1. Quality of services of a primary nucleotide sequence database
Hideaki Sugawara
Center for Information Biology and DDBJ, National Institute of Genetics
SOKEN-DAI.

2. Primary Biological Databases
International Nucleotide Sequence Database Collaboration (INSDC) serves communities
JPO: Japan Patent Office
DDBJ: DNA Data Bank of Japan
NIG: National Institute of Genetics
EMBL: European Molecular Biology Laboratory
EBI: Europen Bioinfomatics Institute
NCBI: National Center for Biotechnology Information
EPO: European Patent Office
USTPO: US Trademark and Patent Offrice
JPO
DDBJ
NIG
GenBank
NCBI
EMBL
EBI
EPO
USTPO
2006/10/23
Primary Biological Databases

3. Infrastructure for the services
Informatics
Contents
Mangement
2006/10/23
Primary Biological Databases

4. Today I will introduce:
Web services
(programmable interface)
Derived databases
Q&A databases
A bug tracking
system
2006/10/23
Primary Biological Databases

5. Mash-up of biological data resources
It has bee feasible to mash-up diverse databases and tools by hand to some extent.
①　Connect to a journal database to find accession numbers of INSDC
A. Web site of Journal database
Users with
Web browser
②. Move to INSDC to search amino acid sequences by the accession numbers found in the step ①
B. Web site of Nucleotide database
C. Web site of Protein Data Bank (PDB)
③. Move to the Protein Data Bank (PDB) to get the 3D structure.
The three steps in the above are accomplished only by click, copy and paste by mouse. However, It is very difficult when dealing with a lot of data.
Nucleic Acid ResearchのDB特別号に掲載されるデータベース数は、500をこえ、
自分の目的に沿った様々な生物情報資源を効率的に取得するためには、
Webブラウザを通じて、様々なデータベースや解析ツールにアクセスする必要がある。
例えば、この例で示しておりますのは、論文に記載されている配列データについて、タンパク立体構造が
判明しているのかどうかを調べたいときのアクセス例を示しております。
一件だけならば、このようなアクセスでも苦はないのですが、複数ですと、
取得は非常に困難になってしまます。より・・・ありと考えております。
Large-scale databases are produced in OMICS. Therefore, a
machanical mash-up by program will be required more and more.
2006/10/23
Primary Biological Databases

6. Web services for the biological problem solving environment
We implemented a SOAP (Simple Object Access Protocol) server and Web services that provide a program-friendly interface.
We propose standardization of bioinformatics services to improve the interoperability of desperately diverse biological data resources.
2006/10/23
Primary Biological Databases

7. The world of Web services
「Web services」 provider
EBI, NCBI, -----
DDBJ
UDDI
Registration of Web services
SOAP
(HTTP/HTTPS)
Search Web services
WSDL
XML
Users
UDDI (Universal Description, Discovery and Integration)
WSDL
(Web Services Description Language).
英語で説明を書く！！！
具体的なサービスの例として、Web servicesの提供者によって、UDDI()と呼ばれる利用可能なサービスの登録
が行われ、利用者はそのUDDIから自分が利用したいサービスの検索を行います。
次に、自分の利用したいサービスを見つけた利用者は、サービスが登録されている提供者から、
そのサイトで提供さえているサービスの一覧とアクセス方法が記載されているWSDLを取得し、
実際に、SOAPを介したサービスを使用することが出来ます。
WSDLとSOAPを介したサービスの提供はXMLファイルにより交換が行われます。
2006/10/23
Primary Biological Databases

8. XML Central of DDBJ: http://xml.nig.ac.jp/index.html
As DDBJ Web services server, We open the SOAP interfaces with the several services which offer by DDBJ.
2006/10/23
Primary Biological Databases

9. Workflow composed of methods provided by Web services
141 methods in 15 services are available.
Blast
ClustalW
DDBJ
Ensembl
Fasta
Getentry
GIB
GTOP
NCBI Genome
Annotation
PML
RefSeq
SPS
SRS
TxSearch
VecScreen
Open
OMIM
Gene Ontology
Rfam
developed
Custom made workflows
2006/10/23
Primary Biological Databases

10. Easy to bind Web services methods from your program
Download ActivePerl for WINDOWS and you can call methods in DDBJ Web services.
Specify a WDSL file and call a method:
In the following example, the method of getXML_DDBJEntry(accession) of the Getentry Web services is called from a Perl program:
1. Include the package needed to use SOAP service.
2. Specifies WSDL file of SOAP service you want to use.
3. Call the service you want to use.
2006/10/23
Primary Biological Databases

11. Workflow with Online Mendelian Inheritance in Man (OMIM)
Web services
tool
Data
Local program
閾値に関しては、記載する。
This workflow reveal homology relationship between
human disease genes and genes of other eukaryotes.
2006/10/23
Primary Biological Databases

12. Environmental DNA sequences automatic annotation workflow
In order to make full use of gene information included in nucleotide sequence database, we developed workflow of gene finding of DNA fragments obtained from pooled genome samples of uncultured microbes in environmental samples.
BLASTP
Environmental
Sample
EMBOSS
(getorf)
ORF
Selected
ORF
Attached
Product
InterPro
Scan
FLOW
Execute getorf in EMBOSS which finds and outputs the sequences of open reading frames (ORFs). The ORFs can be defined as regions of a specified minimum size (longer than 300bp) between START and STOP codons. (The minimum size cut-off for getorf was )
The protein product was attacheded by searching the bacterial division of DDBJ release 62 against predicted ORFs using BLASTP. (The e-value cut-off for BLASTP was 1e-40.)
遺伝子情報
We observed DNA fragments obtained from pooled genome samples of uncultured microbes in environmental and clinical samples.
Evidence-based gene finding. Evidence in the form of protein alignments was used to
determine the most likely coding frame. The approximate start and stop positions were
likewise determined from the bounding coordinates of the alignments and refined to
identify specific start and stop codons. This methodology was applied to several different
lines of evidence in a series of steps as follows: First evidence was generated by
searching the bacterial portion of the nraa dataset against the Sargasso data using tblastn.
All blast searches were performed using NCBI blastall version Unless stated
otherwise all blast searches were performed with these generic parameters: -Y
F "m L" -U T -v 5. The NCBI non-redundant amino acid dataset (nraa)
was downloaded from NCBI on September 2nd and contains 1,510,260 peptides of which
626,877 were classified as bacterial. The tblastn searches using the bacterial subset of
nraa were performed with these additional parameters: -e 1e-4 -b K The
blastx searches against all of nraa were done with these additional parameters: -e 1e-3 -b
5000 -K The tblastx searches used these parameters: -e 1e-3 -b K The
search of the Sargasso data against itself using blastn used these parameters: -W 20 -K
b q -9 -e 1e-40.
2006/10/23
Primary Biological Databases

13. Access frequency of DDBJ Web services
　Examples of methods frequently used: getentry, blast, RefSeq、GO 、
2006/10/23
Primary Biological Databases

14. Web services in the world
EBI, NCBI, ----
DDBJ
SOAP(HTTP/HTTPS)
利用者
(Web services map prepared by EBI)
2006/10/23
Primary Biological Databases

15. Primary Biological Databases
Derived databases
2006/10/23
Primary Biological Databases

16. >60M entries
>100G bps
~ 0.5 projects
INSDC is not only large-scale but also complex. It is not easy to retrieve what you want to anapyze.

17. Primary Biological Databases
A simple derived database: subset of complete microorganisms genomes from INSDC
2006/10/23
Primary Biological Databases

18. Gene prediction programs used
2006/10/23
Primary Biological Databases

19. Parameters: the cutoff length used
number
＞20
＞（＝）30
>33.3aa　（100bp)
>40aa
>50aa
>60aa
>66.6aa　(200bp)
>80
>100aa
>150aa
>200aa
>300aa
>400aa 1 25 3 7 4 2 6
2006/10/23
Primary Biological Databases

20. COG ID in the qualifier of /note
Annotation: description of products
~ Hahella chejuensis KCTC 2396
CDS
/gene="argB"
/locus_tag="HCH_01027"
/EC_number=" "
/note="COG0548"
/codon_start=1
/transl_table=11
/product="Acetylglutamate kinase"
/protein_id="ABC "
/db_xref="GI: "
/translation="MLDRDNALQVAAVLSKALPYIQRFAGKTIVIKYGGNAMTDEELK
NSFARDVVMMKLVGINPIVVHGGGPQIGDLLQRLNIKSSFINGLRVTDSETMDVVEMV
LGGSVNKDIVALINRNGGKAIGLTGKDANFITARKLEVTRATPDMQKPEIIDIGHVGE
VTGVRKDIITMLTDSDCIPVIAPIGVGQDGASYNINADLVAGKVAEVLQAEKLMLLTN
IAGLMNKEGKVLTGLSTKQVDELIADGTIHGGMLPKIECALSAVKNGVHSAHIIDGRV
PHATLLEIFTDEGVGTLITRKGCDDA"
COG ID in the qualifier of /note
~ Archaeoglobus fulgidus DSM 4304
CDS complement( )
/locus_tag="AF_1280"
/note="similar to GB:L77117 SP:Q60382 PID: percent
identity: 56.06; identified by sequence similarity;
putative"
/codon_start=1
/transl_table=11
/product="acetylglutamate kinase (argB)"
/protein_id="AAB "
/db_xref="GI: "
/translation="MENVELLIEALPYIKDFHSTTMVIKIGGHAMVNDRILEDTIKDI
VLLYFVGIKPVVVHGGGPEISEKMEKFGLKPKFVEGLRVTDKETMEVVEMVLDGKVNS
KIVTTFIRNGGKAVGLSGKDGLLIVARKKEMRMKKGEEEVIIDLGFVGETEFVNPEII
RILLDNGFIPVVSPVATDLAGNTYNLNADVVAGDIAAALKAKKLIMLTDVPGILENPD
DKSTLISRIRLSELENMRSKGVIRGGMIPKVDAVIKALKSGVERAHIIDGSRPHSILI
ELFTKEGIGTMVEP"
Some details of homology analysis in the /note qualifier
~ Agrobacterium tumefaciens C58 circular chromosome
CDS complement( )
/gene="AGR_C_666"
/note="acetylglutamate kinase PA5323 {imported} -
Pseudomonas aeruginosa (strain PAO1)"
/codon_start=1
/transl_table=11
/product="AGR_C_666p"
/protein_id="AAK "
/db_xref="GI: "
/translation="MTSSESEIQARLLAQALPFMQKYENKTIVVKYGGHAMGDSTLGK
AFAEDIALLKQSGINPIVVHGGGPQIGAMLSKMGIESKFEGGLRVTDAKTVEIVEMVL
AGSINKEIVALINQTGEWAIGLCGKDGNMVFAEKAKKTVIDPDSNIERVLDLGFVGEV
VEVDRTLLDLLAKSEMIPVIAPVAPGRDGATYNINADTFAGAIAGALHATRLLFLTDV
PGVLDKNKELIKELTVSEARALIKDGTISGGMIPKVETCIDAIKAGVQGVVILNGKTP
HSVLLEIFTEGAGTLIVP"
Product ID in the qualifier of /product
Product name in the qualifier of /note
2006/10/23
Primary Biological Databases

21. Annotation: description of products
unknown
hypothetical protein
probable orf
predicted protein
putative protein
hypothetical conserved protein
uncharacterized protein
conserved domain protein
conseved hypthetical
2006/10/23
Primary Biological Databases

22. Annotation: inconsistency or biological variation
Agrobacterium tumefaciens C58 circular chromosome (Cereon)
Agrobacterium tumefaciens C58 circular chromosome (U. Washington)
2006/10/23
Primary Biological Databases

23. Primary Biological Databases
A derived database: contents of complete microorganisms genomes were evaluated
Gene Trek in Procayote Space
2006/10/23
Primary Biological Databases

24. Primary Biological Databases
Features of GTPS
Simultaneous prediction and evaluation of all the possible protein coding genes (ORFs) in prokaryote genomes
All the ORFs are graded
The criteria and evidence data are available
Web site
Updated once a year
2006/10/23
Primary Biological Databases

25. Primary Biological Databases
Short history of GTPS
ver. (data froze in)
strains
Archaea
Bacteria
2003 (Jul 2003)
123 14
109
2004 (Sep 2004)
183 17
166
2005 (Feb 2006)
302 25
277
2006/10/23
Primary Biological Databases

26. Not putative membrane nor unknown protein
Grade
blastp hit
InterProScan hit
Coverage
Subject
Subject
alignment
subject query
AAAA
BBBB
Function known motif
Not putative membrane nor unknown protein
≧ 70%
AAA
BBB
Unknown motif
& or AA BB
No hit
AAAA1-D3 grades
Potential genes
alignment/
ORF
Putative membrane
or unknown
protein
Function known
or unknown
motif
≧ 70% A B
≧ 70%
Putative membrane
protein
No hit C or
Function known
or unknown motif
No hit
≧ 70%
Unknown
protein D
No hit
&
≧ 70%
Unknown
protein E
No hit or X
No hit
No hit
2006/10/23
Primary Biological Databases

27. Potential genes (AAAA1 - D3)
Number of ORFs sorted by each grade
GTPS ver. 2003
GTPS ver. 2004
GTPS ver. 2005
Grades
AAAA-A
BBBB-B C D E X
Total
283,247
7,208
4,680
79,779
6,788
466,681
848,383
431,672
10,250
7,511
107,382
10,225
687,110
1,254,150
752,186
20,755
16,227
137,656
17,739
278,075
1,222,638
Glimmer2 and RBS finder Glimmer 3
Potential genes (AAAA1 - D3)
370,876
551,246
903,845
INSDC
362,828
537,312
904,530
2006/10/23
Primary Biological Databases

28. Comparison of the number of protein coding genes between GTPS and INSDC
GTPS ver. 2003
GTPS ver. 2004
GTPS ver. 2005
Identical
3' matched
New ORFs (Not annotated in INSDC)
Not predicted by Glimmer
Total (potential genes)
261,720
92,206
14,954
1,996
370,876
70.6%
24.9%
4.0%
0.5%
390,557
133,146
23,935
3,608
551,246
70.8%
24.2%
4.3%
0.7%
576,762
252,365
31,438
43,280
903,845
63.8%
27.9%
3.5%
4.8%
Glimmer 3 was used for ver.2005.
2006/10/23
Primary Biological Databases

29. How many genes are out there?
The ratio of the clustered ORFs to all the ORFs among the sampled genomes increased with the increasing number of the genomes. The ratio is not yet saturated at the 180 genomes (29.5% of the genes among 180 genomes were clustered and presumed to have the same function.).
2006/10/23
Primary Biological Databases

30. Primary Biological Databases
Summary
It is the long term mission of the primary database to archive all the data published for the long term.
At the same time, it is requested to provide objective/reliable data and tools for the mash-up.
It proposes and maintains standards of biological data processing for the long term.
2006/10/23
Primary Biological Databases