Bioinformatics in China: a personal perspective.

Full text

Page 1

Perspective
Bioinformatics in China: A Personal Perspective
Liping Wei1*, Jun Yu2*
1Center for Bioinformatics, National Laboratory of Protein Engineering and Plant Genetic Engineering, College of Life Sciences, Peking University, Beijing, People’s
Republic of China, 2CAS Key Laboratory in Genome Sciences and Information, Beijing Institute for Genomics, Chinese Academy of Sciences, Beijing, People’s Republic of
China
In this personal perspective, we recall
the history of bioinformatics and compu-
tational biology in China, review current
research and education, and discuss future
prospects and challenges. The field of
bioinformatics in China has grown signif-
icantly in the past decade despite a delayed
and patchy start at the end of the 1980s by
a few scientists from other disciplines, most
noticeably physics
where China’s traditional strength has
been. In the late 1990s and early 2000s,
rapid expansion of the field was fueled by
the Internet boom and genomics boom
worldwide and in China. Today bioinfor-
matics research in China is characterized
by a great variety of biological questions
addressed and the close collaborative
efforts between computational scientists
and biologists, with a full spectrum of
focuses ranging from database building
and algorithm development to hypothesis
generation and biological discoveries. Al-
though challenges remain, the future of
bioinformatics in China is promising
thanks to advances in both computing
infrastructure and experimental biology
research, a steady increase of governmen-
tal funding, and most importantly a
critical mass of bioinformatics scientists
consisting of not only converts from other
disciplines but also formally trained over-
seas returnees and a new generation of
domestically trained
Ph.D.s.
andmathematics,
bioinformatics
Introduction
The field of bioinformatics has enjoyed
significant growth in China. A rough yet
useful indicator of the field is the number
of bioinformatics and computational biol-
ogy publications from China indexed in
PubMed at NCBI. As shown in Figure 1A,
this number has been increasing signifi-
cantly over the past decade. The number
of all publications from China indexed in
PubMed hasalso
(Figure 1B), but if we plot the percentage
of bioinformatics publications from China
among all PubMed publications from
China, we observe that the percentage
itselfhasbeen
(Figure 1C), indicating that the contribu-
been increasing
increasingrapidly
tion of bioinformatics research is becom-
ing more significant within the life sciences
in China. Comparing it with the situation
worldwide, we observe that the number of
bioinformatics publications from China is
growing faster than the number of bioin-
formaticspublications
(Figure 1A versus 1D). Additionally, the
number of PubMed publications from
China has also been growing faster than
the total number of PubMed publications
(Figure 1B versus 1E). Furthermore, we
observe that, very interestingly, for each
year starting from 2003, the percentage of
bioinformaticspublications
PubMed publications is greater in China
than worldwide (Figure 1C versus 1F),
indicating that bioinformatics as a field
within the life sciences is enjoying a faster
growth rate in China than elsewhere in the
world.
These numbers and trends, although
only rough estimates, are fascinating and
deserve a closer look. In this article, we
(see Box 1 for Authors’ Biographies) take a
historical, as well as a horizontal, survey of
bioinformatics in China. Due to the nature
of personal perspectives and space limita-
tions, what we present is by no means
exhaustive, but merely suggestive.
worldwide
among all
The Early Years
China had a late and rather sluggish
start in bioinformatics, largely due to lack
of institutional support for bioinformatics
and governmental funding during the
early years. It was not until 1996 that the
first Center for Bioinformatics in China
was established within the College of Life
Sciences at Peking University and the first
government funding dedicated to bioinfor-
matics was established as part of the
National High Technology Development
863 Program by the Ministry of Science
and Technology (MOST). Despite the
difficulties, starting from the end of the
1980s bioinformatics research was pio-
neered by a few scientists from other
disciplines, most noticeably physics and
mathematics where China’s traditional
strength has been, applying theoretical
frameworks and analytical tools from their
original specialty to study biological ques-
tions.
A great example of these early scientists
was Bailin Hao, who, trained in the former
Soviet Union, was at the time already an
accomplished theoretical physicist. Fasci-
nated by computable and predictable
characteristics of biological systems, he
first studied protein structures and then
devoted himself to the analysis of DNA
and protein sequences. In visualizing very
long DNA sequences, including the com-
plete genomes of several bacteria and yeast
and segments of human genome, his group
observed fractal-like patterns [1]; subse-
quently they proposed the description of a
genome using statistics of K-strings (by
counting a series of K-tuples in a linear
sequence of symbols), enabling a new
paradigm for phylogenetic analysis with-
out sequence alignments [2]. Chunting
Zhang, another established physicist, also
started his bioinformatics research at the
end of the 1980s working on protein
structure prediction [3]. His group later
created the concept of Z-curve to display
DNA composition dynamics in geometric
spaces [4,5]. Z-curve has been successfully
applied to genome sequence analyses such
as prediction of coding genes in yeast
genome [6], isochores in human genome
[7], and replication origins in archaeal
genomes [8]. Runsheng Chen, with a
background in biophysics, was involved
in the early phases of Chinese genomics
research and later pioneered small RNA
Citation: Wei L, Yu J (2008) Bioinformatics in China: A Personal Perspective. PLoS Comput Biol 4(4): e1000020.
doi:10.1371/journal.pcbi.1000020
Published April 25, 2008
Copyright: ? 2008 Wei, Yu. This is an open-access article distributed under the terms of the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium,
provided the original author and source are credited.
* E-mail: weilp@mail.cbi.pku.edu.cn (LW); junyu@genomics.org.cn (JY)
Editor: Philip E. Bourne, University of California San Diego, United States of America
PLoS Computational Biology | www.ploscompbiol.org1April 2008 | Volume 4 | Issue 4 | e1000020

Page 2

research in China [9–17]. His group
identified hundreds of novel small non-
coding RNAs in the nematode C. elegans
and characterized their sequence features
and expression patterns [11,13]. A few
other early scientists with important con-
tributions include Yanda Li [18–24],
Luhua Lai [25–34], Yunyu Shi [35],
Liaofu Luo [36–40], Dafu Ding [41–43],
and Zhirong Sun [44–48].
In the early to mid-1990s, many
Chinese biologists were still not familiar
with international biological databases,
and some in remote cities did not even
have reliable Internet connections to
overseas Web sites. To promote bioinfor-
matics as well as biology research, since
1995 Jingchu Luo and Xiaocheng Gu at
the Center for Bioinformatics, Peking
University, dedicated themselves to setting
up official mirror sites of major biological
databases, providing bioinformatics servic-
es, and organizing bioinformatics training
workshops. From them, many Chinese
scientists received their first exposure to
biologicaldata and
Through close collaboration with the
NCBI, EBI, and EMBnet, Luo’s group
continues to maintain the largest online
bioinformaticsresource
http://www.cbi.pku.edu.cn.
analysis tools.
inChina at
The Internet Boom and the
Genomic Boom
Two significant developments in the
second half of the 1990s proved critical to
the expansion of bioinformatics in China.
They were the Internet boom and the
genomic boom. It was not until 1994 that
full TCP/IP Internet connection between
China and the rest of the world was
established. At that time, the only estab-
lished connection was an existing link, a
dial-up X.25 connection, between the
Institute of High-Energy Physics (IHEP)
in Beijing and the Stanford Linear Accel-
erator Center (SLAC) at Stanford Univer-
sity. On May 17, 1994, the official
connection to FIX-West was announced,
andthe U.S.-based
Network (ESnet) agreed to carry China
IP traffic. Despite the late start, China
caught up quickly with the worldwide
Internet boom and established fast and
broad Internet connections which were
critical for biological data exchange.
At about the same time, the genomic
boom in the late 1990s to early 2000s
exemplifiedbythe
Project [49] started to generate huge and
exponentially growing amounts of data.
After earlier efforts such as sequencing of
Energy Sciences
HumanGenome
full-length cDNAs [50], Chinese scientists
officially participated in the Human Ge-
nome Project in 1999 and sequenced 1%
of the human genome [9,49]. Since then
the Chinese genome centers have inde-
pendently sequenced several large ge-
nomes including those of rice (Oryza sativa
L. ssp. indica) [51], silkworm (Bombyx mori)
[52], and numerous microbial genomes
such as that of Thermoanaerobacter tengcon-
gensis, a rod-shaped, gram-negative, anaer-
obic eubacterium isolated from a freshwa-
ter hot spring in Tengchong, China [53].
Data is a major driving force for
bioinformatics—it demands new methods
for data storage and analyses and motivated
the computational discovery of biological
patterns. In all genomic projects, bioinfor-
matics teams played significant roles in data
analysis.Databaseswerecreated tostorethe
genomic data, for instance BGI-RIS [54]
and SilkDB [55] for rice and silkworm
genomic sequence data and related biolog-
ical information, and ChickVD [56] for
chickengenomicsequence
Methods and software packages were de-
veloped to analyze the huge amount of
genomic data, including for instance BGF,
an ab initio gene prediction method
developed for the rice genome based on
Hidden Markov Model (HMM) and dy-
namic programming [57], a nonsupervised
gene prediction algorithm for bacterial and
archaeal genomes [58], and a method for
genome comparison using Gene Ontology
(GO) with statistical testing [59].
variations.
Present Day: A Diversified
Landscape
Since the end of the 1990s, bioinfor-
matics in China has experienced signifi-
cant growth fueled by not only the
aforementioned Internet and genomic
boom but also by official institutional
support, steadily increasing government
funding, and very importantly, a growing
number of Chinese scientists returning to
China after formal overseas training and
research experience in bioinformatics and
genomics. Today the field is characterized
by a great variety of biological questions
addressed and the close collaborations
between computational
bench biologists, with a full spectrum of
focuses ranging from providing resources
and services, developing databases and
algorithms, to generating hypotheses and
making biological discoveries.
Providing biological data resources and
services is still an important part of
bioinformatics. A good example is the
Shanghai Center for Bioinformation Tech-
nology (http://www.scbit.org/), led by
scientistsand
Box 1. Authors’ Biographies
Liping Wei is Professor and Director of the Center for Bioinformatics at Peking
University and Associate Director of the National Laboratory of Protein
Engineering and Plant Genetic Engineering in China. She received her
undergraduate training in Electrical Engineering and Information Sciences from
the University of Science and Technology of China. She holds a Master’s degree in
Applied Mathematics from Brown University and a PhD in Medical Informatics
(now called Biomedical Informatics) from Stanford University. She worked in the
biotech industry for four years, first at Exelixis, Inc., and then at Nexus Genomics,
Inc., while continuing to serve as a Consulting Assistant Professor of bioinfor-
matics in the Department of Medicine at Stanford University. She moved back to
China in early 2004 to resume an academic career in bioinformatics at Peking
University. Her current research interests include the regulation by, and of,
noncoding RNAs and antisense transcripts and the signaling and regulatory
networks underlying neurobiological disorders such as autism and addiction.
Jun Yu is a professor in the Beijing Institute of Genomics, Chinese Academy of
Sciences (CAS). Dr. Yu obtained a B.S. degree in biochemistry from Jilin University
in 1983 and a PhD in biomedical sciences from New York University School of
Medicine in 1990. He joined the University of Washington Genome Center in 1993
and attuned his primary research interests toward genomics and bioinformatics.
He started to work in China in 1998 and has led many major genome projects
there, such as the International Human Genome Project (the Chinese effort), the
Superhybrid Rice Genome Project, and the Silkworm Genome Project. His current
research interests include comparative genome analysis, transcriptome modeling,
human genetic diversity, and model organisms for phenotypic plasticity. He has
published more than 120 scientific papers and won numerous academic awards,
such as Scientific Leader of the Year, 2002 (Scientific American), 100-Talent Plan
(CAS, 2002–2005), and China–US Biology Examination and Application (CUSBEA,
1983).
Bioinformatics in China
PLoS Computational Biology | www.ploscompbiol.org2April 2008 | Volume 4 | Issue 4 | e1000020

Page 3

Yixue Li and founded in 2002. It has served
as the central repository for biological data
generated throughout Shanghai and neigh-
boring areas. It has also provided extensive
public online bioinformatics resources. Fur-
thermore, many new primary databases
with raw data and value-added secondary
databaseswerecreated byChinesescientists
(Table 1), and many locally developed
methods were implemented and incorpo-
rated into Web servers (Table 2). These
databases and tools address a broad range
of biological questions and are open to the
entire international scientific community,
and are enjoyed by a large worldwide user
base.
Independently or in collaboration with
bench scientists, bioinformatics scientists
have played significant roles in important
biological discoveries. In 2003, China and
many other countries suffered from the
severe acute respiratory syndrome (SARS)
epidemic. A team of bioinformaticians
quickly joined the Chinese SARS Molec-
ular Epidemiology Consortium, led by
Guo-Ping Zhao, to sequence and analyze
SARS coronavirus genomic sequences
Figure 1. The number and percentage of bioinformatics publications from China and in all of PubMed in the past decade. (A) The
number of bioinformatics and computational biology publications from China in PubMed, retrieved from NCBI by the Entrez query ‘‘China[affiliation]
AND (bioinformatics OR computational biology)’’. MESH has a heading ‘‘Computational Biology’’ of which ‘‘Bioinformatics’’ is an Entry Term. The
numbers provide a rough indication of growth of the field. They tend to be underestimated because some bioinformatics publications cannot be
retrieved by keywords ‘‘bioinformatics OR computational biology’’. However adding more keywords would increase the false positive rate
significantly. (B) The number of all publications from China in PubMed, retrieved from NCBI by the Entrez query ‘‘China[affiliation]’’. (C) The
percentage of bioinformatics publications in China among all publications in China, in PubMed. (D) The number of bioinformatics and computational
biology publications in all of PubMed, retrieved from NCBI by the Entrez query ‘‘bioinformatics OR computational biology’’. (E) The number of all
publications in PubMed. (F) The percentage of bioinformatics publications among all publications, in PubMed.
doi:10.1371/journal.pcbi.1000020.g001
Bioinformatics in China
PLoS Computational Biology | www.ploscompbiol.org3April 2008 | Volume 4 | Issue 4 | e1000020

Page 4

derived from the early, middle, and late
phases of the SARS epidemic as well as
viral sequences from palm civets. Their
work uncovered the molecular evolution
of the SARS coronavirus and the viral
invasion from animal to human [60,61].
Another group proposed a mathematical
model to estimate the evolution rate of the
SARS coronavirus genome (0.16 base/
day) and the time of the last common
ancestor of the sequenced SARS strains
(August or September of 2002) [62]. To
identify potential anti-viral drug targets,
proteomic and bioinformatic methods
were used to investigate key SARS viral
proteins including structural proteins [63],
spike protein [64], and 3C-like proteinase
[30]. Genomic packaging signals, which
may be used to design antisense RNA and
RNA interfere (RNAi) drugs treating
SARS, were predicted by comparative
genomics methods [65]. The three-dimen-
sional structure of 3C-like proteinase was
constructed by homology modeling, based
on which virtual screening of chemical
databases was performed to search for
potential inhibitors [31]. Chinese SARS
patients were statistically studied to eluci-
date the association of symptom combina-
tions with different outcome and thera-
peutic effects [66] and the association
betweenmannose-binding lectingene
polymorphisms and susceptibility to SARS
virus infection [67].
The large Chinese population provided
invaluable resources for genetic studies.
Patient and control groups were geno-
typed to establish associations between
genetic variations and susceptibility to
diseases such as esophageal squamous cell
carcinoma [68], hepatocellular carcinoma
in a particularly high-risk region of China
[69], hypertension [70], and coronary
heart disease [71]. A link between an
Asian-enriched SNP in human sialidase
and severe adverse reactions to the anti-
viral drug Tamiflu (oseltamivir carboxyl-
ate) was proposed and confirmed by
Table 1. Examples of biological databases developed and maintained in China.
Category NameMain contentURLReference
Genome resources BGI-RIS An integrated information resource and
comparative analysis workbench for rice genomics
http://rise.genomics.org.cn/[54]
SilkDBA knowledgebase for silkworm biology and
genomics
http://silkworm.genomics.org.cn[55]
ChickVDA sequence variation database for the chicken
genome
http://chicken.genomics.org.cn[56]
The Z curve database A graphic representation of genome sequenceshttp://tubic.tju.edu.cn/zcurve/[148]
MED-Start Predicted translation initiation sites in microbial
genomes
http://ctb.pku.edu.cn/main/SheGroup/
MED_Start.htm
[100]
ProTISA A comprehensive resource for translation initiation
site annotation in prokaryotic genomes
http://mech.ctb.pku.edu.cn/protisa[149]
DoriC A database of oriC regions in bacterial genomes http://tubic.tju.edu.cn/doric/[150]
Transcription
regulation resources
DRTFA database of rice transcription factorshttp://drtf.cbi.pku.edu.cn/[97]
DATFA database of Arabidopsis transcription factorshttp://datf.cbi.pku.edu.cn[98]
NATsDBA database of natural antisense transcripts
identified in ten genomes
http://natsdb.cbi.pku.edu.cn/[151]
NONCODEAn integrated knowledge database of noncoding
RNAs
http://noncode.bioinfo.org.cn[13,152]
NPInterA database of noncoding RNA and protein
interaction.
http://bioinfo.ibp.ac.cn/NPInter [14]
ATID A collection of publicly available alternative
translational initiation events
http://bioinfo.au.tsinghua.edu.cn/atie/[153]
dbRESA database for annotated RNA editing sites http://bioinfo.au.tsinghua.edu.cn/dbRES/ [154]
PASDB A collection of genes reported to be alternatively
spliced in plants, spanning 44 plant species
http://pasdb.genomics.org.cn[155]
Protein resourcesMPSS An integrated database of protein annotationshttp://www.scbit.org/mpss/ [156]
SPDA secreted protein databasehttp://spd.cbi.pku.edu.cn [157]
SynDBA synapse protein database based on a synapse
ontology
http://syndb.cbi.pku.edu.cn[158]
DBSub-LocDatabase of protein subcellular localizations http://www.bioinfo.tsinghua.edu.cn/dbsubloc.
html
[45]
dbNEIA database for neuro-endocrine-immune
interaction.
http://bioinfo.au.tsinghua.edu.cn/dbNEIweb/[159]
SPIDerSaccharomyces protein–protein interaction
database
http://cmb.bnu.edu.cn/SPIDer/index.html [130]
InterDomA database of putative interacting protein
domains for validating predicted protein
interactions and complexes
http://InterDom.lit.org.sg [160]
doi:10.1371/journal.pcbi.1000020.t001
Bioinformatics in China
PLoS Computational Biology | www.ploscompbiol.org4April 2008 | Volume 4 | Issue 4 | e1000020

Page 5

bioinformatic methods and enzymatic
assays [72]. A number of interesting works
studied the migration history of Chinese
and East Asian populations by the se-
quencing and phylogentic analysis of Y-
chromosomes [73–77]. Recently the Chi-
nese HapMap Consortium participated in
the international HapMap project to
genotype the Chinese Han population
[78]. As more and more genetic data were
generated, new algorithms continued to be
developed, for instance a Java-based
method to analyze linkage disequilibrium
[79], software for haplotype block parti-
tion and htSNPs selection [80], a method
Table 2. Examples of Web servers and software packages developed and maintained in China.
CategoryName Main Function URLReference
Genome analysis toolsBGF Prediction of genes in the rice genomehttp://tlife.fudan.edu.cn/bgf/[57]
CVTreeA phylogenetic tree reconstruction tool based on
whole genome sequences without alignment
http://cvtree.cbi.pku.edu.cn [161]
ZCURVEA system for recognizing protein coding genes in
bacterial and archaeal genomes
http://tubic.tju.edu.cn/Zcurve_B/ [5]
Zplotter onlineA program to draw and manipulate the Z curve
online based on a user’s input DNA sequence
http://tubic.tju.edu.cn/zcurve/[148]
GS-Finder A program to find bacterial gene start sites with a
self-training method
http://tubic.tju.edu.cn/GS-Finder/[162]
GC-ProfileA Web-based tool for visualizing and analyzing the
variation of GC content in genomic sequences
http://tubic.tju.edu.cn/GC-Profile/[163]
FGF A Web tool for Fishing Gene Family in a whole
genome database
http://fgf.genomics.org.cn/ [164]
BPhyOGAn interactive server for genome-wide inference of
bacterial phylogenies based on overlapping genes
http://cmb.bnu.edu.cn/BPhyOG/[165]
SAPRED Using new structural and sequence attributes and
Support Vector Machine to predict possible disease
association of single amino acid polymorphism
http://sapred.cbi.pku.edu.cn/ [82]
Expression regulation
analysis tools
GBAEST-based digital gene expression profiling http://gba.cbi.pku.edu.cn[166]
CEASAn online server to analyze transcription factor
binding sites based on ChIP-chip data
http://ceas.cbi.pku.edu.cn[99]
OTFBSOver-represented Transcription Factor Binding Site
Prediction Tool
http://www.bioinfo.tsinghua.edu.cn/
%7Ezhengjsh/OTFBS/index.html
[48]
SVAP Identification and expression analysis of alternatively
spliced isoforms
http://svap.cbi.pku.edu.cn
CPC Prediction of the protein-coding potential of
transcripts using sequence features and support
vector machine
http://cpc.cbi.pku.edu.cn[107]
RDfolderA Web server for prediction of RNA secondary
structure
http://rna.cbi.pku.edu.cn[108]
miRASA data processing system for miRNA expression
profiling study
http://e-science.tsinghua.edu.cn/miras/ [106]
MiPredClassification of real and pseudo microRNA
precursors using random forest prediction model
with combined features
http://www.bioinf.seu.edu.cn/miRNA/ [103]
RFRCDB-siRNAImproved design of siRNAs by random forest
regression model coupled with database searching
http://www.bioinf.seu.edu.cn/siRNA/index.htm[167]
Protein analysis toolsEasyGO Gene Ontology-based annotation and functional
enrichment analysis tool for agronomical species
http://bioinformatics.cau.edu.cn/easygo/ [168]
CTKPredAn SVM-based method for the prediction and
classification of the cytokine superfamily
http://www.bioinfo.tsinghua.edu.cn/%7Ehn/
CTKPred/index.html
[169]
GNBSLA new integrative system to predict the subcellular
location for Gram-negative bacteria proteins
http://166.111.24.5/webtools/GNBSL/index.htm [127]
MeMo A Web tool for prediction of protein methylation
modifications
http://www.bioinfo.tsinghua.edu.cn/,tigerchen/
memo/contact.html
[170]
KOBASA Web-based platform for pathway identificationhttp://kobas.cbi.pku.edu.cn[171]
IntNetDBAn integrated protein–protein interaction network
database generated by a probabilistic model
http://hanlab.genetics.ac.cn/IntNetDB.htm[172]
PlatformsBOD A customizable bioinformatics on demand system
accommodating multiple steps and parallel tasks
http://e-science.tsinghua.edu.cn/bod/index.jsp[173]
ABCGridApplication for Bioinformatics Computing Grid http://abcgrid.cbi.pku.edu.cn[174]
doi:10.1371/journal.pcbi.1000020.t002
Bioinformatics in China
PLoS Computational Biology | www.ploscompbiol.org5April 2008 | Volume 4 | Issue 4 | e1000020