Flat Files What is a “flat file” ? • Flat file is a term used to refer to when data is stored in a plain ordinary file on the hard disk • Example RefSEQ – See CD-ROM – FILE: hs.GBFF • Hs: Homo Sapiens • GBFF: Genbank File Format • (associated with textpad, use monospaced font eg. Courier)
Nucleotide DatabasesEMBL Nucleotide Sequence Database (European Molecular BiologyLaboratory) http://www.ebi.ac.uk/ebi_docs/embl_db/ebi/topembl.htmlGenBank at NCBI (National Center for Biotechnology Information)http://www.ncbi.nlm.nih.gov/Web/Genbank/index.htmlDDBJ (DNA Database of Japan) http://www.ddbj.nig.ac.jp/DDBJ,the Center for operating DDBJ, National Institute of Genetics (NIG),Japan,established inApril 1995.http://www.ncbi.nlm.nih.gov/Genbank/genbankstats.htmlRelease Notes (ftp://ftp.ncbi.nih.gov/genbank/gbrel.txt)Genetic Sequence Data Bank - August 15 2003NCBI-GenBank Flat File Release 137.0Distribution Release Notes33 865 022 251 bases, from 27 213 748 reported sequences
GenBank FormatLOCUS LISOD 756 bp DNA BCT 30-JUN-1993DEFINITION L.ivanovii sod gene for superoxide dismutase.ACCESSION X64011.1 GI:37619753NID g44010KEYWORDS sod gene; superoxide dismutase.SOURCE Listeria ivanovii.ORGANISM Listeria ivanovii Eubacteria; Firmicutes; Low G+C gram-positive bacteria; Bacillaceae; Listeria.REFERENCE 1 (bases 1 to 756) AUTHORS Haas,A. and Goebel,W. TITLE Cloning of a superoxide dismutase gene from Listeria ivanovii by functional complementation in Escherichia coli and characterization of the gene product JOURNAL Mol. Gen. Genet. 231 (2), 313-322 (1992) MEDLINE 92140371REFERENCE 2 (bases 1 to 756) AUTHORS Kreft,J. TITLE Direct Submission JOURNAL Submitted (21-APR-1992) J. Kreft, Institut f. Mikrobiologie, Universitaet Wuerzburg, Biozentrum Am Hubland, 8700 Wuerzburg, FRG
Example of location descriptorsLocation Description476 Points to a single base in the presented sequence340..565 Points to a continuous range of bases bounded by and including the starting and ending bases<345..500 The exact lower boundary point of a feature is unknown.(102.110) Indicates that the exact location is unknown but that it is one of the bases between bases 102 and 110.(23.45)..600 Specifies that the starting point is one of the bases between bases 23 and 45, inclusive, and the end base 600123^124 Points to a site between bases 123 and 124145^177 Points to a site anywhere between bases 145 and 177J00193:hladr Points to a feature whose location is described in another entry: the feature labeled hladr in the entry (in this database) with primary accession J00193
batch download (ftp server)• Data available via website is most of the time also available via an ftp server to download a complete batch.• Examples: –ftp://ftp.ncbi.nih.gov/ –ftp://ftp.ebi.ac.uk/pub/
Sequence file format tips • When saving a sequence for use in an email message or pasting into a web page, use an unannotated text format such as FASTA • When retrieving from a database or exchanging between programs, use an annotated text format such as Genbank • When using sequence again with the same program, use that program’s annotated binary format (or annotated text if binary not available) – Asn-1 (NCBI) – Gbff (sanger) – XML
Expressed Sequence Tags • Sequence that codes for protein is < 5% of the genome. • Coding sequence can be obtained from mRNA by reverse transcription. • Tags for that sequence can be obtained by end- sequencing. • Incyte and HGS gambled on this being the useful part: – Search for homologies to known proteins, motifs. – Search for changed levels of expression and tissue specificity (“virtual/electronic northern” used in GeneCards) • ESTs have driven the huge expansion of GenBank: – Unigene now contains some sequence from most genes. – > 4,000,000 human est sequences – http://www.ncbi.nlm.nih.gov/dbEST/
Traces <-> strings • Traces contain much more information – TraceDB: http://www.ncbi.nlm.nih.gov/Traces/Example
Traces <-> strings • Phrep – base calling, vector trimming, end of sequence read trimming • Phrap – Phrap uses Phred’s base calling scores to determine the consensus sequences. Phrap examines all individual sequences at a given position, and uses the highest scoring sequence (if it exists) to extend the consensus sequence • Consend – graphical interface extension that controls both Phred and Phrap
What is Phred?• Phred is a program that observes the base trace, makesbase calls, and assigns quality values (qv) of bases in thesequence.• It then writes base calls and qv to output files that will beused for Phrap assembly. The qv will be useful forconsensus sequence construction.• For example, ATGCATGC string1 ATTCATGC string2 AT-CATGC superstring• Here we have a mismatch ‘G’ and ‘T’, the qv willdetermine the dash in the superstring. The base with higherqv will replaces the dash.
How Phred calculates qv?• From the base trace Phred know number of peaks and actual peak locations.• Phred predicts peaks locations.• Phred reads the actual peak locations from base trace.• Phred match the actual locations with the predicted locations by using Dynamic Programming.• The qv is related to the base call error probability (ep) by the formula qv = -10*log_10(ep)• Example 1:10000 = qv 40
Why Phred? • Output sequence might contain errors. • Vector contamination might occur. • Dye-terminator reaction might not occur. • Segment migration abnormal in gel electrophoresis. • Weak or variable signal strength of peak corresponding to a base.
End of Sequence Cropping• It is common that the end of sequencing reads have poor data. This is due to the difficulties in resolving larger fragment ~1kb (it is easier to resolve 21bp from 20bp than it is to resolve 1001bp from 1000bp).• Phred assigns a non-value of ‘x’ to this data by comparing peak separation and peak intensity to internal standards. If the standard threshold score is not reached, the data will not be used.
NCBI reference sequencesRefSeq database is a non-redundant set of reference standards that includes chromosomes, complete genomic molecules, intermediate assembled genomic contigs, curated genomic regions, mRNAs, RNAs, and proteins.
RefSeq nomenclature - modelsXM_#### mRNAXR_#### RNAXP_#### proteinAutomated Homo sapiens models provided by the Genome Annotation process; sequence corresponds to the genomic contig.
Open reading frame• Definition: – A stretch of triplet codons with an initiator codon at one end and a stop codon sat the other, as identifiable by nucleotide sequences.• Example – http://www.ncbi.nlm.nih.gov/entrez/query.fcgi? cmd=Retrieve&db=nucleotide&list_uids=6688 473&dopt=GenBank&term=Y18948.1&qty=1
Protein sequence databaseSWISS-PROT & TREMBLSwissProt - http://expasy.hcuge.ch/sprot/ SWISS-PROT is an annotated protein sequence database The sequences are translated from the EMBL Nucleotide Sequence Database Sequence entries are composed of different lines. For standardization purposes the format of SWISS-PROT follows as closely as possible that of the EMBL Nucleotide Sequence Database. Continuously updated (daily).
Different Features of SWISS-PROT • Format follows as closely as possible that of EMBL’s • Curated protein sequence database • Three differences: 1. Strives to provide a high level of annotations 2. Minimal level of redundancy 3. High level of integration with other databases
1. Annotation Three Distinct Criteria The sequence data; the citation information (bibliographical references) and the taxonomic data (description of the biological source of the protein) such as protein functions,post-translational modifications ,domains and sites,secondary structure,quaternary structure,similarities to other proteins,diseases associated with deficiencies in the protein,sequence conflicts, variants, etc.
2. Minimal Redundancy any sequence databases contain, for a given protein sequence, separate entries which correspond to different literature reports. SWISS- PROT is as much as possible to merge all these data so as to minimize the redundancy. If conflicts exist between various sequencing reports, they are indicated in the feature table of the corresponding entry.
3. Integration With Other Databases • SWISS-PROT and TrEMBL - Protein sequences • PROSITE - Protein families and domains • SWISS-2DPAGE - Two-dimensional polyacrylamide gel electrophoresis • SWISS-3DIMAGE - 3D images of proteins and other biological macromolecules • SWISS-MODEL Repository - Automatically generated protein models • CD40Lbase - CD40 ligand defects • ENZYME - Enzyme nomenclature • SeqAnalRef - Sequence analysis bibliographic references
TREMBL- http://expasy.hcuge.ch/sprot/ Translated EMBL sequences not (yet) inSwissprot. Updated faster than SWISS-PROT.TREMBL - two parts1. SP-TREMBL Will eventually be incorporated into Swissprot Divided into FUN, HUM, INV, MAM, MHC, ORG, PHG, PLN, PRO, ROD, UNC, VRL and VRT.2. REM-TREMBL (remaining) Will NOT be incorporated into Swissprot Divided into:Immunoglobins and T-cell receptors,Synthetic sequences,Patent application sequences,Small fragments,CDS not coding for real proteins
SWISS-PROT/TrEMBL • TrEMBL is a computer-annotated supplement of SWISS-PROT that contains all the translations of EMBL nucleotide sequence entries not yet integrated in SWISS-PROT • SWISS-PROT Release 39.15 of 19- Mar-2001: 94,152 entries TrEMBL Release 16.2 of 23-Mar- 2001: 436,924 entries
CC -!- FUNCTION: CYTOKINE WITH A WIDE VARIETY OF FUNCTIONS: IT CANCC CAUSE CYTOLYSIS OF CERTAIN TUMOR CELL LINES, IT IS IMPLICATEDCC IN THE INDUCTION OF CACHEXIA, IT IS A POTENT PYROGEN CAUSINGCC FEVER BY DIRECT ACTION OR BY STIMULATION OF IL-1 SECRETION, ITCC CAN STIMULATE CELL PROLIFERATION & INDUCE CELL DIFFERENTIATIONCC UNDER CERTAIN CONDITIONS. CommentsCC -!- SUBUNIT: HOMOTRIMER.CC -!- SUBCELLULAR LOCATION: TYPE II MEMBRANE PROTEIN. ALSO EXISTS ASCC AN EXTRACELLULAR SOLUBLE FORM.CC -!- PTM: THE SOLUBLE FORM DERIVES FROM THE MEMBRANE FORM BYCC PROTEOLYTIC PROCESSING.CC -!- DISEASE: CACHEXIA ACCOMPANIES A VARIETY OF DISEASES, INCLUDINGCC CANCER AND INFECTION, AND IS CHARACTERIZED BY GENERAL ILLCC HEALTH AND MALNUTRITION.CC -!- SIMILARITY: BELONGS TO THE TUMOR NECROSIS FACTOR FAMILY.DR EMBL; X02910; G37210; -. Database Cross-referencesDR EMBL; M16441; G339741; -.DR EMBL; X01394; G37220; -.DR EMBL; M10988; G339738; -.DR EMBL; M26331; G339764; -.DR EMBL; Z15026; G37212; -.DR PIR; B23784; QWHUN.DR PIR; A44189; A44189.DR PDB; 1TNF; 15-JAN-91.DR PDB; 2TUN; 31-JAN-94.
Protein searching3-levels of Protein Searching1. Swissprot Little Noise Annotated entries2. Swissprot + TREMBL More Noisy All probable entries3. Translated EMBL - tblast or tfasta Most Noisy All possible entries
New initiatiaves • IPI: International Protein Index – http://www.ebi.ac.uk/IPI/IPIhelp.ht ml • UNIPROT: Universal Protein Knowledgebase – http://www.pir.uniprot.org/ • HPRD: Human Protein Reference Database – http://www.hprd.org/
UniProt UniProt Consortium • European Bioinformatics Institute (EBI) • Swiss Institute of Bioinformatics (SIB) • Protein Information Resource (PIR) Uniprot Databases •UniProt Knowledgebase (UniProtKB) •UniProt Reference Clusters (UniRef) •UniProt Archive (UniParc) UniprotKB •Swiss-Prot (annotated protein sequence db, golden standard) •trEMBL (translated EMBL + automated electronic annotations)
understanding molecular structure is critical to the understanding of biologybecause because structure determines function
From Structure to Function• the drug morphine has chemical groups that are functionally equivalent to the naturalendorphins found in the human body
From Structure to Function• the drug morphine has chemical groups that are functionally equivalent to the naturalendorphins found in the human body • the receptor molecules located at the synapse (between two neurons) bind morphine much the same way as endorphins • therefore, morphine is able to attenuate the pain response
Structure databasesProtein Data Bank (PDB)Protein Data Bank - http://www.rcsb.org/pdbDiffraction 7373 structures determined by X-ray diffractionNMR 388 structures determined by NMR spectroscopyTheoretical Model 201 structures proposed by modeling
• PDB is three-dimensional structure of proteins,some nuclei acids involved• PDB is operated by RCSB(Research Collaboratory for Structural Bioinformatics),funded by NSF, DOE, and two units of NIH:NIGMS National Institute Of General Medical Sciences and NLM National Library Of Medicine.• Established at BNL Brookhaven National Laboratories in 1971,as an archive for biological macromolecular crystal structures• In 1980s, the number of deposited structures began to increase dramatically.• October 1998, the management of the PDB became the responsibility of RCSB.• Website http://www.rcsb.org
PDB Holdings List: 27-Mar-2001 Molecule Type Proteins, Protein/ Peptides, Nucleic Nuclei Carbohydrate total and Viruses Acid c s Complexes Acids X-ray 11045 526 552 14 12137Exp. Diffraction and otherTech. NMR 1832 71 366 4 2273 Theoretica 281 19 21 0 321 l Modeling total 13158 616 939 18 14731 5032 Structure Factor Files 968 NMR Restraint Files
Other structure databasesBioMagResBank http://www.bmrb.wisc.edu/A Repository for Data from NMR Spectroscopy on Proteins, Peptides, and NucleicAcidsBiological Macromolecule Crystallization Database (BMCD) http://h178133.carb.nist.gov:4400/bmcd/bmcd.htmlContains crystal data and the crystallization conditions, which have been compiledfrom literatureNucleic Acid Database (NDB) http://ndbserver.rutgers.edu:80/Assembles and distributes structural information about nucleic acidsStructural Classification of Proteins (SCOP) http://scop.mrc-lmb.cam.ac.uk/scop/Structure similarity search. Hierarchic organization.MOOSE http://db2.sdsc.edu/moose/Macromolecular Structure QueryCambridge Structural Database (CSD) http://www.ccdc.cam.ac.uk/Small molecules.
Protein Splicing?• Protein splicing is defined as the excision of an intervening protein sequence (the INTEIN) from a protein precursor and the concomitant ligation of the flanking protein fragments (the EXTEINS) to form a mature extein protein and the free intein• http://www.neb.com/inteins/intein_intro.ht ml
Biological databases • NAR Database Issue – Every year: NAR DB Issue – The 2006 update includes 858 databases – Citation top 5 are: • Pfam • Gene Ontology • UniProt • SMART • KEGG – Primary Nucleotide DB’s and PDB are not cited anymore
Why biological databases ? • Explosive growth in biological data • Data (sequences, 3D structures, 2D gel analysis, MS analysis….) are no longer published in a conventional manner, but directly submitted to databases • Essential tools for biological research, as classical publications used to be !
Problems with Flat files … • Wasted storage space • Wasted processing time • Data control problems • Problems caused by changes to data structures • Access to data difficult • Data out of date • Constraints are system based • Limited querying eg. all single exon GPCRs (<1000 bp)
Relational • The Relational model is not only very mature, but it has developed a strong knowledge on how to make a relational back-end fast and reliable, and how to exploit different technologies such as massive SMP, Optical jukeboxes, clustering and etc. Object databases are nowhere near to this, and I do not expect then to get there in the short or medium term. • Relational Databases have a very well-known and proven underlying mathematical theory, a simple one (the set theory) that makes possible – automatic cost-based query optimization, – schema generation from high-level models and – many other features that are now vital for mission-critical Information Systems development and operations.
• What is a relational database ? – Sets of tables and links (the data) – A language to query the datanase (Structured Query Language) – A program to manage the data (RDBMS)• Flat files are not relational – Data type (attribute) is part of the data – Record order mateters – Multiline records – Massive duplication • Bv Organism: Homo sapeinsm Eukaryota, … – Some records are hierarchical • Xrefs – Records contain multiple “sub-records” – Implecit “Key”
The Benefits of Databases • Redundancy can be reduced • Inconsistency can be avoided • Conflicting requirements can be balanced • Standards can be enforced • Data can be shared • Data independence • Integrity can be maintained • Security restrictions can be applied
Relational Database Terminology• Each row of data in a table is uniquely identified by a primary key (PK)• Information in multiple tables can be logically related by foreign keys (FK) Table Name: CUSTOMER Table Name: EMP ID NAME PHONE EMP_ID ID LAST_NAME FIRST_NAME 201 Unisports 55-2066101 12 10 Havel Marta 202 Simms Atheletics 81-20101 14 11 Magee Colin 203 Delhi Sports 91-10351 14 12 Giljum Henry 204 Womansport 1-206-104-0103 11 14 Nguyen Mai Primary Key Foreign Key Primary Key
• RDBM products – Free • MySQL, very fast, widely usedm easy to jump into but limited non standard SQL • PostrgreSQL – full SQLm limited OO, higher learning curve than MySQL – Commercial • MS Access – Great query builder, GUI interfaces • MS SQL Server – full SQL, NT only • Oracle, everything, including the kitchen sink • IBM DB2, Sybase
A simple datamodel (tables and relations) Prot_id name seq Species_id 1 GTM1_HUMA MGTDHG… 1 N 2 GTM1_RAT MGHJADSW.. 2 3 GTM2_HUMA MVSDBSVD.. 1 N Species_id name Full Lineage 1 human Homo Sapiens … 2 rat Rattus rattus
Relational Database Fundamentals • Basic SQL – SELECT – FROM – WHERE – JOIN – NATURAL, INNER, OUTER • Other SQL functions – COUNT() – MAX(),MIN(),AVE() – DISTINCT – ORDER BY – GROUP BY – LIMIT
• Query: een opdracht om gegevens uit een databaase op te vragen noemt men een query• eg. MyGPCRdb – Bioentry – Taxid (include full lineage) – Linking table (bioentry_tax)
MyGPCR;Geef me allE GPCR die korter zijn dan 1000bpselect * from bioentry;select count(*) from bioentry;select * from bioentry inner join biosequence on bioentry.bioentry_id=biosequence.bioentry_id ;select * from bioentry inner join biosequence on bioentry.bioentry_id=biosequence.bioentry_id where length(biosequence_str)<1000;
Example 3-tier model in biological databaseExample of different interface to the same back-end database (MySQL) http://www.bioinformatics.be
Overview • DataBases – FF • *.txt • Indexed version – Relational (RDBMS) • Access, MySQL, PostGRES, Oracle – OO (OODBMS) • AceDB, ObjectStore – Hierarchical • XML – Frame based systemOverview • Eg. DAML+OIL – Hybrid systems
Object • The Object paradigm is already proven for application design and development, but it may simply not be an adequate paradigm for the data store. • Object Database are modelled by graphs. The graph theory plays a great role on computer science, but is also a great source of unbeatable problems, the NP-complex class: problems for which there are no computationally efficient solution, as theres no way to escape from exponential complexity. This is not a current technological limit. Its a limit inherent to the problem domain. • Hybrid Object-Relational databases will probably be the long term solution for the industry. They put a thin object layer above the relational structure, thus providing a syntax and semantics closer to the object oriented design and programming tools. They simply make it easier to build the data layer classes
Conclusions • A database is a central component of any contemporary information system • The operations on the database and the mainenance of database consistency is handled by a DBMS • There exist stand alone query languages or embedded languages but both deal with definition (DDL) and manipulation (DML) aspects • The structural properties, constraints and operations permitted within a DBMS are defined by a data model - hierarchical, network, relational • Recovery and concurrency control are essential • Linking of heterogebous datasources is central theme in modern bioinformatics
• How do you know which database exists ?• NAR list• Weblinks op Nexus – Searchable – Maintainable
• Tools available in public domain for simultaneous access – entrez – srs• Batch queries for offload in local databases for subsequent analysis (see further)
• What if you want to search the complete human genome (golden path coordinates) instead of separate NCBI entries ?• ENSEMBL
BioMart • Joined project between EBI and CSHL, http://www.biomart.org/ • Aim is to develop a generic, query-oriented data management system capable of integrating distributed data sources • 3 step system: – Start by selecting a dataset to query – Filter this dataset by applying the appropriate filters – Generate the output by selecting the attributes and output format • Available public biomart websites: http://www.biomart.org/biomart/martview
BioMart - Single access point - Generic interface
BioMart – 3 step system Dataset Attribute Filter
BioMart - 3 step system Name, chromosomeDataset position, descriptionAttribute for all Ensembl genesFilter located on chromosome 1, expressed in lung, associated with human homologues
BioMart - EnsMart • The first in line was EnsMart, a powerful data mining toolset for retrieving customized data sets from annotated genomes. EnsMart integrates data from Ensembl and various worldwide data sources. • EnsMart provides .... – Gene and protein annotation – Disease information – Cross-species analyses – SNPs affecting proteins – Allele frequency data – Retrieval by external identifiers – Retrieval by Gene Ontology – Customized sequence datasets – Microarray annotation tools
Other BioMart implementations • Other data resources also implemented a BioMart interface: – Wormbase – Gramene – HapMap – DictyBase – euGenes
BioBar • A toolbar for browsing biological data and databases http://biobar.mozdev.org/ • The following databases are included http://biobar.mozdev.org/Databases.ht ml • a toolbar for Mozilla-based browsers including Firefox and Netscape 7+
Weblems Weblems Online (example posting) W2.1. Which isolate of Tabac was used in record accession Z71230, and human sample in the genbank entry with accession AJ311677 ? W2.2: Find all structures of GFP in the Protein Data Bank and draw a histogram of their dates of deposition ? W2.3: What is the chromosomal location of the human gene for insulin ? W2.4: How many different human NHR (nuclear hormone receptors) s exist ? How many of these are single exon genes ? Are there any drugs working on this class of receptors ? W2.5: The gene for Berardinelli-Seip syndrome was initially localized between two markers on chromosome band 11q13- D11S4191 and D11S987. a. How many base pairs are there in the interval between these two markers ? b. How many known genes are there ? c. List the gene ontology terms for that region ?