transmit signals into the cell.
GPCR-associated proteins may play at least the follow-
ing four distinct roles in receptor signaling [144-147]: (1)
directly mediate receptor signaling, as in the case of G
proteins; (2) regulate receptor signaling through control-
ling receptor localization and/or trafficking; (3) act as a
scaffold, physically linking the receptor to various ef-
fectors; (4) act as an allosteric modulator of receptor
conformation, altering receptor pharmacology and/or
other aspects of receptor function.
Much effort has been invested for studying GPCRs
by both academic institutions and pharmaceutical in-
dustries. Today, approximately one third of the world
small molecule drug markets are GPCR agonists and
antagonists.
The functions of many of GPCRs are still unknown,
and it is both time-consuming and costly to determine
their ligands and signaling pathways. Particularly, as
membrane proteins, GPCRs are very difficult to crystal-
lize and most of them will not dissolve in normal sol-
vents. Accordingly, so far very few crystal GPCR struc-
tures have been determined. Although the recently de-
veloped state-of-the-art NMR technique is a very pow-
erful in determining the 3D structures of membrane pro-
teins [87,92-94,148], it is time-consuming and costly. In
order to timely obtain the protein 3D structures for ra-
tional drug design, the approach of structural bioinfor-
matics has been often adopted (see, e.g., [84,149-153]).
Unfortunately, such an approach fails to work in most
GPCR-related cases because very few GPCRs have suf-
ficiently high sequence similarity with existing struc-
ture-known proteins, an indispensable condition for de-
veloping a reasonable starting structure via structural
bioinformatics [2,3]. Consequently, it is highly desired to
develop automated methods that can fast and effectively
identify the functional families of GPCRs according to
their sequence information because the information thus
obtained can help classifying drugs, a technique called
“evolutionary pharmacology” quite useful for drug de-
velopment.
During the last 7 years or so, a number of methods
were proposed in this regard [154-159]. Some of them
were developed for identifying the main functional classes
of GPCRs (see, e.g., [157]) and some for the sub- func-
tional classes (see, e.g., [155]). None of these methods
has provided a web-server for the public usage, and
hence their practical application value is quite limited.
(a1)
(a2)
(b1)
(b2)
Figure 12. The cellular automaton image generated according to Eqs.2-5 for (a1) the rhodopsin like family member with ac-
cession number P41595; (a2) the rhodopsin like family member with accession number P18599; (b1) the secretin like family
member with accession number O95838; and (b2) the secretin like family member with accession number Q02644. Panels
(a1) and (a2) share a quite similar texture because the protein sequences from which the cellular automaton images were de-
rived belong to a same GPCR family. And the same is true for panels (b1) and (b2).
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
76
(a)
(b)
(c)
(d)
(e)
(f)
Figure 13. The cellular automaton image generated according to Eqs.2-5 for a protein taken from (a) A-rhodopsin like
family, (b) B-secretin like family, (c) C-metabotrophic/glutamate/pheromone family; (d) D-fungal pheromone family,
(e) E-cAMP receptor family, and (f) F-Frizzled/Smoothemed family, respectively. The six panels have completely dif-
ferent textures because they represent six different GPCR family members.
Recently, a web-server predictor was developed [160]
with the name as GPCR-CA, where “CA” stands for
“Cellular Automaton” [161], meaning that the cellular
automaton images have been utilized to reveal the pattern
features hidden in piles of long and complicated protein
sequences. Cellular automata are discrete dynamical sys-
tems whose behavior is completely specified in terms of
a local relation. A cellular automaton can be thought of as
a stylized universe consisting of a regular grid of cells,
each of which is in one of a finite number of possible
states, updated synchronously in discrete time steps ac-
cording to a local, identical interaction rule [162].
The procedures of generating the cellular automaton
images for protein sequences can be briefed as follows.
As a first step, each of the 20 native amino acids in a
protein sequence is represented by a 5-digit strain ac-
cording to the binary coding as defined in [163]. Thus, a
protein consisting of N amino acids can be converted
to a sequence with 5N digits (or grids); i.e,,
12 5
g()g()g()g(), (0)
NN
ttttt (2)
where g()0 or 1
it (1, 2, , 5)
iN as defined in
[163]. Suppose the time for each updated step is con-
secutively expressed by 0, 1, 2, ,
t, we have
12 5
12 5
12 5
g(0) g(0) g(0) g(0)
g(1) g(1) g(1) g(1)
g(2) g(2) g(2) g(2)



NN
NN
NN
12 5
g()g()g() g()
 

NN
(3)
where
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
77
-1 1
-1 1
-1 1
-1 1
0, if g()0, g()0, g()0
0, if g()0, g()0, g()1
1, if g()0, g()1, g()0
0, if g()0, g()1, g()1
g(1)1,
 
 
 
 

iii
iii
iii
iii
i
tt t
tt t
tt t
tt t
t
-1 1
-1 1
-1 1
-1 1
(0, 1, , )
if g()1, g()0, g()0
0, if g()1, g()0, g()1
1, if g()1, g()1, g()0
0, if g()1, g()1, g()1
 
 
 
 
iii
iii
iii
iii
t
tt t
tt t
tt t
tt t
(4)
with the spatially periodic boundary conditions; i.e.,
05
g()g ()N
tt and 511
g()g()
Ntt (5)
Suppose: g()
it, the thi grid at t, is filled with
white color if g() 0
itand black if g() 1
it. Accord-
ingly, each row of Eq.3 corresponds to a narrow ribbon
mixed with white and black colors. Scanning these rib-
bons successively on to a screen or sheet will generate a
2D (2-dimensional) black-and-white image. It has been
observed that the image texture is basically steady after
100t. The image thus evolved is called the cellu-
lar automaton image for the protein sequence concerned.
The advantage of using the cellular automaton image to
represent the protein is that it can help us visualize some
special features hidden in its long and complex sequence
[163]. For instance, the cellular automata images for
proteins from a same GPCR family share a similar textu-
re pattern (Figure 12), while those from different GPCR
families have different texture patterns (Figure 13).
Subsequently, the gray-level co-occurrence matrix
factors extracted from the cellular automaton images
were used to represent the samples of proteins through
their pseudo amino acid composition [18,53], followed
by utilizing the augmented covariant-discriminant clas-
sifier [12,164] to operate the prediction of GPCR-CA.
GPCR-CA is a 2-layer predictor: the 1st layer predic-
tion engine is for identifying a query protein as GPCR
on non-GPCR; if it is a GPCR protein, the process will
be automatically continued with the 2nd-layer prediction
engine to further identify its type among the following
six functional classes: (1) rhodopsin-like, (2) secretin-
like, (3) metabotrophic/glutamate/pheromone; (4) fungal
pheromone, (5) cAMP receptor, and (6) Frizzled/Smoo-
themed family. GPCR-CA is freely accessible at
http://218.65.61.89:8080/bioinfo/GPCR-CA, by which
one can get the desired 2-layer results for a query protein
sequence within about 20 seconds.
2.9. HIVcleave
During the past 17 years, the following two strategies
have often been utilized to find drugs against AIDS (ac-
quired immunodeficiency syndrome). One is to target
the HIV (human immunodeficiency virus) reverse tran-
scriptase (see, e.g., [165-171]); the other is to design
HIV protease inhibitors [128,136,138,139,172-174].
Functioning as a dimer, the HIV protease is made up
of two identical subunits, each having 99 residues, but
with only one active site [136,174]. The essential func-
tion of HIV protease is to cleave the precursor polypro-
tens; loss of the cleavage-ability will stop the life cycle
of infectious HIV, the culprit [175,176] of AIDS.
To find the effective inhibitors against HIV protease,
it is very helpful to understand the mechanism of how it
cleaves the polyproteins and utilize the “distorted key”
theory [136] to approach the problem, as illustrated be-
low. HIV protease is a member of the aspartyl proteases
that is highly substrate-selective and cleavage-specific.
The HIV protease-susceptible sites in a given protein
extend to an octapeptide region [177], with its amino
acid residues sequentially symbolized by eight
subsites 4
R, 3
R, 2
R, 1
R, 1'
R, 2'
R, 3'
R, 4'
R
[178], as shown in Figure 14. The scissile bond is lo-
cated between the subsites 1
R and 1'
R. Occasionally,
the susceptible sites in some proteins may contain one
subsite less or one subsite more, corresponding to the
case of a heptapeptide or nonapeptide, respectively.
However, in investigating the cleavability of peptide
sequences by HIV proteases, heptapeptides and nona-
peptides need to be considered very rarely. This might be
the result of a compromise between the following two
factors. On one hand, according to the “rack mecha-
nism” [179], the active site of HIV protease can be lik-
ened to a “rack” during the peptide cleaving process.
Thus, it appears that the more residues that are bound to
the rack of enzyme, the more strained the peptide, and
hence the more efficient the cleavage process. On the
other hand, however, the active site of an HIV protease
can hardly accommodate more than 8 residues. Conse-
quently, for most cases, the protease-susceptible sites in
proteins are strings of octapeptides as observed [135].
Thus, according to the “lock-and-key” mechanism in
enzymology, an HIV protease-cleavable peptide must
satisfy the substrate specificity, i.e., a good fit for bind-
ing to the active site. However, such a peptide, after a
modification of its scissile bond with some chemical
procedure, will completely lose its cleavability but it can
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
78
HIV Protease
R
2
R
4
R
3
R
1
R
1
R
3
R
2
R
4
N
H
H
N
N
H
H
N
C
||
O
O
||
C
C
||
O
O
||
C
H
N
H
N
N
H
N
H
O
||
C
C
||
O
C
||
O
O
||
C
S
4
S
2
S
1
S
3
S
1
S
2
S
3
S
4
HIV Protease
R
2
R
4
R
3
R
1
R
1
R
3
R
2
R
4
N
H
H
N
N
H
H
N
C
||
O
O
||
C
C
||
O
O
||
C
H
N
H
N
N
H
N
H
O
||
C
C
||
O
C
||
O
O
||
C
S
4
S
2
S
1
S
3
S
1
S
2
S
3
S
4
Figure 14. Schematic representation of substrate bound to HIV protease based an analysis of protease-inhibitor crystal struc-
tures. The active site of enzyme is composed of eight extended “subsites”, S4, S3, S2, S1, S1’, S2’, S3’, S4’, and their counterparts
in a substrate extend to an octapeptide region, sequentially symbolized by R4, R3, R2, R1, R1’, R1’, R2’, R3’, R4’, respectively. The
scissile bond is located between the subsites R1 and R1’. Reproduced with permission from Figure 3 of K.C. Chou [136].
(a)
R
1
R
2
R
3
R
4
R
1
R
2
R
3
R
4
HIV Protease
R
1
R
2
R
3
R
4
R
1
R
2
R
3
R
4
HIV Protease
(b)
R
1
R
2
R
3
R
4
R
1
R
2
R
3
R
4
HIV Protease
R
1
R
2
R
3
R
4
R
1
R
2
R
3
R
4
HIV Protease
Figure 15. Schematic illustration to show (a) a cleavable oc-
tapeptide is chemically effectively bound to the active site of
HIV protease, and (b) although still bound to the active site,
the peptide has lost its cleavability after its scissile bond is
modified from a hybrid peptide bond [254] to a single bond by
some simple routine procedure. The eight residues of the pep-
tide is sequentially symbolized R4, R3, R2, R1, R1’, R1’, R2’, R3’,
R4’. The scissile bond is located between R1 and R1’. Adapted
from [136] with permission.
still bind to the active site of an enzyme. Actually, the
molecule thus modified can be deemed as a “distorted
key”, which can be inserted into a lock but can neither
open the lock nor be pulled out from it. That is why a
molecule modified from a cleavable peptide can sponta-
neously become a competitive inhibitor against the en-
zyme. An illustration about such a concept is given in
Figure 15, where panel (a) shows an effective binding
of a cleavable peptide to the active site of HIV protease,
while panel (b) shows that the peptide has become a
non-cleavable one after its scissile bond is modified al-
though it can still tightly bind to the active site. Such a
modified peptide, or “distorted key”, will automatically
become an inhibitor candidate of HIV protease. Even for
non-peptide inhibitors, it can also provide useful insights
about the key binding groups, hydrophobic or hydro-
philic environment, fitting conformation, et al. Accord-
ingly, in search for the potential inhibitors, a matter of
paramount importance is to discern what kind of pep-
tides can be cleaved by HIV protease and what kind
cannot be. Even if limited in the range of an octapeptide,
it is by no means easy to address the question. This is
because the number of possible octapeptides formed
from 20 amino acids runs into 88log2010
20102.56 10
.
It would be exhausting to experimentally test out such an
astronomical number of octapeptides. However, if one
could find an effective computational method for pre-
dicting the cleavage sites in proteins by HIV protease,
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
79
the pace in search for the proper inhibitors of HIV pro-
tease would be significantly expedited. Actually, during
the last decade or so, various prediction methods have
been developed in this regard [128,135,137-139,180-
186].
Recently, based on the discriminant function algo-
rithm [136], a web server called HIVcleave [187] was
established at the website
http://chou.med.harvard.edu/bioinf/HIV/. For a given
protein sequence, one can use HIVcleave to predict its
cleavage sites by HIV-1 and HIV-2 proteases, respec-
tively.
2.10. QuatIdent
As the chief actors of various biological processes in a
cell, proteins have the following four different structural
levels: primary, secondary, tertiary, and quaternary [188].
The primary structure refers to the constituent amino
acid sequence; the secondary, to the local spatial ar-
rangement of a polypeptide’s backbone without regard to
the conformations of its side chains; the tertiary, to the
three-dimensional structure of an entire polypeptide; and
the quaternary, to how many polypeptide chains (sub-
units) involved in forming a protein and the spatial ar-
rangement of its subunits. The concept of quaternary
structure is derived from the fact that many proteins are
composed of two or more subunits which associate with
each other through non-covalent interactions and, in
some cases, disulfide bonds. According to the number of
subunits aggregated together in an oligomeric complex,
protein quaternary structures can be classified into:
monomer, dimer, trimer, tetramer, pentamer, and so forth
[189]. A statistical distribution of different quaternary
structural types is shown in Figure 16, from which we
can see that the nature prefers those oligomers with even
and/or small number of subunits, fully consistent with
the findings by the previous investigators [190,191]. If
the subunits in a complex are identical, then the complex
is called homo-oligomer; otherwise hetero-oligomer. For
example, the sodium channel is formed by a monomer
[192] while the potassium channel by a homo-tetramer
[88]; the phospholamban is formed by homo-pentamer
[93,193] while the Gamma-aminobutyric acid type A
(GABAA) receptor by a hetero-pentamer [84,194]; the
M2 proton channel is formed by a homo-tetramer [87]
while hemoglobin by a hetero-tetramer [195].
Facing the explosion of newly generated protein se-
quences, we are challenged to develop an automated
method for rapidly and reliably identify the quaternary
structural attributes of uncharacterized proteins because
they are closely relevant to the functions and mecha-
nisms of proteins (see, e.g., [87,195]. Besides, the in-
formation thus obtained is very useful in screening the
candidates of proteins for their 3D structure determina-
tion. It is known that many functionally important pro-
teins exist in vivo as oligomers rather than single indi-
vidual chains. For example, hemoglobin is a hetero-
tetramer of two α chains and two β chains, and the
four chains must be aggregated into one construct to
perform its cooperative function during the oxygen-
transporting process [195]. Also, the novel allosteric
drug-inhibition mechanism for the M2 proton channel
was recently revealed by the NMR observations [87,92].
It has been found through an in-depth analysis that such
a subtle mechanism is closely correlated with a unique
packing arrangement of four transmembrane helices
from four identical protein chains [90,91,196]. For this
kind of proteins, determination of their individual chains
independently would be less interesting or should be
avoided. Therefore, developing an effective method to
predict the quaternary structural attributes of proteins
based on their sequence information alone would pro-
vide useful clues for both basic research and drug de-
velopment.
To address the challenge, the web-server predictor
called “QuatIdent” [197] was developed recently by
fusing the functional domain and sequential evolution
information. QuatIdent is a 2-layer predictor. The 1st
layer is for identifying a query protein as belonging to
which one of the following ten main quaternary struc-
tural attributes: (1) monomer, (2) dimer, (3) trimer, (4)
tetramer, (5) pentamer, (6) hexamer, (7) heptamer, (8)
octamer, (9) decamer, and (10) dodecamer. If the result
thus obtained turns out to be anything but monomer, the
process will be automatically continued to further iden-
tify it belonging to a homo-oligomer or hetero-oligomer.
QuatIdent is freely accessible to the public as a web
server via the site at
http://www.csbio.sjtu.edu.cn/bioinf/Quaternary/, by whi-
ch one can get the desired 2-level results for a query
protein sequence in around 25 seconds. And the longer
the sequence is, the more time that is needed.
2.11. PQSA-Pred
This is another web-server predictor [198] developed by
hybridizing the functional domain composition approach
and pseudo amino acid composition approach for pre-
dicting protein quaternary structural attribute based on
the sequence information alone. PQSA-Pred can be
used to predict a query protein among the following
three quaternary attributes according to its sequence in-
formation: monomer, homo-oligomer, and heterooligo-
mer. As a useful tool for crystallographic scientists in
screening for their targets, PQSA-Pred is freely accessi-
ble to the public via the website at
http://218.65.61.89:8080/bioinfo/pqsa-pred .
Besides QuatIdent [197] and PQSA-Pred [198], some
other efforts were also made in this regard [189,199,200].
However, none of these methods provide a web-server
that can be easily used by the public.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
80
Monomer
14%
Dimer 50%
Trimer
6%
Tetramer
21%
Pentamer
2%
Heptamer
5%
0.2%
Octamer 1%
Decamer 0.2%
0.3%
0.3%Dodecamer
Others
Hexamer
Monomer
14%
Dimer 50%
Trimer
6%
Tetramer
21%
Pentamer
2%
Heptamer
5%
0.2%
Octamer 1%
Decamer 0.2%
0.3%
0.3%Dodecamer
Others
Hexamer
Figure 16. A pie chart to show the statistical distribution of different quaternary structural types in the nature derived from
version 55.3 of Swiss-Prot database released 29-April-2008. Reproduced with permission from [197].
2.12. PFP-Pred
A protein can function properly only if it is folded into a
very special and individual shape or conformation, i.e.,
has the correct secondary, tertiary and quaternary struc-
ture [201]. Failure to fold into the intended 3D structure
usually produces inactive proteins or misfolded proteins
[202] that may cause cell death and tissue damage [203]
and be implicated in prion diseases such as bovine
spongiform encephalopathy (BSE, also known as “mad
cow disease”) in cattle and Creutzfeldt-Jakob disease
(CJD) in humans. All prion diseases are currently un-
treatable and are always fatal [204].
Although the X-ray crystallography is a powerful tool
in determining protein 3D structures, it usually takes
months or even years to determine the structure of a sin-
gle protein. Also, the determination might fail for those
proteins (particularly membrane proteins) that are diffi-
cult to crystallize. Although the nuclear magnetic reso-
nance (NMR) technique is very powerful in determining
membrane protein structures [87,93,94,148], it requires
expensive equipments and take equally long or even
longer time. The avalanche of protein sequences gener-
ated in the Post Genomic Age has challenged us for de-
veloping computational methods by which the structural
information can be timely extracted from sequence da-
tabases. Although the direct prediction of the 3D struc-
ture of a protein from its sequence based on the least free
energy principle [201,205] is scientifically quite sound
and some encouraging results already obtained in eluci-
dating the handedness problems and packing arrange-
ments in proteins (see, e.g., [206-211]), it is far from
successful yet for predicting its 3D structure owing to
the notorious local minimum problem except for some
very special cases or by utilizing some additional infor-
mation from experiments (see, e.g., [212,213]). Actually,
it is even not successful yet for simply predicting the
overall fold of a query protein based on its sequence
alone. For further information about protein folding,
refer to a recent review [214] and the references cited
therein. Again, although it is quite successful to predict
the 3D structure of a protein according to the homology
modeling approach [2,215] as reflected by a series of
homology-modeled proteins for drug development
[84,147,149-151,153,216-226], a hurdle exists when the
query protein does not have any structure-known ho-
mologous protein in the existing databases [3].
Facing this kind of situation, a different strategy, the
so-called taxonomic approach [227] was developed to
address the problem. According to such a strategy, pre-
dicting the 3D structure of a protein may be first con-
verted to a problem of classification; i.e., identifying
which fold pattern it belongs to. Its underpinning is
based on the assumption that the number of protein folds
is limited [228-231].
The fold pattern of a protein is one level deeper than
its structural classification [98,99,229], and hence is
more challenging and complicated for prediction.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
81
PFP-Pred [232] is one of these kinds of predictors. It
was formed by a set of basic classifiers, with each
trained in different parameter systems, such as predicted
secondary structure, hydrophobicity, van der Waals vol-
ume, polarity, polarizability, as well as different dimen-
sions of pseudo amino acid composition, that were ex-
tracted from a training dataset. The operation engine for
the constituent individual classifiers was OET-KNN
(Optimized Evidence-Theoretic K-Nearest Neighbors)
rule [32,113,233]. Their outcomes were combined thru a
weighted voting to give a final determination for classi-
fying a query protein. The recognition was to find the
true fold among the 27 possible patterns. The web-server
of PFP-Pred is available to the public via the site
http://chou.med.harvard.edu/bioinf/PFP-Pred/.
2.13. PFP-FunDSeqE
This is an improved version of PFP-Pred by combining
the functional domain information and the sequential
evolution information through a fusion ensemble classi-
fier [234], as reflected by parts of its name where
“FunD” stands for “functional domain” while “SeqE”
for “sequential evolution”. Compared with the other ex-
isting methods for predicting the protein fold patterns,
PFP-FunDSeqE can usually yield better results [234].
Its web-server is available at
http://www.csbio.sjtu.edu.cn/bioinf/PFP-FunDSeqE/.
2.14. Pred-PFR
Since each protein begins as a polypeptide translated
from a sequence of mRNA as a linear chain of amino
acids, it is interesting to study the folding rates of pro-
teins from their primary sequences. Actually, protein
chains can fold into the functional 3D structures with
quite different rates, varying from several microseconds
[235] to even an hour [236]. Since the 3D structure of a
protein is determined by its primary sequence, we can
assume the same is true for its folding rate. In view of
this, we are challenged by an interesting question: Given
a protein sequence, can we find its folding rate? Al-
though the answer can be found by conducting various
biochemical experiments, doing so is both time- con-
suming and expensive. Also, although a number of pre-
diction methods were proposed [237-242], they need the
input from the 3D structure of the protein concerned, and
hence the prediction is feasible only after its 3D struc-
ture has been determined. However, according to data
released on5-May-2009 by the RCSB Protein Data Bank
(http://www.rcsb.org/pdb), the number of proteins with
3D structure known is only about 1.34% of the number
of sequence-known proteins. Therefore, it is highly de-
sired to develop an automated method that can rapidly
and approximately predict the folding rates of proteins
according to their sequence information alone. Some ef-
forts have been made in this regard (see, e.g., [243,244]).
Since the experimentally observed folding rate for a
protein chain usually represents the “apparent folding
rate constant” [245] as denoted by f
K
, it is instructive
to unravel its relationship with the detailed rate constants,
as given below.
The apparent folding rate constant f
K
for a protein
chain is defined via the following differential equation
unfold
funfold
fold
f unfold
dP( )P()
d
dP()P()
d
t
K
t
t
tKt
t

(6)
where unfold
P()t and fold
P()t represent the concentra-
tions of its unfolded state and folded state, respectively.
Suppose the total protein concentration is 0
C, and ini-
tially only the unfolded protein is present; i.e.,
unfold 0
P()tC
and fold
P()0t when 0t. Subse-
quently, the protein system is subjected to a sudden
change in temperature, solvent, or any other factor that
causes the protein to fold. Obviously, the solution for
Eq.6 is


unfold 0f
fold 0f
P()exp
P()1exp
tC Kt
tC Kt



(7)
It can be seen from the above equation that the larger the
f
K
, the faster the folding rate will be. Given the value of
f
K
, the half-life of an unfolded protein chain can be
expressed by
1/2 f
f
ln1/ 20.693TK
K
  (8)
which can also be used to reflect the time that is needed
for a protein chain to be half folded. However, the actual
folding process is much more complicated than the one
as described by Eq.6 even if the reverse rate for the
folding system concerned can be ignored. As an illustra-
tion, let us consider the following three-state folding
mechanism
23
12
unfold interfold
PPP
k
k
 (9)
where inter
P()t represents the concentration of an in-
termediate state between the unfolded and folded states,
12
k is the rate constant for unfold
P converting to inter
P,
and 23
k the rate constant for inter
P converting to fold
P.
Thus we have the following kinetic equation
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
82
unfold
12 unfold
inter
12 unfold23 inter
fold
23 inter
dP()P()
d
dP()P() P()
d
dP( )P()
d
tkt
t
tktkt
t
tkt
t


(1 0)
To get the solution of Eq.10, let us use an intuitive dia-
gram called “directed graph” or “digraph” (Figure
17a) [245,246] to represent Eq.9. To reflect the variation
of the concentrations of the three protein states with time,
the digraph is further transformed to the phase di-
graph
[245,246] as shown in Figure 17b, where s
is an interim parameter associated with the Laplace
transform as shown in Eq.11.



unfold unfold
0
inter inter
0
fold fold
0
P() P()expd
P() P()expd
P()P()expd
s
ttst
s
ttst
s
ttst



(11)
where unfold
P
, inter
P
and fold
P
are the phase concentra-
tions of unfold
P, inter
P and fold
P, respectively. Thus, ac-
cording to the phase digraph
of Figure 17b and
using the graphic rule 4 [245,246], which is also called
the graphic rule for non-steady-state kinetics” in litera-
tures (see, e.g., [247]), we can directly write out the fol-
lowing phase concentrations:



23023 00
unfold
12 2312
231212 23
P() sk sCsk CC
s
s
ksk sk
ssk skskk


 



(12.1)
 
12012 0
inter
12 23
231212 23
P() ksC kC
s
s
ksk
ssk skskk




(12.2)
 
12 23012 230
fold
12 23
231212 23
P() kkC kkC
s
s
sk sk
ssk skskk




(12.3)
Through the above phase concentrations and using
Laplace transform table (see, e.g., [248] or any standard
mathematical tables), we can immediately obtain the
desired concentrations for unfold
P, inter
P and fold
P of
Eq.10, as given by Eq.13.
unfold 0
12 0
inter
23 12
0
fold12 230
23 12
12
23
12
23 12
P()e
P()e e
P()e e
tC
kC
tkk
C
tkkC
kk
kt
kt
kt
kt kt








(13)
Accordingly, it follows from Eq.13 that
23 12
fold1223012 23
23
12
unfold
23 1223 12
dP()ee1e P
d
k
tkkC kk
t
t
tkk kk
kt
k
k











(14)
Comparing Eq.14 with Eq.6, we obtain the following equivalent relation
23 12
12 23
f
23 12
1e k
kk
Kkk
kt




(15)
(a) (b)
unfold
Pinter
P
12
k
fold
P
23
k
unfold
Pinter
P
12
k
fold
P
23
k
unfold
Pinter
P
12
k
fold
P
23
k
12
ks
23
kss
unfold
Pinter
P
12
k
fold
P
23
k
12
ks
23
kss
Figure 17. (a) The directed graph or digraph [245,246] for the three-state protein folding mechanism as schematically ex-
pressed by Eq.9 and formulated by Eq.10. (b) The phase digraph
obtained from of panel (a) according to graphic rule 4
for enzyme and protein folding kinetics [245,246], where s is an interim parameter (see the text for further explanation).
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
83
meaning that the apparent folding rate constant f
K
is a
function of not only the detailed rate constants, but also
t. Accordingly, f
K
is actually not a constant but will
change with time. Only when 23 12
kk and 23 1k ,
can Eq.15 be reduced tof12
K
kand Eq.14 to
folded
12 unfoldunfold
dP( )P()P()
df
tktKt
t (16)
and f
K
be treated as a constant.
Even for a two-state protein folding system when the
reverse effect needs to be considered, i.e., the system
described by the following scheme and equation
12
21
unfoldfold
PP
k
k

 (17)
unfold
12unfold21 fold
fold
12unfold21fold
dP( )P()P()
d
dP()P()P()
d
tktkt
t
tktkt
t
 

(18)
where 21
k represents the reverse rate constant convert-
ing fold
P back to unfold
P. With the similar derivation by
using the non-steady state graphic rule [245,246] as de-
scribed above, we can get the following equivalent rela-
tion [249]

 
12 1221
f1221
21 121221
exp
exp
kkk
K
kkt
kk kkt









(19)
indicating that, even for the two-state folding system of
Eq.17, the apparent folding rate constant f
K
can be
treated as a constant only when 12 21
kk and
12 1k .
It can be imagined that for a general multi-state fold-
ing system, f
K
will be much more complicated. Con-
sequently, all the experimental apparent folding rate
constants were actually measured under some special
conditions.
Recently, a web-server, called “Pred-PFR” (Predict-
ing Protein Folding Rate), was developed for predicting
the folding rate of a protein [249]. The predictor is fea-
tured by fusing multiple individual predictors, each of
which is established based on one special feature derived
from the protein sequence. As a user-friendly web-server,
Pred-PFR is freely accessible to the public at
www.csbio.sjtu.edu.cn/bioinf/FoldingRate/.
2.15. FoldRate
This is a different kind of protein folding rate predictor
developed by fusing the folding-correlated features that
can be either directly obtained or easily derived from the
sequences of proteins [250]. FoldRate is freely accessible
to the public at www.csbio.sjtu.edu.cn/bioinf/FoldRate/.
Both Pred-PFR and Fold Rate can be used to predict
the folding rate of a protein according to its sequence
alone. The time by using the two web-server predictors
to get the desired result for a query protein sequence is
around 30 seconds. And the results obtained thus ob-
tained are usually at least comparable with or even better
than the existing methods that, however, need both the
sequence and 3D structure information for prediction.
3. LIST OF WEB SERVERS
For reader’s convenience, a brief description of each of
the 15 web servers introduced in this article as well as its
website address is given in Table 3.
4. CONCLUSION
Web-server is a newly emerging thing in the Internet
Age. Technically speaking, a web-server means a com-
puter program that is responsible for accepting HTTP
(Hypertext Transfer Protocol) requests from clients. By
means of web-servers, many computational prediction
methods, regardless how difficult their mathematics or
how complicated their algorithms are, can be easily used
by the vast majority of scientists without the need to
understand the mathematical details. Written as a labo-
ratory protocol with a “recipe” style, the web-servers
introduced here are user friendly and can be very easily
used. Therefore, they are particularly useful for bench
scientists to generate various data or information in a
timely manner that they may need for their research pro-
jects.
It is anticipated that all these web-servers are con-
stantly evolving with continuously improving the train-
ing datasets and prediction algorithms. To keep the users
timely informed of the development, a short note will be
published or an announcement will be placed in the
relevant website.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
84
Table 3. List of the 15 web servers introduced in this paper as well as their website addresses and targets.
No. Name Website address Target
1 Cell-PLoc package http://chou.med.harvard.edu/bioinf/Cell-PLoc/ Protein subcellular localization [49]
2 Nuc-PLoc http://chou.med.harvard.edu/bioinf/Nuc-PLoc/ Protein subnuclear localization [63]
3 Signal-CF http://chou.med.harvard.edu/bioinf/Signal-CF/ Protein signal peptide [79]
4 Signal-3L http://chou.med.harvard.edu/bioinf/Signal-3L/ Protein signal peptide [82]
5 MemType-2L http://chou.med.harvard.edu/bioinf/MemType/ Membrane protein type [54]
6 EzyPred http://chou.med.harvard.edu/bioinf/EzyPred/ Enzyme functional class [126]
7 ProtIdent http://www.csbio.sjtu.edu.cn/bioinf/Protease/ Protease type [55]
8 GPCR-CA http://218.65.61.89:8080/bioinfo/GPCR-CA GPCR type [160]
9 HIVcleave http://chou.med.harvard.edu/bioinf/HIV/ HIV protease cleavage site [187]
10 QuatIdent www.csbio.sjtu.edu.cn/bioinf/Quaternary/ Protein quaternary structural attribute [197]
11 PQSA-Pred http://218.65.61.89:8080/bioinfo/pqsa-pred Protein quaternary structural attribute [198]
12 PFP-Pred http://www.csbio.sjtu.edu.cn/bioinf/PFP-Pred/ Protein fold pattern [232]
13 PFP-FunDSeqE www.csbio.sjtu.edu.cn/bioinf/PFP-FunDSeqE/ Protein fold pattern [234]
14 Pred-PFR www.csbio.sjtu.edu.cn/bioinf/FoldingRate/ Protein folding rate [249]
15 FoldRate www.csbio.sjtu.edu.cn/bioinf/FoldRate/ Protein folding rate [250]
REFERENCES
[1] Chou, K.C. (2002) A new branch of proteomics: predic-
tion of protein cellular attributes. In Weinrer, P. W. and
Lu, Q. (eds.), Gene Cloning & Expression Technologies,
Chapter 4. Eaton Publishing, Westborough, MA, pp.
57-70.
[2] Chou, K.C. (2004) Review: Structural bioinformatics and
its impact to biomedical science. Current Medicinal
Chemistry, 11, 2105-2134.
[3] Chou, K.C. (2006) Structural bioinformatics and its im-
pact to biomedical science and drug discovery. Frontiers
in Medicinal Chemistry, 3, 455-502.
[4] Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and
Watson, J.D. (1994) Molecular Biology of the Cell,
chap.1. 3rd ed. Garland Publishing, New York & Lon-
don.
[5] Lodish, H., Baltimore, D., Berk, A., Zipursky, S.L., Ma-
tsudaira, P. and Darnell, J. (1995) Molecular Cell Biol-
ogy , Chap.3. 3rd ed. Scientific American Books, New
York.
[6] Nakai, K. and Kanehisa, M. (1991) Expert system for
predicting protein localization sites in Gram-negative
bacteria. Proteins: Structure, Function and Genetics, 11,
95-110.
[7] Nakashima, H. and Nishikawa, K. (1994) Discrimination
of intracellular and extracellular proteins using amino
acid composition and residue-pair frequencies. J Mol
Biol, 238, 54-61.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
85
[8] Cedano, J., Aloy, P., P'erez-Pons, J.A. and Querol, E.
(1997) Relation between amino acid composition and
cellular location of proteins. J Mol Biol, 266, 594-600.
[9] Nakai, K. and Horton, P. (1999) PSORT: a program for
detecting sorting signals in proteins and predicting their
subcellular localization. Trends in Biochemical Science,
24, 34-36.
[10] Chou, K.C. and Elrod, D.W. (1998) Using discriminant
function for prediction of subcellular location of pro-
karyotic proteins. BBRC, 252, 63-68.
[11] Reinhardt, A. and Hubbard, T. (1998) Using neural net-
works for prediction of the subcellular location of pro-
teins. Nucleic Acids Research, 26, 2230-2236.
[12] Chou, K.C. and Elrod, D.W. (1999) Protein subcellular
location prediction. Protein Engineering, 12, 107-118.
[13] Yuan, Z. (1999) Prediction of protein subcellular loca-
tions using Markov chain models. FEBS Letters, 451,
23-26.
[14] Nakai, K. (2000) Protein sorting signals and prediction of
subcellular localization. Advances in Protein Chemistry,
54, 277-344.
[15] Murphy, R.F., Boland, M.V. and Velliste, M. (2000) To-
wards a systematics for protein subcellular location:
quantitative description of protein localization patterns
and automated analysis of fluorescence microscope im-
ages. Proc Int Conf Intell Syst Mol Biol, 8, 251-259.
[16] Chou, K.C. (2000) Review: Prediction of protein struc-
tural classes and subcellular locations. Current Protein
and Peptide Science, 1, 171-208.
[17] Emanuelsson, O., Nielsen, H., Brunak, S. and von Heijne,
G. (2000) Predicting subcellular localization of proteins
based on their N-terminal amino acid sequence. Journal
of Molecular Biology, 300, 1005-1016.
[18] Chou, K.C. (2001) Prediction of protein cellular attrib-
utes using pseudo amino acid composition. PROTEINS:
Structure, Function, and Genetics (Erratum: ibid, 2001,
Vol44, 60), 43, 246-255.
[19] Feng, Z.P. (2001) Prediction of the subcellular location of
prokaryotic proteins based on a new representation of the
amino acid composition. Biopolymers, 58, 491-499.
[20] Hua, S. and Sun, Z. (2001) Support vector machine ap-
proach for protein subcellular localization prediction.
Bioinformatics, 17, 721-728.
[21] Feng, Z.P. and Zhang, C.T. (2001) Prediction of the sub-
cellular location of prokaryotic proteins based on the hy-
drophobicity index of amino acids. Int J Biol Macromol,
28, 255-261.
[22] Feng, Z.P. (2002) An overview on predicting the subcel-
lular location of a protein. In Silico Biol, 2, 291-303.
[23] Chou, K.C. and Cai, Y.D. (2002) Using functional do-
main composition and support vector machines for pre-
diction of protein subcellular location. J Biol Chem, 277,
45765-45769.
[24] Zhou, G.P. and Doctor, K. (2003) Subcellular location
prediction of apoptosis proteins. PROTEINS: Structure,
Function, and Genetics, 50, 44-48.
[25] Pan, Y.X., Zhang, Z.Z., Guo, Z.M., Feng, G.Y., Huang,
Z.D. and He, L. (2003) Application of pseudo amino acid
composition for predicting protein subcellular location:
stochastic signal processing approach. Journal of Protein
Chemistry, 22, 395-402.
[26] Park, K.J. and Kanehisa, M. (2003) Prediction of protein
subcellular locations by support vector machines using
compositions of amino acid and amino acid pairs. Bioin-
formatics, 19, 1656-1663.
[27] Gardy, J.L., Spencer, C., Wang, K., Ester, M., Tusnady,
G.E., Simon, I., Hua, S., deFays, K., Lambert, C., Nakai,
K. et al. (2003) PSORT-B: Improving protein subcellular
localization prediction for Gram-negative bacteria. Nu-
cleic Acids Research, 31, 3613-3617.
[28] Huang, Y. and Li, Y. (2004) Prediction of protein subcel-
lular locations using fuzzy k-NN method. Bioinf ormatic s,
20, 21-28.
[29] Xiao, X., Shao, S., Ding, Y., Huang, Z., Huang, Y. and
Chou, K.C. (2005) Using complexity measure factor to
predict protein subcellular location. Amino Acids, 28,
57-61.
[30] Gao, Y., Shao, S.H., Xiao, X., Ding, Y.S., Huang, Y.S.,
Huang, Z.D. and Chou, K.C. (2005) Using pseudo amino
acid composition to predict protein subcellular location:
approached with Lyapunov index, Bessel function, and
Chebyshev filter. Amino Acids, 28, 373-376.
[31] Lei, Z. and Dai, Y. (2005) An SVM-based system for
predicting protein subnuclear localizations. BMC Bioin-
formatics, 6, 291.
[32] Shen, H.B. and Chou, K.C. (2005) Predicting protein
subnuclear location with optimized evidence-theoretic
K-nearest classifier and pseudo amino acid composition.
Biochem Biophys Res Comm, 337, 752-756.
[33] Garg, A., Bhasin, M. and Raghava, G.P. (2005) Support
vector machine-based method for subcellular localization
of human proteins using amino acid compositions, their
order, and similarity search. J Biol Chem, 280,
14427-14432.
[34] Matsuda, S., Vert, J.P., Saigo, H., Ueda, N., Toh, H. and
Akutsu, T. (2005) A novel representation of protein se-
quences for prediction of subcellular location using sup-
port vector machines. Protein Sci, 14, 2804-2813.
[35] Gao, Q.B., Wang, Z.Z., Yan, C. and Du, Y.H. (2005)
Prediction of protein subcellular location using a com-
bined feature of sequence. FEBS Lett, 579, 3444-3448.
[36] Chou, K.C. and Shen, H.B. (2006) Predicting protein
subcellular location by fusing multiple classifiers. Jour-
nal of Cellular Biochemistry, 99, 517-527.
[37] Guo, J., Lin, Y. and Liu, X. (2006) GNBSL: A new inte-
grative system to predict the subcellular location for
Gram-negative bacteria proteins. Proteomics, 6,
5099-5105.
[38] Xiao, X., Shao, S.H., Ding, Y.S., Huang, Z.D. and Chou,
K.C. (2006) Using cellular automata images and pseudo
amino acid composition to predict protein subcellular lo-
cation. Amino Acids, 30, 49-54.
[39] Hoglund, A., Donnes, P., Blum, T., Adolph, H.W. and
Kohlbacher, O. (2006) MultiLoc: prediction of protein
subcellular localization using N-terminal targeting se-
quences, sequence motifs and amino acid composition.
Bioinformatics, 22, 1158-1165.
[40] Lee, K., Kim, D.W., Na, D., Lee, K.H. and Lee, D. (2006)
PLPD: reliable protein localization prediction from im-
balanced and overlapped datasets. Nucleic Acids Res, 34,
4655-4666.
[41] Zhang, Z.H., Wang, Z.H., Zhang, Z.R. and Wang, Y.X.
(2006) A novel method for apoptosis protein subcellular
localization prediction combining encoding based on
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
86
grouped weight and support vector machine. FEBS Lett,
580, 6169-6174.
[42] Shi, J.Y., Zhang, S.W., Pan, Q., Cheng, Y.-M. and Xie, J.
(2007) Prediction of protein subcellular localization by
support vector machines using multi-scale energy and
pseudo amino acid composition. Amino Acids, 33, 69-74.
[43] Chou, K.C. and Shen, H.B. (2007) Large-scale plant
protein subcellular location prediction. Journal of Cellu-
lar Biochemistry, 100, 665-678.
[44] Shen, H.B. and Chou, K.C. (2007) Hum-mPLoc: An
ensemble classifier for large-scale human protein sub-
cellular location prediction by incorporating samples
with multiple sites. Biochem Biophys Res Commun, 355,
1006-1011.
[45] Shen, H.B., Yang, J. and Chou, K.C. (2007) Euk-PLoc:
an ensemble classifier for large-scale eukaryotic protein
subcellular location prediction. Amino Acids, 33, 57-67.
[46] Chen, Y.L. and Li, Q.Z. (2007) Prediction of apoptosis
protein subcellular location using improved hybrid ap-
proach and pseudo amino acid composition. Journal of
Theoretical Biology, 248, 377–381.
[47] Chen, Y.L. and Li, Q.Z. (2007) Prediction of the subcel-
lular location of apoptosis proteins. Journal of Theoreti-
cal Biology, 245, 775-783.
[48] Mundra, P., Kumar, M., Kumar, K.K., Jayaraman, V.K.
and Kulkarni, B.D. (2007) Using pseudo amino acid
composition to predict protein subnuclear localization:
Approached with PSSM. Pattern Recognition Letters, 28,
1610-1615.
[49] Chou, K.C. and Shen, H.B. (2007) Review: Recent pro-
gresses in protein subcellular location prediction. Ana-
lytical Biochemistry, 370, 1-16.
[50] Chou, K.C. and Shen, H.B. (2008) Cell-PLoc: A package
of web-servers for predicting subcellular localization of
proteins in various organisms. Nature Protocols, 3,
153-162.
[51] Chou, K.C. and Shen, H.B. (2007) Euk-mPLoc: a fusion
classifier for large-scale eukaryotic protein subcellular
location prediction by incorporating multiple sites.
Journal of Proteome Research, 6, 1728-1734.
[52] Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D.,
Butler, H., Cherry, J.M., Davis, A.P., Dolinski, K.,
Dwight, S.S., Eppig, J.T. et al. (2000) Gene ontology:
tool for the unification of biology. Nature Genetics, 25,
25-29.
[53] Chou, K.C. (2005) Using amphiphilic pseudo amino acid
composition to predict enzyme subfamily classes. Bioin-
formatics, 21, 10-19.
[54] Chou, K.C. and Shen, H.B. (2007) MemType-2L: A Web
server for predicting membrane proteins and their types
by incorporating evolution information through
Pse-PSSM. Biochem Biophys Res Comm, 360,
339-345.
[55] Chou, K.C. and Shen, H.B. (2008) ProtIdent: A web
server for identifying proteases and their types by fusing
functional domain and sequential evolution information.
Biochem Biophys Res Comm, 376, 321-325.
[56] Shen, H.B. and Chou, K.C. (2009) A top-down approach
to enhance the power of predicting human protein sub-
cellular localization: Hum-mPLoc 2.0. Analytical Bio-
chemistry, in press.
[57] Shen, H.B. and Chou, K.c. (2009) Gpos-mPLoc: A
top-down approach to improve the quality of predicting
subcellular localization of Gram-positive bacterial pro-
teins. Protein & Peptide Letters, submitted.
[58] Shen, H.B. and Chou, K.C. (2009) Gneg-mPLoc: A
top-down strategy to enhance the quality of predicting
subcellular localization of Gram-negative bacterial pro-
teins, to be submitted.
[59] Chou, K.C. and Shen, H.B. (2009) Plant-mPLoc: A
top-down strategy to augment the power for predicting
plant protein subcellular localization, to be submitted.
[60] Shen, H.B. and Chou, K.C. (2007) Gpos-PLoc: an en-
semble classifier for predicting subcellular localization of
Gram-positive bacterial proteins. Protein Engineering,
Design, and Selection, 20, 39-46.
[61] Chou, K.C. and Shen, H.B. (2006) Large-scale predic-
tions of Gram-negative bacterial protein subcellular loca-
tions. Journal of Proteome Research, 5, 3420-3428.
[62] Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K.
and Walter, P. (2002) Molecular biology of the cell, 4th
edition. Garland Science, New York.
[63] Shen, H.B. and Chou, K.C. (2007) Nuc-PLoc: A new
web-server for predicting protein subnuclear localization
by fusing PseAA composition and PsePSSM. Protein
Engineering, Design & Selection, 20, 561-567.
[64] Rapoport, T.A. (1992) Transport of proteins across the
endoplasmic reticulum membrane. Science, 258,
931-936.
[65] Zheng, N. and Gierasch, L.M. (1996) Signal sequences:
the same yet different. Cell, 86, 849-852.
[66] Chou, K.C. (2001) Prediction of signal peptides using
scaled window. Peptides, 22, 1973-1979.
[67] McGeoch, D.J. (1985) On the predictive recognition of
signal peptide sequences. Virus Res, 3, 271-286.
[68] von Heijne, G. (1986) A new method for predicting signal
sequence cleavage sites. Nucleic Acids Research, 14,
4683-4690.
[69] Folz, R.J. and Gordon, J.I. (1987) Computer-assisted
predictions of signal peptidase processing sites. Biochem
Biophys Res Comm, 146, 870-877.
[70] Ladunga, I., Czako, F., Csabai, I. and Geszti, T. (1991)
Improving signal peptide prediction accuracy by simu-
lated neural network. Comput Appl Biosci, 7, 485-487.
[71] Arrigo, P., Giuliano, F., Scalia, F., Rapallo, A. and
Damiani, G. (1991) Identification of a new motif on nu-
cleic acid sequence data using Kohonen's self-organizing
map. Comput Appl Biosci, 7, 353-357.
[72] Schneider, G. and Wrede, P. (1993) Signal analysis of
protein targeting sequences. Protein Seq Data Anal, 5,
227-236.
[73] Nielsen, H., Engelbrecht, J., Brunak, S. and von Heijne,
G. (1997) Identification of prokaryotic and eukaryotic
signal peptides and prediction of their cleavage sites.
Protein Engineering, 10, 1-6.
[74] Emanuelsson, O., Nielsen, H. and von Heijne, G. (1999)
ChloroP, a neural network-based method for predicting
chloroplast transit peptides and their cleavage sites. Pro-
tein Science, 8, 978-984.
[75] Bendtsen, J.D., Nielsen, H., von Heijne, G. and Brunak, S.
(2004) Improved prediction of signal peptides: SignalP
3.0. J Mol Biol, 340, 783-795.
[76] Hiller, K., Grote, A., Scheer, M., Munch, R. and Jahn, D.
(2004) PrediSi: prediction of signal peptides and their
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
87
cleavage positions. Nucleic Acids Res, 32, W375-379.
[77] Chou, K.C. (2002) Review: Prediction of protein signal
sequences. Current Protein and Peptide Science, 3,
615-622.
[78] Chou, K.C. (2001) Using subsite coupling to predict
signal peptides. Protein Engineering, 14, 75-79.
[79] Chou, K.C. and Shen, H.B. (2007) Signal-CF: a sub-
site-coupled and window-fusing approach for predicting
signal peptides. Biochem Biophys Res Comm, 357,
633-640.
[80] Hiss, J.A. and Schneider, G. (2009) Architecture, function
and prediction of long signal peptides. Brief Bioinform,
10, 569-578.
[81] Kall, L., Krogh, A. and Sonnhammer, E.L. (2007) Ad-
vantages of combined transmembrane topology and sig-
nal peptide prediction--the Phobius web server. Nucleic
Acids Res, 35, W429-432.
[82] Shen, H.B. and Chou, K.C. (2007) Signal-3L: a 3-layer
approach for predicting signal peptide. Biochem Bio-
phys Res Comm, 363, 297-303.
[83] Reynolds, S.M., Kall, L., Riffle, M.E., Bilmes, J.A. and
Noble, W.S. (2008) Transmembrane topology and signal
peptide prediction using dynamic bayesian networks.
PLoS Comput Biol, 4, e1000213.
[84] Chou, K.C. (2004) Modelling extracellular domains of
GABA-A receptors: subtypes 1, 2, 3, and 5. Biochemical
and Biophysical Research Communications, 316,
636-642.
[85] Chou, K.C. (1993) Conformational change during
photocycle of bacteriorhodopsin and its proton-pumping
mechanism. Journal of Protein Chemistry, 12, 337-350.
[86] Chou, K.C. (1994) Mini Review: A molecular piston
mechanism of pumping protons by bacteriorhodopsin.
Amino Acids, 7, 1-17.
[87] Schnell, J.R. and Chou, J.J. (2008) Structure and mecha-
nism of the M2 proton channel of influenza A virus. Na-
ture, 451, 591-595.
[88] Doyle, D.A., Morais, C.J., Pfuetzner, R.A., Kuo, A.,
Gulbis, J.M., Cohen, S.L., Chait, B.T. and MacKinnon, R.
(1998) The structure of the potassium channel: molecular
basis of K+ conduction and selectivity. Science, 280,
69-77.
[89] Chou, K.C. (2004) Insights from modelling
three-dimensional structures of the human potassium and
sodium channels. Journal of Proteome Research, 3,
856-861.
[90] Huang, R.B., Du, Q.S., Wang, C.H. and Chou, K.C.
(2008) An in-depth analysis of the biological functional
studies based on the NMR M2 channel structure of in-
fluenza A virus. Biochem Biophys Res Comm, 377,
1243-1247.
[91] Du, Q.S., Huang, R.B., Wang, C.H., Li, X.M. and Chou,
K.C. (2009) Energetic analysis of the two controversial
drug binding sites of the M2 proton channel in influenza
A virus. Journal of Theoretical Biology, 259, 159-164.
[92] Pielak, R.M., Jason R. Schnell, J.R. and Chou, J.J. (2009)
Mechanism of drug inhibition and drug resistance of in-
fluenza A M2 channel. Proceedings of National Acad-
emy of Science, USA, 106, 7379-7384.
[93] Oxenoid, K. and Chou, J.J. (2005) The structure of
phospholamban pentamer reveals a channel-like archi-
tecture in membranes. Proc Natl Acad Sci U S A, 102,
10870-10875.
[94] Douglas, S.M., Chou, J.J. and Shih, W.M. (2007)
DNA-nanotube-induced alignment of membrane proteins
for NMR structure determination. Proc Natl Acad Sci U
S A, 104, 6644-6648.
[95] Nakashima, H., Nishikawa, K. and Ooi, T. (1986) The
folding type of a protein is relevant to the amino acid
composition. J Biochem, 99, 152-162.
[96] Klein, P. and Delisi, C. (1986) Prediction of protein
structural class from amino acid sequence. Biopolymers,
25, 1659-1672.
[97] Klein, P. (1986) Prediction of protein structural class by
discriminant analysis. Biochim Biophys Acta, 874,
205-215.
[98] Chou, K.C. and Zhang, C.T. (1994) Predicting protein
folding types by distance functions that make allowances
for amino acid interactions. J Biol Chem, 269,
22014-22020.
[99] Chou, K.C. (1995) A novel approach to predicting pro-
tein structural classes in a (20-1)-D amino acid composi-
tion space. Proteins: Structure, Function & Genetics, 21,
319-344.
[100] Liu, W. and Chou, K.C. (1998) Prediction of protein
structural classes by modified Mahalanobis discriminant
algorithm. Journal of Protein Chemistry, 17, 209-217.
[101] Chou, K.C., Liu, W., Maggiora, G.M. and Zhang, C.T.
(1998) Prediction and classification of domain structural
classes. PROTEINS: Structure, Function, and Genetics,
31, 97-103.
[102] Chou, K.C. and Maggiora, G.M. (1998) Domain struc-
tural class prediction. Protein Engineering, 11, 523-538.
[103] Chou, K.C. (1999) A key driving force in determination
of protein structural classes. Biochemical and Biophysi-
cal Research Communications, 264, 216-224.
[104] Chou, K.C. and Elrod, D.W. (1999) Prediction of mem-
brane protein types and subcellular locations. PROTEINS:
Structure, Function, and Genetics, 34, 137-153.
[105] Cai, Y.D., Liu, X.J. and Chou, K.C. (2001) Artificial
neural network model for predicting membrane protein
types. Journal of Biomolecular Structure and Dynamics,
18, 607-610.
[106] Guo, Z.M. (2002) Prediction of Membrane protein types
by using pattern recognition method based on pseudo
amino acid composition. Master Thesis, Bio-X Life Sci-
ence Research Center, Shanghai Jiaotong University.
[107] Cai, Y.D., Zhou, G.P. and Chou, K.C. (2003) Support
vector machines for predicting membrane protein types
by using functional domain composition. Biophysical
Journal, 84, 3257-3263.
[108] Cai, Y.D., Pong-Wong, R., Feng, K., Jen, J.C.H. and
Chou, K.C. (2004) Application of SVM to predict mem-
brane protein types. Journal of Theoretical Biology, 226,
373-376.
[109] Wang, M., Yang, J., Liu, G.P., Xu, Z.J. and Chou, K.C.
(2004) Weighted-support vector machines for predicting
membrane protein types based on pseudo amino acid
composition. Protein Engineering, Design, and Selection,
17, 509-516.
[110] Chou, K.C. and Cai, Y.D. (2005) Prediction of membrane
protein types by incorporating amphipathic effects.
Journal of Chemical Information and Modeling, 45,
407-413.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
88
[111] Liu, H., Wang, M. and Chou, K.C. (2005)
Low-frequency Fourier spectrum for predicting mem-
brane protein types. Biochem Biophys Res Commun, 336,
737-739.
[112] Wang, M., Yang, J., Xu, Z.J. and Chou, K.C. (2005)
SLLE for predicting membrane protein types. Journal of
Theoretical Biology, 232, 7-15.
[113] Shen, H.B. and Chou, K.C. (2005) Using optimized evi-
dence-theoretic K-nearest neighbor classifier and pseudo
amino acid composition to predict membrane protein
types. Biochemical & Biophysical Research Communica-
tions, 334, 288-292.
[114] Shen, H.B., Yang, J. and Chou, K.C. (2006) Fuzzy KNN
for predicting membrane protein types from pseudo
amino acid composition. Journal of Theoretical Biology,
240, 9-13.
[115] Wang, S.Q., Yang, J. and Chou, K.C. (2006) Using
stacked generalization to predict membrane protein types
based on pseudo amino acid composition. Journal of
Theoretical Biology, 242, 941-946.
[116] Shen, H.B. and Chou, K.C. (2007) Using ensemble clas-
sifier to identify membrane protein types. Amino Acids,
32, 483-488.
[117] Yang, X.G., Luo, R.Y. and Feng, Z.P. (2007) Using amino
acid and peptide composition to predict membrane pro-
tein types. Biochem Biophys Res Commun, 353, 164-169.
[118] Pu, X., Guo, J., Leung, H. and Lin, Y. (2007) Prediction
of membrane protein types from sequences and posi-
tion-specific scoring matrices. J Theor Biol, 247,
259–265.
[119] Afjehi-Sadat, L. and Lubec, G. (2007) Identification of
enzymes and activity from two-dimensional gel electro-
phoresis. Nature Protocols, 2, 2318-2324.
[120] Chou, K.C. and Elrod, D.W. (2003) Prediction of enzyme
family classes. Journal of Proteome Research, 2,
183-190.
[121] Chou, K.C. and Cai, Y.D. (2004) Predicting enzyme fam-
ily class in a hybridization space. Protein Science, 13,
2857-2863.
[122] Cai, C.Z., Han, L.Y., Ji, Z.L. and Chen, Y.Z. (2004) En-
zyme family classification by support vector machines.
PROTEINS: Structure, Function, and Bioinformatics, 55,
66-76.
[123] 1Cai, Y.D. and Chou, K.C. (2005) Predicting enzyme
subclass by functional domain composition and pseudo
amino acid composition. Journal of Proteome Research,
4, 967-971.
[124] Huang, W.L., Chen, H.M., Hwang, S.F. and Ho, S.Y.
(2006) Accurate prediction of enzyme subfamily class
using an adaptive fuzzy k-nearest neighbor method. Bio-
systems, 90, 405-413.
[125] Zhou, X.B., Chen, C., Li, Z.C. and Zou, X.Y. (2007)
Using Chou's amphiphilic pseudo-amino acid composi-
tion and support vector machine for prediction of enzyme
subfamily classes. Journal of Theoretical Biology, 248,
546–551.
[126] Shen, H.B. and Chou, K.C. (2007) EzyPred: A top-down
approach for predicting enzyme functional classes and
subclasses. Biochem Biophys Res Comm, 364, 53-59.
[127] Bairoch, A. (2000) The ENZYME Database in 2000.
Nucleic Acids Research, 28, 304-305.
[128] Poorman, R.A., Tomasselli, A.G., Heinrikson, R.L. and
Kezdy, F.J. (1991) A cumulative specificity model for
proteases from human immunodeficiency virus types 1
and 2, inferred from statistical analysis of an extended
substrate data base. J Biol Chem, 266, 14554-14561.
[129] Qin, H., Srinvasula, S.M., Wu, G., Fernandes-Alnemri, T.,
Alnemri, E.S., and Shi, Y. (1999) Structural basis of pro-
caspase-9 recruitment by the apoptotic prote-
ase-activating factor 1. Nature, 399, 549-557.
[130] Chou, J.J., Li, H., Salvessen, G.S., Yuan, J. and Wagner,
G. (1999) Solution structure of BID, an intracellular am-
plifier of apoptotic signalling. Cell, 96, 615-624.
[131] Watt, W., Koeplinger, K.A., Mildner, A.M., Heinrikson,
R.L., Tomasselli, A.G. and Watenpaugh, K.D. (1999) The
atomic resolution structure of human caspase-8, a key
activator of apoptosis. Structure, 7, 1135-1143.
[132] Chou, K.C., Wei, D.Q. and Zhong, W.Z. (2003) Binding
mechanism of coronavirus main proteinase with ligands
and its implication to drug design against SARS. (Erra-
tum: ibid., 2003, Vol.310, 675). Biochem Biophys Res
Comm, 308, 148-151.
[133] Puente, X.S., Sanchez, L.M., Overall, C.M. and Lo-
pez-Otin, C. (2003) Human and mouse proteases: a
comparative genomic approach. Nat Rev Genet, 4,
544-558.
[134] Chou, K.C., Wei, D.Q., Du, Q.S., Sirois, S., Shen, H.B.
and Zhong, W.Z. (2009) Study of inhibitors against
SARS coronavirus by computational approaches. In
Lendeckel, U. and Hooper, N. (eds.), Viral proteases and
antiviral protease inhibitor therapy. Proteases in Biology
and Disease, Springer Publishing, 8.
[135] Chou, K.C. (1993) A vectorized sequence-coupling
model for predicting HIV protease cleavage sites in pro-
teins. J Biol Chem, 268, 16938-16948.
[136] Chou, K.C. (1996) Review: Prediction of HIV protease
cleavage sites in proteins. Analytical Biochemistry, 233,
1-14.
[137] You, L., Garwicz, D. and Rognvaldsson, T. (2005) Com-
prehensive bioinformatic analysis of the specificity of
human immunodeficiency virus type 1 protease. J Virol,
79, 12477-12486.
[138] Rognvaldsson, T., You, L. and Garwicz, D. (2007) Bio-
informatic approaches for modeling the substrate speci-
ficity of HIV-1 protease: an overview. Expert Rev Mol
Diagn, 7, 435-451.
[139] Liang, G.Z. and Li, S.Z. (2007) A new sequence repre-
sentation as applied in better specificity elucidation for
human immunodeficiency virus type 1 protease. Bio-
polymers, 88, 401-412.
[140] Rawlings, N.D., Tolle, D.P. and Barrett, A.J. (2004)
MEROPS: the peptidase database. Nucleic Acids Re-
search, 32, D160-D164.
[141] Chou, K.C. and Cai, Y.D. (2006) Prediction of protease
types in a hybridization space. Biochem Biophys Res
Comm, 339, 1015-1020.
[142] Zhou, G.P. and Cai, Y.D. (2006) Predicting protease types
by hybridizing gene ontology and pseudo amino acid
composition. PROTEINS: Structure, Function, and Bio-
informatics, 63, 681-684.
[143] Shen, H.B. and Chou, K.C. (2009) Identification of pro-
teases and their types. Analytical Biochemistry, 385,
153-160.
[144] Heuss, C. and Gerber, U. (2000) G-protein-independent
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
89
signaling by G-protein-coupled receptors. Trends Neuro-
sci, 23, 469-475.
[145] Milligan, G. and White, J.H. (2001) Protein-protein in-
teractions at G-protein-coupled receptors. Trends Phar-
macol Sci, 22, 513-518.
[146] Hall, R.A. and Lefkowitz, R.J. (2002) Regulation of G
protein-coupled receptor signaling by scaffold proteins.
Circ Res, 91, 672-680.
[147] Chou, K.C. (2005) Coupling interaction between throm-
boxane A2 receptor and alpha-13 subunit of guanine nu-
cleotide-binding protein. Journal of Proteome Research,
4, 1681-1686.
[148] Call, M.E., Schnell, J.R., Xu, C., Lutz, R.A., Chou, J.J.
and Wucherpfennig, K.W. (2006) The structure of the
zetazeta transmembrane dimer reveals features essential
for its assembly with the T cell receptor. Cell, 127,
355-368.
[149] Chou, K.C. (2004) Insights from modelling the 3D
structure of the extracellular domain of alpha7 nicotinic
acetylcholine receptor. Biochemical and Biophysical Re-
search Communication, 319, 433-438.
[150] Chou, K.C. (2004) Molecular therapeutic target for
type-2 diabetes. Journal of Proteome Research, 3,
1284-1288.
[151] Wei, D.Q., Du, Q.S., Sun, H. and Chou, K.C. (2006)
Insights from modeling the 3D structure of H5N1 influ-
enza virus neuraminidase and its binding interactions
with ligands. Biochem Biophys Res Comm, 344,
1048-1055.
[152] Wang, S.Q., Du, Q.S. and Chou, K.C. (2007) Study of
drug resistance of chicken influenza A virus (H5N1)
from homology-modeled 3D structures of neuramini-
dases. Biochem Biophys Res Comm, 354, 634-640.
[153] Wang, S.Q., Du, Q.S., Huang, R.B., Zhang, D.W. and
Chou, K.C. (2009) Insights from investigating the inter-
action of oseltamivir (Tamiflu) with neuraminidase of the
2009 H1N1 swine flu virus. Biochem Biophys Res Com-
mun, 386, 432-436.
[154] Elrod, D.W. and Chou, K.C. (2002) A study on the corre-
lation of G-protein-coupled receptor types with amino
acid composition. Protein Engineering, 15, 713-715.
[155] Chou, K.C. and Elrod, D.W. (2002) Bioinformatical
analysis of G-protein-coupled receptors. Journal of Pro-
teome Research, 1, 429-433.
[156] Bhasin, M. and Raghava, G.P. (2005) GPCRsclass: a web
tool for the classification of amine type of
G-protein-coupled receptors. Nucleic Acids Research, 33,
W143-147.
[157] Chou, K.C. (2005) Prediction of G-protein-coupled re-
ceptor classes. Journal of Proteome Research, 4,
1413-1418.
[158] Wen, Z., Li, M., Li, Y., Guo, Y. and Wang, K. (2007)
Delaunay triangulation with partial least squares projec-
tion to latent structures: a model for G-protein coupled
receptors classification and fast structure recognition.
Amino Acids, 32, 277-283.
[159] Gao, Q.B. and Wang, Z.Z. (2006) Classification of
G-protein coupled receptors at four levels. Protein Eng
Des Sel, 19, 511-516.
[160] Xiao, X., Wang, P. and Chou, K.C. (2009) GPCR-CA: A
cellular automaton image approach for predicting
G-protein-coupled receptor functional classes. Journal of
Computational Chemistry, 30, 1414-1423.
[161] Wolfram, S. (1984) Cellular automation as models of
complexity. Nature, 311, 419-424.
[162] Wolfram, S. (2002) A New Kind of Science. Wolfram
Media Inc., Champaign, IL.
[163] Xiao, X., Shao, S., Ding, Y., Huang, Z., Chen, X. and
Chou, K.C. (2005) Using cellular automata to generate
Image representation for biological sequences. Amino
Acids, 28, 29-35.
[164] Chou, K.C. (2000) Prediction of protein subcellular loca-
tions by incorporating quasi-sequence-order effect. Bio-
chemical & Biophysical Research Communications, 278,
477-483.
[165] Althaus, I.W., Chou, J.J., Gonzales, A.J., Diebel, M.R.,
Chou, K.C., Kezdy, F.J., Romero, D.L., Aristoff, P.A.,
Tarpley, W.G. and Reusser, F. (1993) Steady-state kinetic
studies with the non-nucleoside HIV-1 reverse transcrip-
tase inhibitor U-87201E. J Biol Chem, 268, 6119-6124.
[166] Althaus, I.W., Gonzales, A.J., Chou, J.J., Diebel, M.R.,
Chou, K.C., Kezdy, F.J., Romero, D.L., Aristoff, P.A.,
Tarpley, W.G. and Reusser, F. (1993) The quinoline
U-78036 is a potent inhibitor of HIV-1 reverse transcrip-
tase. J Biol Chem, 268, 14875-14880.
[167] Althaus, I.W., Chou, J.J., Gonzales, A.J., Diebel, M.R.,
Chou, K.C., Kezdy, F.J., Romero, D.L., Aristoff, P.A.,
Tarpley, W.G. and Reusser, F. (1993) Kinetic studies with
the nonnucleoside HIV-1 reverse transcriptase inhibitor
U-88204E. Biochemistry, 32, 6548-6554.
[168] Althaus, I.W., Chou, J.J., Gonzales, A.J., Diebel, M.R.,
Chou, K.C., Kezdy, F.J., Romero, D.L., Aristoff, P.A.,
Tarpley, W.G. and Reusser, F. (1994) Steady-state kinetic
studies with the polysulfonate U-9843, an HIV reverse
transcriptase inhibitor. Ex perien tia, 50, 23-28.
[169] Althaus, I.W., Chou, J.J., Gonzales, A.J., Diebel, M.R.,
Chou, K.C., Kezdy, F.J., Romero, D.L., Thomas, R.C.,
Aristoff, P.A., Tarpley, W.G. et al. (1994) Kinetic studies
with the non-nucleoside HIV-1 reverse transcriptase in-
hibitor U-90152E. Biochemical Pharmacology, 47,
2017-2028.
[170] Althaus, I.W., Chou, K.C., Franks, K.M., Diebel, M.R.,
Kezdy, F.J., Romero, D.L., Thomas, R.C., Aristoff, P.A.,
Tarpley, W.G. and Reusser, F. (1996) The benzyl-
thio-pyrididine U-31,355 is a potent inhibitor of HIV-1
reverse transcriptase. Biochemical Pharmacology, 51,
743-750.
[171] Chou, K.C., Kezdy, F.J. and Reusser, F. (1994) Review:
Steady-state inhibition kinetics of processive nucleic acid
polymerases and nucleases. Analytical Biochemistry, 221,
217-230.
[172] McQuade, T.J., Tomasselli, A.G., Liu, L., Karacostas, V.,
Moss, B., Sawyer, T.K., Heinrikson, R.L. and Tarpley,
W.G. (1990) A synthetic HIV-1 protease inhibitor with
antiviral activity arrests HIV-like particle maturation.
Science, 247, 454-456.
[173] Meek, T.D., Lambert, D.M., Dreyer, G.B., Carr, T.J.,
Tomaszek, T.A., Jr., Moore, M.L., Strickler, J.E., De-
bouck, C., Hyland, L.J., Matthews, T.J. et al. (1990) In-
hibition of HIV-1 protease in infected T-lymphocytes by
synthetic peptide analogues. Nature, 343, 90-92.
[174] Wlodawer, A. and Erickson, J.W. (1993) Structure-based
inhibitors of HIV-1 protease. Annu Rev Biochem, 62,
543-585.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
90
[175] Barre-Sinoussi, F., Chermann, J.C., Rey, F., Nugeyre,
M.T., Chamaret, S., Gruest, J., Dauguet, C., Axler-Blin,
C., Vezinet-Brun, F., Rouzioux, C. et al. (1983) Isolation
of a T-lymphotropic retrovirus from a patient at risk for
acquired immune deficiency syndrome (AIDS). Science,
220, 868-871.
[176] Gallo, R.C., Salahuddin, S.Z., Popovic, M., Shearer,
G.M., Kaplan, M., Haynes, B.F., Palker, T.J., Redfield, R.,
Oleske, J., Safai, B. et al. (1984) Frequent detection and
isolation of cytopathic retroviruses (HTLV-III) from pa-
tients with AIDS and at risk for AIDS. Science, 224,
500-503.
[177] Miller, M., Schneider, J., Sathyanarayana, B.K., Toth,
M.V., Marshall, G.R., Clawson, L., Selk, L., Kent, S.B.
and Wlodawer, A. (1989) Structure of complex of syn-
thetic HIV-1 protease with a substrate-based inhibitor at
2.3 A resolution. Science, 246, 1149-1152.
[178] Schechter, I. and Berger, A. (1967) On the size of the
active site in protease. I. Papain. Biochem Biophys Res
Comm, 27, 157-162.
[179] Chou, K.C., Chen, N.Y. and Forsen, S. (1981) The bio-
logical functions of low-frequency phonons: 2. Coopera-
tive effects. Chemica Scripta, 18, 126-132.
[180] Chou, K.C., Zhang, C.T. and Kezdy, F.J. (1993) A vector
approach to predicting HIV protease cleavage sites in
proteins. Proteins: Structure, Function, and Genetics, 16,
195-204.
[181] Chou, J.J. (1993) Predicting cleavability of peptide se-
quences by HIV protease via correlation-angle approach.
Journal of Protein Chemistry, 12, 291-302.
[182] Chou, K.C. and Zhang, C.T. (1993) Studies on the speci-
ficity of HIV protease: an application of Markov chain
theory. Journal of Protein Chemistry, 12, 709-724.
[183] Chou, J.J. (1993) A formulation for correlating properties
of peptides and its application to predicting human im-
munodeficiency virus protease-cleavable sites in proteins.
Biopolymers, 33, 1405-1414.
[184] Zhang, C.T. and Chou, K.C. (1993) An alter-
nate-subsite-coupled model for predicting HIV protease
cleavage sites in proteins. Protein Engineering, 7, 65-73.
[185] Thompson, T.B., Chou, K.C. and Zheng, C. (1995) Neu-
ral network prediction of the HIV-1 protease cleavage
sites. Journal of Theoretical Biology 177, 369-379.
[186] Chou, K.C., Tomasselli, A.L., Reardon, I.M. and Hein-
rikson, R.L. (1996) Predicting HIV protease cleavage
sites in proteins by a discriminant function method.
PROTEINS: Structure, Function, and Genetics, 24,
51-72.
[187] Shen, H.B. and Chou, K.C. (2008) HIVcleave: a
web-server for predicting HIV protease cleavage sites in
proteins. Analytical Biochemistry, 375, 388-390.
[188] Klotz, I.M., Darnell, D.W. and Langerman, N.R. (1975)
Quaternary structure of proteins. In Neurath, H. and Hill,
R. L. (eds.), The Proteins (3rd ed). Academic Press, New
York, 1, 226-241.
[189] Chou, K.C. and Cai, Y.D. (2003) Predicting protein qua-
ternary structure by pseudo amino acid composition.
PROTEINS: Structure, Function, and Genetics, 53,
282-289.
[190] Goodsell, D.S. and Olson, A.J. (2000) Structural symme-
try and protein function. Annu Rev Biophys Biomol Struct,
29, 105-153.
[191] Levy, E.D., Boeri Erba, E., Robinson, C.V. and
Teichmann, S.A. (2008) Assembly reflects evolution of
protein complexes. Nature, 453, 1262-1265.
[192] Chen, Z., Alcayaga, C., Suarez-Isla, B.A., O'Rourke, B.,
Tomaselli, G. and Marban, E. (2002) A "minimal" sodium
channel construct consisting of ligated S5-P-S6 segments
forms a toxin-activatable ionophore. J Biol Chem, 277,
24653-24658.
[193] Oxenoid, K., Rice, A.J. and Chou, J.J. (2007) Comparing
the structure and dynamics of phospholamban pentamer
in its unphosphorylated and pseudo-phosphorylated
states. Protein Sci, 16, 1977-1983.
[194] Tretter, V., Ehya, N., Fuchs, K. and Sieghart, W. (1997)
Stoichiometry and assembly of a recombinant GABAA
receptor subtype. Journal of Neuroscience, 17,
2728-2737.
[195] Perutz, M.F. (1964) The Hemoglobin Molecule. Scien-
tific American, 211 , 65-76.
[196] Wei, H., Wang, C.H., Du, Q.S., Meng, J. and Chou, K.C.
(2009) Investigation into adamantane-based M2 inhibi-
tors with FB-QSAR. Medicinal Chemistry, 5, 305-317.
[197] Shen, H.B. and Chou, K.C. (2009) QuatIdent: A web
server for identifying protein quaternary structural attrib-
ute by fusing functional domain and sequential evolution
information. Journal of Proteome Research, 8,
1577–1584.
[198] Xiao, X., Wang, P. and Chou, K.C. (2009) Predicting
protein quaternary structural attribute by hybridizing
functional domain composition and pseudo amino acid
composition. Journal of Applied Crystallography, 42,
169-173.
[199] Garian, R. (2001) Prediction of quaternary structure from
primary structure. Bioinformatics, 17, 551-556.
[200] Zhang, S.W., Chen, W., Yang, F. and Pan, Q. (2008) Us-
ing Chou's pseudo amino acid composition to predict
protein quaternary structure: a sequence-segmented
PseAAC approach. Amino Acids, 35, 591-598.
[201] Anfinsen, C.B. and Scheraga, H.A. (1975) Experimental
and theoretical aspects of protein folding. Adv Protein
Chem, 29, 205-300.
[202] Aguzzi, A. (2008) Unraveling prion strains with cell
biology and organic chemistry. Proc Natl Acad Sci U S A,
105, 11-12.
[203] Dobson, C.M. (2001) The structural basis of protein
folding and its links with human disease. Philos Trans R
Soc Lond B Biol Sci, 356, 133-145.
[204] Prusiner, S.B. (1998) Prions. Proc Natl Acad Sci U S A,
95, 13363-13383.
[205] Anfinsen, C.B. (1973) Principles that govern the folding
of protein chains. Science, 181, 223-230.
[206] Chou, K.C. and Scheraga, H.A. (1982) Origin of the
right-handed twist of beta-sheets of poly-L-valine chains.
Proceedings of National Academy of Sciences, USA, 79,
7047-7051.
[207] Chou, K.C., Maggiora, G.M., Nemethy, G. and Scheraga,
H.A. (1988) Energetics of the structure of the
four-alpha-helix bundle in proteins. Proceedings of Na-
tional Academy of Sciences, USA, 85, 4295-4299.
[208] Chou, K.C., Nemethy, G. and Scheraga, H.A. (1990)
Review: Energetics of interactions of regular structural
elements in proteins. Accounts of Chemical Research, 23,
134-141.
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
91
[209] Chou, K.C., Nemethy, G. and Scheraga, H.A. (1984)
Energetic approach to packing of a-helices: 2. General
treatment of nonequivalent and nonregular helices.
Journal of American Chemical Society, 106, 3161-3170.
[210] Chou, K.C., Nemethy, G., Pottle, M. and Scheraga, H.A.
(1989) Energy of stabilization of the right-handed
beta-alpha-beta crossover in proteins. Journal of Mo-
lecular Biology, 205, 241-249.
[211] Chou, K.C. and Carlacci, L. (1991) Energetic approach
to the folding of alpha/beta barrels. Proteins: Structure,
Function, and Genetics, 9, 280-295.
[212] Chou, K.C. (1992) Energy-optimized structure of anti-
freeze protein and its binding mechanism. Journal of
Molecular Biology, 223, 509-517.
[213] Carlacci, L., Chou, K.C. and Maggiora, G.M. (1991) A
heuristic approach to predicting the tertiary structure of
bovine somatotropin. Biochemistry, 30, 4389-4398.
[214] Scheraga, H.A., Khalili, M. and Liwo, A. (2007) Pro-
tein-folding dynamics: overview of molecular simulation
techniques. Annu Rev Phys Chem, 58, 57-83.
[215] Holm, L. and Sander, C. (1999) Protein folds and fami-
lies: sequence and structure alignments. Nucleic Acids
Research, 27, 244-247.
[216] Chou, K.C. (1995) The convergence-divergence duality
in lectin domains of the selectin family and its implica-
tions. FEBS Letters, 363, 123-126.
[217] Chou, K.C., Jones, D. and Heinrikson, R.L. (1997) Pre-
diction of the tertiary structure and substrate binding site
of caspase-8. FEBS Letters, 419, 49-54.
[218] Chou, J.J., Matsuo, H., Duan, H. and Wagner, G. (1998)
Solution structure of the RAIDD CARD and model for
CARD/CARD interaction in caspase-2 and caspase-9 re-
cruitment. Cell, 94, 171-180.
[219] Chou, K.C., Tomasselli, A.G. and Heinrikson, R.L. (2000)
Prediction of the Tertiary Structure of a Cas-
pase-9/Inhibitor Complex. FEBS Letters, 470, 249-256.
[220] Chou, K.C. and Howe, W.J. (2002) Prediction of the
tertiary structure of the beta-secretase zymogen. BBRC,
292, 702-708.
[221] Du, Q.S., Wang, S., Wei, D.Q., Sirois, S. and Chou, K.C.
(2005) Molecular modelling and chemical modification
for finding peptide inhibitor against SARS CoV Mpro.
Analytical Biochemistry, 337, 262-270.
[222] Zhang, R., Wei, D.Q., Du, Q.S. and Chou, K.C. (2006)
Molecular modeling studies of peptide drug candidates
against SARS. Medicinal Chemistry, 2, 309-314.
[223] Wang, J.F., Wei, D.Q., Li, L., Zheng, S.Y., Li, Y.X. and
Chou, K.C. (2007) 3D structure modeling of cytochrome
P450 2C19 and its implication for personalized drug de-
sign. Biochem Biophys Res Commun (Corrigendum: ibid,
2007, Vol357, 330), 355, 513-519.
[224] Wang, J.F., Wei, D.Q., Lin, Y., Wang, Y.H., Du, H.L., Li,
Y.X. and Chou, K.C. (2007) Insights from modeling the
3D structure of NAD(P)H-dependent D-xylose reductase
of Pichia stipitis and its binding interactions with NAD
and NADP. Biochem Biophys Res Comm, 359,
323-329.
[225] Wang, J.F., Wei, D.Q., Chen, C., Li, Y. and Chou, K.C.
(2008) Molecular modeling of two CYP2C19 SNPs and
its implications for personalized drug design. Protein &
Peptide Letters, 15, 27-32.
[226] Wang, J.F., Wei, D.Q., Du, H.L., Li, Y.X. and Chou, K.C.
(2008) Molecular modeling studies on
NADP-dependence of Candida tropicalis strain xylose
reductase. The Open Bioinformatics Journal, 2, 72-79.
[227] Ding, C.H. and Dubchak, I. (2001) Multi-class protein
fold recognition using support vector machines and neu-
ral networks. Bioinformatics, 17, 349-358.
[228] Finkelstein, A.V. and Ptitsyn, O.B. (1987) Why do
globular proteins fit the limited set of folding patterns?
Prog Biophys Mol Biol, 50, 171-190.
[229] Chou, K.C. and Zhang, C.T. (1995) Review: Prediction
of protein structural classes. Critical Reviews in Bio-
chemistry and Molecular Biology, 30, 275-349.
[230] Dubchak, I., Muchnik, I., Mayor, C., Dralyuk, I. and Kim,
S.H. (1999) Recognition of a protein fold in the context
of the Structural Classification of Proteins (SCOP) clas-
sification. PROTEINS: Structure, Function, and Genetics,
35, 401-407.
[231] Murzin, A.G., Brenner, S.E., Hubbard, T. and Chothia, C.
(1995) SCOP: a structural classification of protein data-
base for the investigation of sequence and structures.
Journal of Molecular Biology, 247, 536-540.
[232] Shen, H.B. and Chou, K.C. (2006) Ensemble classifier
for protein fold pattern recognition. Bioinformatics, 22,
1717-1722.
[233] Chou, K.C. and Shen, H.B. (2006) Predicting eukaryotic
protein subcellular location by fusing optimized evi-
dence-theoretic K-nearest neighbor classifiers. Journal of
Proteome Research, 5, 1888-1897.
[234] Shen, H.B. and Chou, K.C. (2009) Predicting protein
fold pattern with functional domain and sequential evo-
lution information. Journal of Theoretical Biology, 256,
441-446.
[235] Qiu, L.L., Pabit, S.A., Roitberg, A.E. and Hagen, S.J.
(2002) Smaller and faster: The 20-residue Trp-cage pro-
tein folds in 4 microseconds. Journal of American
Chemical Society, 124, 12952-12953.
[236] Goldberg, M.E., Semisotnov, G.V., Friguet, B., Kuwajima,
K., Ptitsyn, O.B. and Sugai, S. (1990) An early immuno-
reactive folding intermediate of the tryptophan synthease
beta 2 subunit is a 'molten globule'. FEBS Lett, 263,
51-56.
[237] Plaxco, K.W., Simons, K.T. and Baker, D. (1998) Contact
order, transition state placement and the refolding rates of
single domain proteins. J Mol Biol, 277, 985-994.
[238] Ivankov, D.N., Garbuzynskiy, S.O., Alm, E., Plaxco,
K.W., Baker, D. and Finkelstein, A.V. (2003) Contact or-
der revisited: influence of protein size on the folding rate.
Protein Science, 12, 2057-2062.
[239] Zhou, H. and Zhou, Y. (2002) Folding rate prediction
using total contact distance. Biophys Journal, 82,
458-463.
[240] Gromiha, M.M. and Selvaraj, S. (2001) Comparison
between long-range interactions and contact order in de-
termining the folding rate of two-state proteins: applica-
tion of long-range order to folding rate prediction. J Mol
Biol, 310, 27-32.
[241] Nolting, B., Schalike, W., Hampel, P., Grundig, F.,
Gantert, S., Sips, N., Bandlow, W. and Qi, P.X. (2003)
Structural determinants of the rate of protein folding. J
Theor Biol, 223, 299-307.
[242] Ouyang, Z. and Liang, J. (2008) Predicting protein fold-
ing rates from geometric contact and amino acid se-
K. C. Chou et al. / Natural Science 1 (2009) 63-92
Copyright © 2009 SciRes. OPEN ACCESS
92
quence. Protein Science, 17, 1256-1263.
[243] Ivankov, D.N. and Finkelstein, A.V. (2004) Prediction of
protein folding rates from the amino acid se-
quence-predicted secondary structure. Proc Natl Acad Sci
USA, 101, 8942-8944.
[244] Gromiha, M.M., Thangakani, A.M. and Selvaraj, S.
(2006) FOLD-RATE: prediction of protein folding rates
from amino acid sequence. Nucleic Acids Res, 34,
W70-74.
[245] Chou, K.C. (1990) Review: Applications of graph theory
to enzyme kinetics and protein folding kinetics. Steady
and non-steady state systems. Biophysical Chemistry, 35,
1-24.
[246] Chou, K.C. (1989) Graphical rules in steady and
non-steady enzyme kinetics. J Biol Chem, 264,
12074-12079.
[247] Lin, S.X. and Neet, K.E. (1990) Demonstration of a slow
conformational change in liver glucokinase by fluores-
cence spectroscopy. J Biol Chem, 265, 9670-9675.
[248] Beyer, W.H. (1988) CRC Handbook of Mathematical
Science, (6th Edition), Chapter 10, page 544. CRC Press,
Inc., Boca Raton, Florida.
[249] Shen, H.B., Song, J.N. and Chou, K.C. (2009) Prediction
of protein folding rates from primary sequence by fusing
multiple sequential features. Journal of Biomedical Sci-
ence and Engineering (JBiSE), 2, 136-143 (open acces-
sible at http://www.srpublishing.org/journal/jbise/).
[250] Chou, K.C. and Shen, H.B. (2009) FoldRate: A
web-server for predicting protein folding rates from pri-
mary sequence. The Open Bioinformatics Journal, 3,
31-50 (open accessible at
http://www.bentham.org/open/tobioij/).
[251] Zhang, Z. and Henzel, W.J. (2004) Signal peptide predic-
tion based on analysis of experimentally verified cleav-
age sites. Protein Sci, 13, 2819-2824.
[252] Spector, D.L. (2001) Nuclear domains. J Cell Sci, 114,
2891-2893.
[253] Spiess, M. (1995) Heads or tails - what determines the
orientation of proteins in the membrane. FEBS Lett, 369,
76-79.
[254] Schulz, G.E. and Schirmer, R.H. (1985) Principles of
Protein Structure, Chapter 2, Springer-Verlag, New York.
17-18.