64 Protein & Peptide Letters, 2010, 17, 64-69

Evaluation of Protein Phosphorylation Site Predictors

Shufu Que1,2,#, Yongfei Wang2,#, Peixiang Chen2, Yu-Rong Tang3, Ziding Zhang3, and
Huaqin He1,2,*

1
Key Laboratory of Ministry of Education for Genetic, Breeding and Multiple Utilization of Crops, Fujian Agriculture
and Forestry University, Fuzhou 350002, China; 2College of Life Science, Fujian Agriculture and Forestry University,
Fuzhou 350002, China; 3State Key Laboratory of AgroBiotechnology, College of Biological Sciences, China Agricul-
tural University, Beijing 100193, China

Abstract: A series of elegant phosphorylation site prediction methods have been developed, which are playing an increas-
ingly important role in accelerating the experimental characterization of phosphorylation sites in phosphoproteins. In this
study, we selected six recently published methods (DISPHOS, NetPhosK, PPSP, KinasePhos, Scansite and PredPhospho)
to evaluate their performance. First, we compiled three testing datasets containing experimentally verified phosphoryla-
tion sites for mammalian, Arabidopsis and rice proteins. Then, we present the prediction performance of the tested meth-
ods on these three independent datasets. Rather than quantitatively ranking the performance of these methods, we focused
on providing an understanding of the overall performance of the predictors. Based on this evaluation, we found the fol-
lowing results: i) current phosphorylation site predictors are not effective for practical use and there is substantial need to
improve phosphorylation site prediction; ii) current predictors perform poorly when used to predict phosphorylation sites
in plant phosphoproteins, suggesting that a rice-specific predictor will be required to obtain confident computational anno-
tation of phosphorylation sites in rice proteomics research; and iii) the tested predictors are complementary to some ex-
tent, implying that establishment of a meta-server might be a promising approach to developing an improved prediction
system.
Keywords: Protein, Phosphorylation site, Prediction, Evaluation.

INTRODUCTION mechanism of phosphorylation dynamics. Nevertheless,

Phosphorylation is one of the most important types of
post-translational modifications [1-4]. More than 30% of all
eukaryotic proteins are estimated to undergo reversible
phosphorylation [5]. Phosphorylation of substrate sites at
serine (Ser), threonine (Thr) or tyrosine (Tyr) residues is
performed by members of the protein kinase family. Bio-
chemically, phosphorylation involves the transfer of a phos-
phate moiety from adenosine triphosphate (ATP) to the ac-
ceptor residue, thereby generating adenosine diphosphate
(ADP). Protein kinases catalyze the phosphorylation events
that are essential for the regulation of cellular processes such
as metabolism, proliferation, differentiation and apoptosis.
Owing to the importance of protein phosphorylation in regu-
lating cellular signaling, a major goal of current proteomic
efforts is to identify phosphoproteins in higher organisms
[6]. Several large-scale phosphoproteomics studies have
been carried out in yeast [7], mouse [8], human [9], plant
[10], and other species.
Traditional experimental identification of protein kinase-
specific phosphorylation sites on substrates in vivo and in
vitro has provided a foundation for understanding the
*Address correspondence to this author at the Key Laboratory of Ministry
of Education for Genetic, Breeding and Multiple Utilization of Crops, Fu-
jian Agriculture and Forestry University, Fuzhou 350002, China and Col-
lege of Life Science, Fujian Agriculture and Forestry University, Fuzhou
350002, China; Tel: +86-591-83789443; Fax: +86-591-83789352;
E-mail: hehq3@yahoo.com.cn
#
These authors contributed equally to this work
these experiments are often time-consuming and costly [11].
Experimental difficulties in the large-scale identification of
protein phosphorylation sites have stimulated the develop-
ment of computational approaches to predict these sites from
protein sequences [12].
Over the past decade, a series of algorithms have been
developed to predict phosphorylation sites from amino acid
sequence. These approaches range from simple motif
searches to more complex machine learning methods, such
as Artificial Neural Networks (ANN) and Support Vector
Machines (SVM). Of the available resources, a few well-
maintained phosphorylation site prediction web sites have
been made freely available to the scientific community, in-
cluding NetPhos [2], NetPhosK [12], KinasePhos [13],
DISPHOS [14], Scansite [15], PPSP [16], GPS [17], Pred-
Phospho [18], and NetPhosYeast [19]. Generally, these
phosphorylation-site predictors can be grouped into two
categories. The first category of predictors focus on forecast-
ing whether a query Ser, Thr or Tyr residue is a phosphoryla-
tion site or not. This group includes the NetPhos and
DISPHOS. The second category of predictors that aim to
identify protein kinase-family-specific phosphorylation sites.
This category includes NetPhosK, PSSP, PredPhospho, and
KinasePhos. Comparatively, the kinase-family-specific
phosphorylation site prediction results are more useful, be-
cause the corresponding kinase family information for every
possible phosphorylation site can be inferred [13, 18]. How-
ever, this type of prediction may be suitable only for those

0929-8665/10 $55.00+.00 © 2010 Bentham Science Publishers Ltd.

Evaluation of Protein Phosphorylation Site Predictors
kinase families in which the number of known phosphoryla-
tion sites is large. Therefore, these kinase-family-specific
prediction systems are inevitably limited to only a few kinase
families [18]. There are currently two important directions in
the field of protein phosphorylation site prediction. With the
increase of experimentally validated phosphorylation sites,
the first direction involves development of organism-specific
predictors. Development of these predictors has been initi-
ated in some model organisms, specifically yeast [19] and
Arabidopsis [20]. The second direction involves combining
phosphorylation site prediction with a network context of
kinases and phosphoproteins. In addition to predicting
kinase-family-specific phosphorylation sites, such a method
would allow for the identification of specific kinases that are
responsible in vivo for the predicted phosphorylation events.
Existing phosphorylation site predictors play an increas-
ingly important role in complementing the experimental
characterization of protein phosphorylation. Generally, a
newly developed predictor must be intensively evaluated
before its acceptance for publication in a peer-reviewed
journal. Owing to the availability of many phosphorylation
site datasets, the accuracy of different predictors reported by
the developers may be overestimated, because these methods
were optimized predominantly for the corresponding training
and testing data. Moreover, the current protein phosphoryla-
tion site predictors are derived primarily from mammalian
phosphorylation site data and they may show lower predic-
tion accuracy for proteins from other organisms [19, 20]. To
our knowledge, different protein phosphorylation site predic-
tors have not been intensively benchmarked via some inde-
pendent datasets in the literature. Sikder and Zomaya (2009)
developed an independent dataset, collected from Phos-
pho.ELM, to compare the performances of KinasePhos, Net-
Phosk, DISPHOS, Scansite and PPSP [21]. However, this
independent dataset contained mainly mammalian proteins.
To explore the current phosphorylation site predictors’ appli-
cation to plant proteomes, we decided to quantitatively in-
vestigate their performance when applied to Arabidopsis or
rice phosphoproteins. To this end, we constructed three in-
dependent phosphorylation site datasets and evaluated the
performance of six phosphorylation site predictors on these
datasets. We believe that this work will provide scientists
Table 1. The Number of Positive and Negative Phosphorylation Sites in the Three Phosphorylation Site Datasets
Datasets Sites
(+) site a
rice
(-) site b
(+) site
Arabidopsis
(-) site
(+) site
mammal
(-) site
a
(+) site: Phosphorylation sites annotated by experimental results
b
(-) site: Non-phosphorylation sites
c
pSer: Ser phosphorylation sites
d
pThr: Thr phosphorylation sites
e
pTyr: Tyr phosphorylation sites
Protein & Peptide Letters, 2010, Vol. 17, No. 1 65
with a quantitative understanding of the overall performance
of the tested predictors. Moreover, we also point out some
existing problems regarding protein phosphorylation site
prediction.
MATERIALS AND METHODS
Datasets
We generated three phosphorylation site datasets consist-
ing of experimentally verified mammalian, Arabidopsis and
rice phosphoproteins, respectively. All the mammalian phos-
phoproteins in Phospho.ELM (release version 7.0), which
contains 4,026 protein entries covering 16,428 phosphoryla-
tion sites, were downloaded [22]. Three hundred mammalian
phosphoproteins were randomly selected from this total set
to compile the initial mammalian phosphoprotein dataset. To
remove sequence redundancy, the initial mammalian phos-
phoproteins were filtered using a 30% sequence identity fil-
ter. Briefly, we performed pairwise BLAST searches for any
sequence pairs in the initial mammalian phosphoprotein
dataset. For sequence pairs with sequence identity > 30%,
only one sequence was left in the dataset. Thus, the phos-
phoprotein dataset size was reduced to 231 total proteins.
The Arabidopsis and rice phosphorylation site datasets were
obtained from recent literature, the feature table of the Swiss-
Prot database (release 42.10) [23], and unpublished rice
phosphorylation data from our laboratory. As in the mam-
malian dataset, we filtered the Arabidopsis and rice phos-
phoproteins using a sequence identity of 30% and compiled
them into the testing datasets for Arabidopsis and rice.
The mammalian dataset consists of a total of 754 phos-
phoserine (pSer) sites, 135 phosphothreonine (pThr) sites
and 45 phosphotyrosine (pTyr) from 231 mammalian pro-
teins (Table 1). The Arabidopsis dataset contains a total of
153 pSer sites, 47 pThr sites and 11 pTyr sites from 98
Arabidopsis proteins. The rice dataset consists of a total of
359 pSer sites, 263 pThr sites and 1 pTyr site from 105 rice
proteins (Table 1). Since all of these phosphorylation sites
are experimentally verified, they are regarded as (+) sites.
The Ser, Thr and Tyr residues that are not annotated as phos-
phorylation sites within the above three datasets are regarded
pSer c pThr d pTyr e Total
359 263 1 623
1961 1216 740 3917
153 47 11 211
4603 2625 1457 8685
754 135 45 934
15709 9825 4267 29801
66 Protein & Peptide Letters, 2010, Vol. 17, No. 1
as ( ) sites (i.e., non-phosphorylation sites). The number of
( ) sites in each dataset is also listed in Table 1. We have
made the three phosphorylation site datasets freely available
at http://210.34.95.200/hlx/, to allow other researchers to
benchmark their methods against our results.
Six Tested Phosphorylation Site Predictors
Each sequence from the three independent datasets was
individually run through all six methods (KinasePhos, Net-
PhosK, DISPHOS, Scansite, PPSP and PredPhospho) using
the web site servers provided by the developers. Aside from
DISPHOS, all the tested predictors belong to kinase-specific
programs. To allow for a fair comparison of these six predic-
tors, we evaluate only whether a candidate site is correctly
predicted. The correctness of the predicted kinase family
information is not considered in our analysis. The algorithms
and parameter settings of the evaluated predictors are briefly
described below.
DISPHOS (DISorder-enhanced PHOSphorylation site
predictor, http://core.ist.temple.edu/pred/) uses position-
specific amino acid composition and predicted structural
disorder information to distinguish phosphorylation and non-
phosphorylation sites. For our work, all the sequences in the
testing datasets were analyzed using DISPHOS (version 1.3).
Since DISPHOS provides the option of different kingdoms
and organisms, the “Eukaryotes” option was selected for
predicting mammalian and rice proteins, and the “A.
thaliana” option was selected for Arabidopsis proteins.
NetPhosK (http://www.cbs.dtu.dk/services/NetPhosK/)
uses an artificial neural network algorithm to predict 17
kinase-family-specific phosphorylation sites. In this work, all
the tested proteins were analyzed with the NetPhosK 1.0
server without a filtering method, and a threshold of 0.5 was
used to judge whether a site was predicted as phosphory-
lated.
KinasePhos (http://kinasephos.mbc.nctu.edu.tw/index.
php) employs a Profile Hidden Markov Model (HMM) to
predict kinase-family-specific phosphorylation sites. In this
evaluation, KinasePhos was run with the option of 100%
prediction specificity. No specific kinase was set, since we
evaluated only the predicted generic phosphorylation sites.
Scansite (http://scansite.mit.edu/) uses scores calculated
from position-specific score matrices (PSSM) to search for
motifs within proteins that are likely to be phosphorylated by
specific protein kinases. In this work, Scansite2.0 was run by
searching all motifs and the “high stringency level” setting
was selected.
PPSP (http://bioinformatics.lcd-ustc.org/PPSP/) utilizes
Bayesian Decision Theory (BDT) to predict phosphorylation
sites for approximately 70 kinase groups. In this work, pre-
diction was performed by considering all kinase groups and
the balance performance option was selected.
PredPhospho (http://pred.ngri.re.kr/PredPhospho.htm)
predicts various kinase-specific phosphorylation sites by
training SVMs. Due to the limitations of phosphorylated
sequence data in public databases, the current PredPhospho
can make classifiers for only four kinase groups and four
Que et al.
kinase families. In this evaluation, prediction was made by
considering all kinase groups and families.
Performance Assessment
Four measurements—Sensitivity (Sn), Specificity (Sp),
Accuracy (Ac) and the Matthews correlation coefficient
(MCC) — were employed to evaluate the performance of the
tested predictors (definitions below):
TP ,
Sn =
TP + FN
TN ,
Sp =
TN + FP
TP + TN ,
Ac =
TP + FP + TN + FN
and
(TP TN ) ( FN FP) .
MCC =
(TP + FN ) (TN + FP) (TP + FP) (TN + FN )
where TP, FP, FN, and TN denote true positives, false posi-
tives, false negatives, and true negatives. Sn and Sp illustrate
the correct prediction ratios of positive and negative data
sets, respectively. When the number of positive data and
negative data differ too much from each other, the MCC
value is calculated to assess the prediction performance. The
value of MCC ranges from -1 to 1, and larger values indicate
better prediction performance.
RESULTS AND DISCUSSION
The Overall Performance of the Six Tested Predictors
For the mammalian dataset, the prediction performance
of the different methods indicated the following trend: Pred-
Phospho>DISPHOS>Scansite>NetPhosK>PPSP>Kinase
Phos (Table 2). PredPhospho demonstrated the best perform-
ance, with an MCC value of 0.1971, and KinasePhos yielded
the poorest prediction (MCC = 0.0076). When tested on the
Arabidopsis and rice datasets, the performance ranking of the
different predictors showed a trend that was similar to that of
the mammalian dataset. For the Arabidopsis dataset, the pre-
diction performance of PredPhospho demonstrated the high-
est MCC (0.1141), DISPHOS was second (MCC = 0.0976),
NetphosK third (MCC = 0.052), PPSP fourth (MCC =
0.0201), Scansite fifth (MCC = 0.0199), and KinasePhos
yielded the lowest MCC (-0.0095). For the rice dataset,
PredPhospho again yielded the highest MCC (0.1061),
whereas KinasePhos resulted in the lowest MCC (-0.0573).
Compared with the other five predictors, PredPhospho is the
most favorable algorithm for phosphorylation site prediction
in mammalian, Arabidopsis and rice proteins.
Generally, all six evaluated phosphorylation site predic-
tors showed poor prediction performance. Using the predic-
tion of mammalian proteins as an example, the MCC values
for all the tested predictors remained below 0.2, suggesting
that these predictors alone were not sufficient for practical
use. The most critical issue contributing to low MCC values
in our study is that the number of non-phosphorylated Ser,
Thr or Tyr residues in a protein is far larger than that of
Evaluation of Protein Phosphorylation Site Predictors Protein & Peptide Letters, 2010, Vol. 17, No. 1 67

Table 2. The Overall Predictive Performance for the Phosphorylation Site Datasets

Tools DATASET Sn Sp Ac MCC

rice 94.86 2.4 15.09 -0.0573

KinasePhos Arabidopsis 95.73 3.17 5.36 -0.0095

mammalian 98.93 1.62 4.58 0.0076

rice 38.04 63 59.56 0.0075

DISPHOS Arabidopsis 63.98 66.47 66.41 0.0976

mammalian 86.51 61.42 62.18 0.1679

rice 59.07 52.41 53.33 0.079

NetPhosK Arabidopsis 66.82 50.26 50.65 0.052

mammalian 82.33 43.63 44.81 0.0901

rice 93.42 4.19 16.43 -0.0397

PPSP Arabidopsis 97.16 5.95 8.12 0.0201

mammalian 98.93 8.03 10.8 0.0445

rice 13.48 87.13 77.03 0.0063

Scansite Arabidopsis 18.48 86.06 84.45 0.0199

mammalian 42.08 85.08 83.78 0.128

rice 25.33 88.44 84.03 0.1061

PredPhospho Arabidopsis 59.61 77.01 76.7 0.1141

mammalian 78.88 77.65 77.68 0.1971

phosphorylated sites, inducing an imbalance of ( ) sites and
(+) sites. Such a phenomenon is further exemplified in the
prediction performance of PredPhospho. When tested in
mammalian proteins, PredPhospho performed reasonably
well (Sn = 78.88% and Sp = 77.65%). However, the ratio of
(+) sites to ( ) sites in the mammalian dataset is approxi-
mately 1:32, indicating that the prediction result of Pred-
Phospho indeed contains too many false positives, resulting
in a relatively low MCC value (0.1971). For a machine
learning method-based phosphorylation site predictor, it is
important to note that an optimized ratio of (+) sites to ( )
sites in the training dataset can often result in a better predic-
tion model.
The Prediction Performance of pSer, pThr and pTyr
Sites
The tested predictors demonstrated different performance
on pSer, pThr and pTyr site predictions (Fig. 1). Generally,
the predictors have similar performance on pSer and pThr
sites and better predictive performance on pTyr site. Because
the number of pTyr sites is limited in all the testing datasets,
we can not conclude that the prediction of pTyr sites is easier
than the prediction of pSer and pThr sites. In terms of pSer
and pThr site prediction on the three datasets, all the tested
predictors indicated better performance (MCC) on the mam-
malian dataset than on either plant (Arabidopsis and rice)
dataset (Fig. 1). However, for pTyr site prediction, Kinase-
Phos, PredPhospho, DISPHOS and NetPhosK showed the
highest MCC on the rice dataset, while the other two predic-
tors still demonstrated their highest MCC values on the
mammalian dataset. Since the number of pTyr sites in the
rice dataset is very limited, we are unable to conclude that
these four methods are favorable for pTyr site prediction.
The Tested Predictors Exhibited Better Performance on
Mammalian Proteins than Plant Proteins
Each predictor exhibited the best performance (i.e., the
highest MCC) on the mammalian dataset and the poorest
performance (i.e., the lowest MCC) on either the rice or
Arabidopsis dataset (Table 2). It is likely that this is caused
by the predictors having been derived primarily from mam-
malian phosphorylation sites. We conclude that the tested
predictors might not adapt well to predicting plant protein
phosphorylation sites. Therefore, a plant-specific predictor
may yield better performance when applied to plant proteins
than a generic predictor. For instance, a plant-specific pre-
dictor called PhosPhAt has been developed to predict pSer
sites in Arabidopsis. PhosPhAt was benchmarked to reveal
better performance than some generic predictors [20]. With
the accumulation of experimentally validated phosphoryla-
tion sites in rice, it is hoped that a rice-specific predictor will
be available in the near future. Such a tool will inevitably
facilitate more reliable computational annotation of phos-
phorylation sites in rice phosphoproteins.
68 Protein & Peptide Letters, 2010, Vol. 17, No. 1
Figure 1. The predictive performance (MCC) of the six tested pre-
dictors on pSer (A), pThr (B) and pTyr (C) sites.
The Complementarity Between Different Predictors
We found that different predictors are often complemen-
tary to some extent. For example, the prediction performance
of KinasePhos was 98.93% (Sn) and 1.62% (Sp) on the
mammalian dataset and outperformed Scansite (Sn = 42.08%
and Sp = 85.08%) with a 56.85% higher sensitivity and an
83.46% lower specificity. This suggested that an effective
combination of KinasePhos and Scansite may result in a new
prediction system with increased performance. Indeed, such
an effort has recently been initiated. Wan et al. (2008) used a
Que et al.
weighted voting method to incorporate 15 phosphorylation-
site predictors into a meta-predictor, the performance of
which can exceed that of all individual predictors [24].
FUTURE PERSPECTIVES
In statistical predictions, three cross-validation methods
are often used to examine a predictor for its effectiveness in
a practical application: the independent dataset test, the sub-
sampling test, and the jackknife test [25]. However, as dem-
onstrated by Chou and Shen [26, 27], the jackknife test is
deemed the most objective of these three tests and can al-
ways yield a unique result for a given benchmark dataset.
For these reasons, the jackknife test has been increasingly
used by investigators to examine the accuracy of various
predictors [28-35]. However, a feasible way to examine the
quality of the existing web-site predictors is the independent
dataset test, as performed in this study.
In this study, we constructed three phosphorylation site
datasets (i.e., the datasets of mammal, Arabidopsis and rice
proteins) to benchmark the prediction performance of six
phosphorylation site predictors. It is worth mentioning that
the three testing datasets can not be regarded as the gold
standard. In particular, the unreliable (-) sites may impact the
evaluation. Some phosphorylation sites that have not been
experimentally validated are likely to be represented as (-)
sites in this study, contributing to the “noise” of our evalua-
tion. Moreover, the size of these three datasets may not be
large enough to represent the complete sequence space of
phosphoproteins. Despite these shortcomings, this evaluation
provides a quantitative understanding of the overall perform-
ance of current phosphorylation-site predictors.
We expect that both the developers of phosphorylation
site predictors and the interested experimental scientists will
benefit from our findings in this work. We would like to em-
phasize the following findings we have obtained. First, the
current phosphorylation site predictors are not effective for
practical use. In particular, the current predictors often yield
overprediction at the whole-sequence level (i.e., a lower frac-
tion of the predicted sites are actually true positives). There-
fore, reduction of overprediction remains an important direc-
tion for development of new phosphorylation-site predictors.
Second, the current predictors perform less well when they
are applied to predicting phosphorylation sites in Arabidop-
sis and rice phosphoproteins. Therefore, it is important to
develop organism-specific phosphorylation site predictors.
With the accumulated phosphorylation site data available in
many model organisms, such predictors have recently been
developed. Finally, we have found that the tested predictors
are complementary to some extent, suggesting that estab-
lishment of a meta-server may result in better predictive per-
formance.
ACKNOWLEDGEMENTS
This study was supported in part by grant 20070019050
of the State Education Ministry, China, the Key Science Pro-
gram of Fujian (no. 2009N0011), the National Natural Sci-
ence Foundation of China (no. 30020170), and the Key Pro-
gram of Ecology, Fujian Province (0608507).
Evaluation of Protein Phosphorylation Site Predictors
REFERENCES
[1] Peck, S.C. Early phosphorylation events in biotic stress. Curr.
Opin. Plant Biol., 2003, 6(4), 334-338.
[2] Blom, N.; Gammeltoft, S.; Brunak, S. Sequence and structure-
based prediction of eukaryotic protein phosphorylation sites. J.
Mol. Biol., 1999, 294(5), 1351-1362.
[3] Chou, K.C.; Watenpaugh, K.D.; Heinrikson, R.L. A model of the
complex between cyclin-dependent kinase 5(Cdk5) and the
activation domain of neuronal Cdk5 activator. Biochem. Biophys.
Res. Commun., 1999, 259, 420-428.
[4] Chou, K.C. Review: Structural bioinformatics and its impact to
biomedical science. Curr. Med. Chem., 2004, 11, 2105-2134.
[5] Hubbard, M.J.; Cohen, P. On target with a new mechanism for the
regulation of protein phosphorylation. Trends Biochem Sci., 1993,
18(5), 172-177.
[6] Khan, M.M.; Jan, A.; Karibe, H.; Komatsu, S. Identification of
phosphoproteins regulated by gibberellin in rice leaf sheath. Plant
Mol. Biol., 2005, 58(1), 27-40.
[7] Ficarro, S.B.; McCleland, M.L.; Stukenberg, P.T.; Burke, D.J.;
Ross, M.M.; Shabanowitz, J.; Hunt, D.F.; White, F.M.
Phosphoproteome analysis by mass spectrometry and its
application to Saccharomyces cerevisiae. Nat. Biotechnol., 2002,
20(3), 301-305.
[8] Ballif, B.A.; Villen, J.; Beausoleil, S.A.; Schwartz, D.; Gygi, S.P.
Phosphoproteomic analysis of the developing mouse brain. Mol.
Cell Proteomics, 2004, 3(11), 1093-1101.
[9] Lim, Y.P.; Diong, L.S.; Qi, R.; Druker, B.J.; Epstein, R.J.
Phosphoproteomic fingerprinting of epidermal growth factor
signaling and anticancer drug action in human tumor cells. Mol.
Cancer Ther., 2003, 2(12), 1369-1377.
[10] Nuhse, T.S.; Stensballe, A.; Jensen, O.N.; Peck, S.C. Phospho-
proteomics of the Arabidopsis plasma membrane and a new
phosphorylation site database. Plant Cell, 2004, 16(9), 2394-2405.
[11] Hjerrild, M.; Gammeltoft, S. Phosphoproteomics toolbox: compu-
tational biology, protein chemistry and mass spectrometry. FEBS
Lett., 2006, 580(20), 4764-4770.
[12] Blom, N.; Sicheritz-Ponten, T.; Gupta, R.; Gammeltoft, S.; Brunak,
S. Prediction of post-translational glycosylation and phospho-
rylation of proteins from the amino acid sequence. Proteomics,
2004, 4(6), 1633-1649.
[13] Huang, H.D.; Lee, T.Y.; Tzeng, S.W.; Horng, J.T. KinasePhos: a
web tool for identifying protein kinase-specific phosphorylation
sites. Nucleic Acids Res., 2005, 33(Web Server issue), W226-229.
[14] Iakoucheva, L.M.; Radivojac, P.; Brown, C.J.; O'Connor, T.R.;
Sikes, J.G.; Obradovic, Z.; Dunker, A.K. The importance of
intrinsic disorder for protein phosphorylation. Nucleic Acids Res.,
2004, 32(3), 1037-1049.
[15] Obenauer, J.C.; Cantley, L.C.; Yaffe, M.B. Scansite 2.0: Proteome-
wide prediction of cell signaling interactions using short sequence
motifs. Nucleic Acids Res., 2003, 31(13), 3635-3641.
[16] Xue, Y.; Li, A.; Wang, L.; Feng, H.; Yao, X. PPSP: prediction of
PK-specific phosphorylation site with Bayesian decision theory.
BMC Bioinformatics, 2006, 7, 163.
[17] Xue, Y.; Zhou, F.; Zhu, M.; Ahmed, K.; Chen, G.; Yao, X. GPS: a
comprehensive www server for phosphorylation sites prediction.
Nucleic Acids Res., 2005, 33(Web Server issue), W184-187.
[18] Kim, J.H.; Lee, J.; Oh, B.; Kimm, K.; Koh, I. Prediction of
phosphorylation sites using SVMs. Bioinformatics, 2004, 20(17),
3179-3184.
Received: March 05, 2009 Revised: May 23, 2009 Accepted: May 27, 2009
Protein & Peptide Letters, 2010, Vol. 17, No. 1 69
[19] Ingrell, C.R.; Miller, M.L.; Jensen, O.N.; Blom, N. NetPhosYeast:
prediction of protein phosphorylation sites in yeast. Bioinformatics,
2007, 23(7), 895-897.
[20] Heazlewood, J.L.; Durek, P.; Hummel, J.; Selbig, J.; Weckwerth,
W.; Walther, D.; Schulze, W.X. PhosPhAt: a database of
phosphorylation sites in Arabidopsis thaliana and a plant-specific
phosphorylation site predictor. Nucleic Acids Res., 2008,
36(Database issue), D1015-1021.
[21] Sikder, A.R.; Zomaya, A.Y. Analysis of protein phosphorylation
site predictors with an independent dataset. Int. J. Bioinform. Res.
Appl., 2009, 5(1), 20-37
[22] Diella, F.; Cameron, S.; Gemund, C.; Linding, R.; Via, A.; Kuster,
B.; Sicheritz-Ponten, T.; Blom, N.; Gibson, T.J. Phospho.ELM: a
database of experimentally verified phosphorylation sites in
eukaryotic proteins. BMC Bioinformatics, 2004, 5, 79.
[23] Boutet, E.; Lieberherr, D.; Tognolli, M.; Schneider, M.; Bairoch,
A. UniProtKB/Swiss-Prot. Methods Mol. Biol., 2007, 406, 89-112.
[24] Wan, J.; Kang, S.; Tang, C.; Yan, J.; Ren, Y.; Liu, J.; Gao, X.;
Banerjee, A.; Ellis, L.B.; Li, T. Meta-prediction of phosphorylation
sites with weighted voting and restricted grid search parameter
selection. Nucleic Acids Res., 2008, 36(4), e22.
[25] Chou, K.C.; Zhang, C.T. Review: Prediction of protein structural
classes. Crit. Rev. Biochem. Mol. Biol., 1995, 30, 275-349.
[26] Chou, K.C.; Shen, H.B. Cell-PLoc: A package of web-servers for
predicting subcellular localization of proteins in various organisms.
Nat. Protocols, 2008, 3, 153-162.
[27] Chou, K.C.; Shen, H.B. Review: Recent progresses in protein
subcellular location prediction. Analy. Biochem., 2007, 370, 1-16.
[28] Xiao, X.; Chou, K.C. Digital coding of amino acids based on
hydrophobic index. Protein Pept. Lett., 2007, 14, 871-875.
[29] Munteanu, C.B.; Gonzalez-Diaz, H.; Magalhaes, A.L. Enzymes/
non-enzymes classification model complexity based on
composition, sequence, 3D and topological indices. J. Theor. Biol.,
2008, 254, 476-482.
[30] Munteanu, C.R.; Gonzalez-Diaz, H.; Borges, F.; de Magalhaes,
A.L. Natural/random protein classification models based on star
network topological indices. J. Theor. Biol., 2008, 254, 775-783.
[31] Rezaei, M.A.; Abdolmaleki, P.; Karami, Z.; Asadabadi, E.B.;
Sherafat, M.A.; Abrishami-Moghaddam, H.; Fadaie, M.;
Forouzanfar, M. Prediction of membrane protein types by means of
wavelet analysis and cascaded neural networks. J. Theor. Biol.,
2008, 254, 817-820.
[32] Kannan, S.; Hauth, A.M.; Burger, G. Function prediction of
hypothetical proteins without sequence similarity to proteins of
known function. Protein Pept. Lett., 2008, 15, 1107-1116.
[33] Chen, C.; Chen, L.; Zou, X.; Cai, P. Prediction of protein
secondary structure content by using the concept of chou's pseudo
amino acid composition and support vector machine. Protein Pept.
Lett., 2009, 16, 27-31.
[34] Shen, H.B.; Chou, K.C. QuatIdent: A web server for identifying
protein quaternary structural attribute by fusing functional domain
and sequential evolution information. J. Proteome Res., 2009, 8,
1577-1584.
[35] Xiao, X.; Wang, P.; Chou, K.C. Predicting protein quaternary
structural attribute by hybridizing functional domain composition
and pseudo amino acid composition. J. Appl. Crystallogr., 2009,
42, 169-173.