1、Can New Oracle10g Search Features Help Bridge the Biological Discovery Gap?,Jake Y. Chen, Ph.D. Head of Computational Proteomics & Principle Bioinformatics Scientist,Marcel Davidson Head of Data Management,Messages,New Informatics Challenges in Protein Interactomics R&D Scale, integration, discovery
2、 issues A data-driven discovery-oriented framework “Enabling” Features in 10g? Biological data integration? Biological data analysis integration?,Outline,Data-driven Discovery-oriented Computational Framework 10g Regular Expression Case Studies 10g BLAST Case Studies,Why Myriad Maps Protein-Protein
3、Interactions,Conventional Drug Discovery,Post-Genomic Drug Discovery,Nucleus,GPCR,enzyme,hormone receptor,Nucleus,target validation,lead discovery, optimization,novel, more specific targets non-specific targets novel, druggable targets,enhanced pre-validation target pool,Principle of the Yeast Two-H
4、ybrid (Y2H) System,Reporter Gene,DNA Binding Domain,Bait,Activation Domain,Prey,( No Reporter Gene Activity ),Scenario B: Human Proteins X and Z do not Interact,Readout: No growth of yeast colonies,Human Protein Z,Human Protein X,DNA,Reporter Gene,DNA Binding Domain,Human Protein X,Reporter mRNA,Rep
5、orter mRNA,Reporter mRNA,Bait,Human Protein Y,Activation Domain,Prey,Scenario A: Human Proteins X and Y do Interact,Readout: Yeast colonies grow,DNA,Data Collected from Y2H System,Perform BLAST Against Human REFSEQ DB,Protein Interaction Network (Snapshot of 8,000 interactions),Knowledge Discovery (
6、KD) Challenges,80,000 unique interactions100 biological data sources,Protein Interaction Data,$, drugs, ,Data-drivenDiscovery-oriented,KD in Interaction-based Proteomics,Bioinformatics DB Framework,Annotation DB RefSeq, LocusLink, GO, OMIM, CGAP, Protein Kinase DB, GPCR DB, Ensemble, Curation, ,Y2H
7、Data Processing and Analysis DB Lab_Seq, Seq_Match, Y2H_Mart,Y2H Interaction Data Mart Y2h_Mart,A Schema Fragment to Manage Sequence Similarity Results,Jake Yue Chen and John Carlis (2003) Genomic Data Modeling. Information Systems Journal, 28(4), p287-310.,Interaction Matrix using Randomly Ordered
8、Locus IDs,12,958 unique Interactions 1955 bait loci 2766 prey loci,Jake Yue Chen, et al (2003) Proceedings of the IEEE Computer Science Society Bioinformatics Conference 2003. Stanford University, Stanford, CA.,Outline,Data-driven Discovery-oriented Computational Framework 10g Regular Expression Cas
9、e Studies 10g BLAST Case Studies,Oracle10g Regular Expressions: Powerful String Processing,RE new tools in Oracle10g Search and manipulate data strings of arbitrary complexity Prior database solutions SQL LIKE operator Java stored procedures, C external libraries Prior non-database solutions: AWK, S
10、ED, GREP, PERL, etc. Done now inside database Facilitates rapid data-centric analysis,Case1: Retrieving Protein data from SGD (Saccharomyces Genome Database),ORF Identifier,Associated Amino Acid Sequence,HTTP Raw Data,Quick Search:Site Map | Help | Full Search | HomeCommunity InfoSubmit DataBLASTPri
11、mersPatMatchGene/Seq ResourcesVirtual LibraryContact SGDSequence for a region of YDR099W/BMH2Send questions or suggestions to SGDBLAST search | FASTA searchProtein translation of the coding sequence.Other Formats Available: GCGYDR099W Chr 4 MSQTREDSVYLAKLAEQAERYEEMVENMKAVASSGQELSVEERNLLSVAYKNVIGARRA
12、S WRIVSSIEQKEESKEKSEHQVELIRSYRSKIETELTKISDDILSVLDSHLIPSATTGESK VFYYKMKGDYHRYLAEFSSGDAREKATNSSLEAYKTASEIATTELPPTHPIRLGLALNFS VFYYEIQNSPDKACHLAKQAFDDAIAELDTLSEESYKDSTLIMQLLRDNLTLWTSDISES GQEDQQQQQQQQQQQQQQQQQAPAEQTQGEPTK*Return to SGD Send a Message to the SGD Curators ,Need to parse out embedded AA S
13、equence,Function to Return AA Sequence Given ORF,create or replace function orf2seq (p_orf in varchar2 ) return varchar2 isv_stream clob;strt number; begin- Retrieve the HTTP stream:v_stream := httpuritype.getclob(httpuritype.createuri(http:/db.yeastgenome.org/cgi-bin/SGD/getSeq?seq=|p_orf|,Web site
14、 URL,RegExp to remove control chars from HTTP stream,Parameterized ORF Id,RegExp to extract AA sequence,Amino Acid Sequence for ORF YDR099W,SQL select orf2seq(YDR099W) from dual;ORF2SEQ(YDR099W) - MSQTREDSVYLAKLAEQAERYEEMVENMKAVASSGQELSVEERNLLSVAYKNVIGARRASWRIVSSIEQKEESKEKSEHQ VELIRSYRSKIETELTKISDDI
15、LSVLDSHLIPSATTGESKVFYYKMKGDYHRYLAEFSSGDAREKATNSSLEAYKTASEI ATTELPPTHPIRLGLALNFSVFYYEIQNSPDKACHLAKQAFDDAIAELDTLSEESYKDSTLIMQLLRDNLTLWTSDISES GQEDQQQQQQQQQQQQQQQQQAPAEQTQGEPTKElapsed: 00:00:01.24,Elapsed time 2 sec. (network latency),SQL insert into pseq (orf_id, sequence) 2 values (YDR099W, orf2seq(Y
16、DR099W);,Case 2: Motif Searching in Proteins,PROSITE database of protein sequence motifsID TYR_PHOSPHO_SITE; PATTERN. AC PS00007; DT APR-1990 (CREATED); APR-1990 (DATA UPDATE); APR-1990 (INFO UPDATE). DE Tyrosine kinase phosphorylation site. PA RK-x(2,3)-DE-x(2,3)-Y. CC /TAXO-RANGE=?E?V; CC /SITE=5,
17、phosphorylation; CC /SKIP-FLAG=TRUE; DO PDOC00007; Source: http:/www.expasy.org/prosite/ps_frequent_patterns.txtTKP Pattern: RK-x(2,3)-DE-x(2,3)-Y. R=Arginine, K=Lysine, D=Aspartate, E=Glutamate, Y=Tyrosine, x=any AA Oracle10g Regular Expression Equivalent RK.2,3DE.2,3Y,TKP,TKP motif pattern,1 Argin
18、ine or Lysine,2 3 Any,1 Aspartate or Glutamate,1 Tyrosine,2 3 Any,SQL Example: Retrieving all Interacting Proteins with TKP,select distinctsubstr(a.refseq_id, 1, 9) refseq_id,length(a.seq_string_varchar) seq_length,regexp_instr(a.seq_string_varchar, RK.2,3DE.2,3Y, 1, 1) motif_offs1,regexp_instr(a.se
19、q_string_varchar, RK.2,3DE.2,3Y, 1, 2) motif_offs2,regexp_instr(a.seq_string_varchar, RK.2,3DE.2,3Y, 1, 3) motif_offs3,regexp_instr(a.seq_string_varchar, RK.2,3DE.2,3Y, 1, 4) motif_offs4 fromtarget_db a,y2h_interaction_p b wherea.refseq_id like NP%and regexp_like(a.seq_string_varchar, RK.2,3DE.2,3Y)
20、and (substr(a.refseq_id,1,9) = b.bait_refseq or substr(a.refseq_id,1,9) = b.prey_refseq) ;,Returns all rows with TKP site,Returns first 4 instances of TKP in each sequence,SQL Example Output,REFSEQ_ID SEQ_LENGTH MOTIF1_OFFS MOTIF2_OFFS MOTIF3_OFFS MOTIF4_OFFS - - - - - - NP_003961 1465 14 202 347 53
21、7 NP_003968 330 241 0 0 0 NP_003983 490 8 50 62 93 NP_004001 3562 3085 0 0 0 .,MHHCKRYRSPEPDPYLSYRWKRRRSYSREHEGRLRYPSRREPPPRRSRSRSHDRLPYQRRYRERRDSDTYRCEERSPSFG EDYYGPSRSRHRRRSRERGPYRTRKHAHHCHKRRTRSCSSASSRSQQSSKRTGRSVEDDKEGHLVCRIGDWLQERYEIVG NLGEGTFGKVVECLDHARGKSQVALKIIRNVGKYREAARLEINVLKKIKEKDKENKFLC
22、VLMSDWFNFHGHMCIAFELLG KNTFEFLKENNFQPYPLPHVRHMAYQLCHALRFLHENQLTHTDLKPENILFVNSEFETLYNEHKSCEEKSVKNTSIRVAD FGSATFDHEHHTTIVATRHYRPPEVILELGWAQPCDVWSIGCILFEYYRGFTLFQTHENREHLVMMEKILGPIPSHMIHR TRKQKYFYKGGLVWDENSSDGRYVKENCKPLKSYMLQDSLEHVQLFDLMRRMLEFDPAQRITLAEALLHPFFAGLTPEER SFHTSRNPSR,RK.2,3DE.2,3Y,Motif #1 a
23、t offset 8,Motif #2 at offset 50,Motif #3 at offset 62,Motif #4 at offset 93,Result: 702 (56%) interacting proteins with TKP site,Is 56% TKP in interacting proteins significant?,Random sample test of all NP entriesN = 33 random samplesSample size 7.4% (1251)Sample mean = 515SD = 17.2Significance lev
24、el 1E-30,Outline,Data-driven Discovery-oriented Computational Framework 10g Regular Expression Case Studies 10g BLAST Case Studies,Similarity Search (Sequence Comparison): A Routine Biology Task,A Query Sequence,n Target Sequences,k Pair-wise Comparison Results,Similarity Search has not been integra
25、ted into the DB system.,Using BLAST can be a laborious process & a data-management hell,Custom setup of BLAST target database Iterate through query sequences: “Batch BLAST” Export/parse/filter/import data DBMS Integration of results with external data,Case 1: Oracle 10g BLASTN as a sequence identifi
26、cation tool,SELECT t.t_seq_id, t.expect FROM TABLE ( BLASTN_MATCH (select sequence FROM query_db where sequence_id = 100),CURSOR (select refseq_id, sequence_string FROM target_db where refseq_id like NM_%) ) t WHERE t.expect 1E-20;,- A sequence fragment with a sequence_id = 100 - Sequence is stored
27、in the query_db table. TACACACCTCGGCGTCGCAGCTCTCGATCATCTCCGGAGCTAACAAGGAAGGCCGGACTGTCCCGTAGAAGCCGCTCTGC,T_SEQ_ID EXPECT - - NM_016333.2 0,Case 2: Discovering “Interlogs”,A,B,X,Y,C,Z,Yeast Protein Interactome,Human Protein Interactome,Homology Mapping,Interlogs: (A|X, B|Y) and (A|X, B|Z),A Computatio
28、nal Intensive Task,Data to use Yeast Protein-Protein Interaction Data Yeast Protein Sequences Human Protein-Protein Interaction Data Human Protein Sequences & Annotations Analysis to prepare Homology search: yeast vs. human proteins Things to consider Collect/parse public data from web Import/export
29、 data for BLAST Connect analysis result to internal data, Missing Data, Laborious,Traditional way? Or inside DBMS?,Pipelining Missing Data Directly into BLASTP Searches,insert into yeast_human_homolog select YDR099W Yeast_ORF_name, t.t_seq_id Human_refseq,t.expect E_Value from TABLE ( BLASTP_MATCH (
30、SELECT orf2seq (YDR099W) FROM dual),CURSOR (SELECT refseq_id, sequence_string FROM target_dbWHERE refseq_id LIKE NP_%) t WHERE t.expect 0.0001 ;, BLAST in DBMS, Online DataIntegration, BLAST Target DB Customization,- Note: Iterate through Yeast ORF Names to perform batch BLAST.,Mission Impossible: A
31、ccomplished,SELECTa.orf_1, a.orf_2, b.human_refseq, b.e_value , c.human_refseq, c.e_value FROM yeast_interaction a, yeast_human_homolog b, yeast_human_homolog c, y2h_interaction_p d WHERE a.orf_1 = b.yeast_ORF_name anda.orf_2 = c.yeast_ORF_name and(b.human_refseq = d.bait_refseq and c.human_refseq =
32、 d.prey_refseq)or(b.human_refseq = d.prey_refseq and c.human_refseq = d.bait_refseq) ;,ORF_1 ORF_2 HUMAN_REFSEQ E_VALUE HUMAN_REFSEQ E_VALUE - - - - - - YCR002C YHR107C NP_xxxxx1 5.9279E-44 NP_yyyyy1 3.7130E-46 YCR002C YJR076C NP_xxxxx2 5.9279E-44 NP_yyyyy2 1.7807E-48 YJR076C YHR107C NP_xxxxx3 1.973
33、4E-39 NP_yyyyy3 3.7130E-46 YCR002C YHR107C NP_xxxxx4 2.3257E-48 NP_yyyyy4 7.4988E-39 YCR002C YJR076C NP_xxxxx5 2.3257E-48 NP_yyyyy5 1.9734E-39 YJR076C YHR107C NP_xxxxx6 1.7807E-48 NP_yyyyy6 7.4988E-39,Conclusion,Data-driven discovery-oriented bioinformatics framework demands rich bio-specific DBMS s
34、upport 10g Regular Expression and BLAST in DBMS features benefit our scientific discovery tasks in interactome studies Additional enhancements,References,Jake Yue Chen, et al (2003) Initial Large-scale Exploration of Protein-protein Interactions in the Human Brain. Proceedings of the IEEE Computer S
35、cience Society Bioinformatics Conference 2003. Stanford University, Stanford, CA.Sudhir Sahasrabudhe and Chen, Jake Yue (2003) Extracting Biological Information from System-scale Protein Interactome Data. Tutorial at the 11th International Conference on Intelligent Systems in Molecular Biology. Brisbane, Australia.Jake Yue Chen and John Carlis (2003) Similar_Join: Extending DBMS with a Bio-specific Operator. Proceedings of the 2003 ACM Symposium on Applied Computing. Melbourne, Florida.Jake Yue Chen and John Carlis (2003) Genomic Data Modeling. Information Systems, Vol 28, issue 4: p287-310.,
