Web Services for N-Glycosylation Process Satya S. Sahoo, Amit P. Sheth, William S. York, John A. Miller Presentation at International Symposium on Web Services For Computational Biology and Bioinformatics, VBI, Blacksburg, VA, May 26-27, 2005 Integrated Technology Resource for Biomedical Glycomics NCRR/NIH Glycomics Study of structure, function and quantity of ‘complex carbohydrate’ synthesized by an organism Carbohydrates added to basic protein structure Glycosylation Folded protein structure (schematic) 2 Glycosylation – why is it important? Genome (comprised of DNA) or Proteome (proteins) are not the only factors in life functions of an organism Carbohydrates attached to different protein structures (by glycosylation) are important for: Identification of foreign entities by immune system cells Markers to accurately diagnose diseases Regulate signaling activities Categorization of glycosylation - the way carbohydrates are attached to proteins. Example: N-glycosylation 3 N-Glycosylation Process (NGP) Cell CultureBy N-glycosylation Process, extract we mean the identification and Glycoprotein Fraction quantification of proteolysis glycopeptides Glycopeptides Fraction 1 n Separation technique I Glycopeptides Fraction n PNGase Peptide Fraction Separation technique II n*m Peptide Fraction Mass spectrometry ms data ms/ms data Data reduction ms peaklist ms/ms peaklist binning Glycopeptide identification and quantification 4 N-dimensional array Signal integration Data reduction Peptide identification Peptide list Data correlation NGP – part of the Bioinformatics core Integrated Technology Resource for Biomedical Glycomics This Resource was established by the National Center for Research Resources The aim is to develop the tools and technology to analyze glycoprotein and glycolipid expression of embryonic stem cells Our research provides bioinformatics support for four research groups: Embryonic Stem Cell Culture Program Glycomic Analysis of Glycoproteins Glycomic Analyses of Glycosphingolipids and Sphingolipids Transcript analysis by kinetic RT-PCR 5 NGP – need in Glycomics Unlike proteomics or genomics, high-throughput experimental protocols are still being established in Glycomics NGP involves a multitude of heterogeneous tasks, including human-mediated tasks NGP attempts to encapsulate particular computational steps as platform-independent, scalable and Web-accessible tools – Web Services Enables glycobiologists to integrate automated data generation tasks with data processing tools (Web Services) end-toend experimental lifecycle 6 N-Glycosylation identification - Problems Extremely difficult to identify glycosylated sequences using standard analytical methods peptide N-glycosylation occurs at particular sites on the protein structure – consensus sequences Asparagine Aspartate Consensus Sequence Peptide N D J X S/T PNGaseF Glycan 7 An example glycopeptide (schematic) NGP - implementation NGP,currently,implements a Web Process constituted of two Web Services: DB Modifier Web Service – modifies the search database by replacing N (in consensus sequences) by J Collator Web Service – identifies a probable N-glycosylated peptide, using three parameters: Calculated molecular mass Presence of ‘J’ in a peptide sequence MASCOT* Score assigned to a hit NGP also involves propriety Mass Spectrometer search engine service (MASCOT*) as an intermediate task Hence, NGP Web Process identifies probable glycosylated peptides – enabling rapid processing of data from high throughput experiment 8 *http://www.matrixscience.com/ NGP – Architecture (current) PEAK LIST FILE ms/ms raw data Primary Sequence Database ModifyDB Web Service MASCOT* Mass Spectrometer Search Engine Collator Web Service Deglycosylated peptide list 9 *http://www.matrixscience.com/ MASCOT* output file (contains both glycosylated and nonglycosylated peptide sequences) NGP Results q1_p1=-1 q2_p1=0,626.349945,-0.023321,2,APGVAGR,18,000000000,1.49,00020000000000000,0,0;"gi|51465537":0:190:196:1 q2_p2=1,626.361191,-0.034567,2,APARGR,18,00000000,1.33,00020000000000000,0,0;"gi|10140845":0:2:7:2 q2_p3=0,626.349945,-0.023321,2,APAVGGR,18,000000000,1.33,00020000000000000,0,0;"gi|51470766":0:212:218:1,"gi|51470768":0:212:218:1 q3_p3=0,634.368973,0.006151,4,DIIFK,12,0000000,25.26,00010020000000000,0,0;"gi|47078238":0:364:368:2,"gi|47078240":0:328:332:2 q3_p4=0,634.351227,0.023897,4,MPLFK,12,0000000,25.24,00010020000000000,0,0;"gi|41197108":0:95:99:1,"gi|4557311":0:1:5:2 q3_p5=0,634.343811,0.031313,3,NNLFK,12,0000000,15.34,00010020000000000,0,0;"gi|31377725":0:539:543:1 q3_p6=0,634.368973,0.006151,3,LDIFK,12,0000000,15.34,00010020000000000,0,0;"gi|39725634":0:891:895:1 q3_p7=0,634.343811,0.031313,3,NNIFK,12,0000000,15.34,00010020000000000,0,0;"gi|7661646":0:212:216:1 q3_p8=0,634.368973,0.006151,3,LDLFK,12,0000000,15.34,00010020000000000,0,0;"gi|51474898":0:237:241:1 q3_p9=0,634.368958,0.006166,3,EVIFK,12,0000000,13.61,00010020000000000,0,0;"gi|28376662":0:67:71:1 q3_p10=0,634.368958,0.006166,3,VELFK,12,0000000,13.61,00010020000000000,0,0;"gi|51467300":0:493:497:1,"gi|51467535":0:99:103:1 q4_p1=-1 q5_p1=0,662.375122,0.004702,5,DLLFR,14,0000000,18.41,00020020000000000,0,0;"gi|21536369":0:84:88:1,"gi|21536367":0:17:21:1,"gi|4557871":0:647:651:1 q5_p2=0,662.375122,0.004702,3,DLFLR,14,0000000,12.81,00010020000000000,0,0;"gi|33695153":0:407:411:1,"gi|4504043":0:330:334:1,"gi|11968045":0:6:10:1 q5_p3=0,662.375122,0.004702,3,DIFIR,14,0000000,12.81,00010020000000000,0,0;"gi|4505725":0:924:928:1,"gi|29788751":0:1170:1174:1 q5_p4=0,662.349960,0.029864,3,NNFIR,14,0000000,11.84,00010020000000000,0,0;"gi|24416002":0:667:671:1 q5_p5=0,662.375122,0.004702,4,IDLFR,14,0000000,9.98,00020020000000000,0,0;"gi|12957488":0:602:606:1,"gi|41148707":0:536:540:1,"gi|51464463":0:646:650:1 q5_p6=0,662.375122,0.004702,4,LDLFR,14,0000000,9.98,00020020000000000,0,0;"gi|42657517":0:335:339:1 q5_p7=0,662.375107,0.004717,4,VELFR,14,0000000,9.98,00020020000000000,0,0;"gi|6912230":0:436:440:1 q5_p8=0,662.375122,0.004702,4,LDIFR,14,0000000,9.98,00020020000000000,0,0;"gi|8922081":0:2699:2703:1 q5_p9=0,662.349960,0.029864,4,NLNFR,64,0000000,5.89,00010020000000000,0,0;"gi|19923416":0:816:820:1 q5_p10=1,662.361191,0.018633,2,NRFAR,14,0000000,3.37,00010020000000000,0,0;"gi|4758704":0:97:101:1 q6_p1=0,674.359863,-0.006639,4,VSDNIK,35,00000000,11.27,00010020000000000,0,0;"gi|32130516":0:935:940:1 q6_p2=0,674.323456,0.029768,5,EGDLGGK,21,000000000,7.97,00020020000000000,0,0;"gi|13569928":0:1058:1064:1 q6_p3=0,674.359848,-0.006624,5,EATVAGK,21,000000000,7.88,00020020000000000,0,0;"gi|51475822":0:527:533:1 q6_p4=1,674.389740,-0.036516,3,QRMLK,14,0000000,7.46,00020010000000000,0,0;"gi|24307905":0:467:471:2,"gi|24307905":0:638:642:2 q6_p5=0,674.359863,-0.006639,5,LSSSPGK,56,000000000,7.38,00000020000000000,0,0;"gi|8922075":0:806:812:1 q6_p6=0,674.338730,0.014494,4,WDLGGK,42,00000000,6.40,00010020000000000,0,0;"gi|13375817":0:123:128:1 q6_p7=0,674.359879,-0.006655,4,QATDLK,56,00000000,6.21,00020010000000000,0,0;"gi|21361684":0:451:456:1 q6_p8=1,674.371094,-0.017870,3,QTNKGK,14,00000000,6.03,00020010000000000,0,0;"gi|41117716":0:85:90:1 q6_p9=1,674.389740,-0.036516,6,QMRIK,28,0000000,5.77,00020020000000000,0,0;"gi|28329439":0:269:273:1,"gi|28558993":0:278:282:1 q6_p10=1,674.389740,-0.036516,6,QMRLK,28,0000000,5.77,00020020000000000,0,0;"gi|40255096":0:300:304:1 q7_p1=0,695.348969,0.007855,4,YDASLK,14,00000000,8.86,00020020000000000,0,0;"gi|4758454":0:2761:2766:1 A typical MASCOT output file is about 3MB! High-throughput experiment protocol generate thousands of such files - manual identification is not feasible 10 NGP Web Services – Adding Semantics Two Ontologies developed as part of the NCRR-Glycomics project: GlycO: a domain Ontology embodying knowledge of the structure and metabolisms of glycans Contains 770 classes – describe structural features of glycans URL: http://lsdis.cs.uga.edu/projects/glycomics/glyco ProPreO: a comprehensive process Ontology modeling experimental proteomics Contains 296 classes Models three phases of experimental proteomics* – Separation techniques, Analytical techniques and, Data analysis URL: http://lsdis.cs.uga.edu/projects/glycomics/propreo 11 *http://pedro.man.ac.uk/uml.html (PEDRO UML schema) ProPreO - Experimental Proteomics Process Ontology ProPreO models the phases of proteomics experiment using five fundamental concepts: Data: (Example: a peaklist file from ms/ms raw data) Data_processing_applications: (Example: MASCOT* search engine) Hardware: embodies instrument types used in proteomics (Example: ABI_Voyager_DE_Pro_MALDI_TOF) Parameter_list: describes the different types of parameter lists associated with experimental phases Task: (Example: chromatography) 12 *http://www.matrixscience.com/ component separation, used in Service description using WSDL-S Formalize description and classification of Web Services using ProPreO concepts <?xml version="1.0" encoding="UTF-8"?> <?xml version="1.0" encoding="UTF-8"?> <wsdl:definitions targetNamespace="urn:ngp" <wsdl:definitions targetNamespace="urn:ngp" …… ….. xmlns: xmlns:xsd="http://www.w3.org/2001/XMLSchema"> wssem="http://www.ibm.com/xmlns/WebServices/WSSemantics" xmlns: <wsdl:types> ProPreO="http://lsdis.cs.uga.edu/ontologies/ProPreO.owl" > <schema targetNamespace="urn:ngp“ xmlns="http://www.w3.org/2001/XMLSchema"> <wsdl:types> ….. <schema targetNamespace="urn:ngp" </complexType> xmlns="http://www.w3.org/2001/XMLSchema"> </schema> …… </wsdl:types> </complexType> <wsdl:message name="replaceCharacterRequest"> </schema> <wsdl:part name="in0" type="soapenc:string"/> </wsdl:types> <wsdl:part name="in1" type="soapenc:string"/> <wsdl:message name="replaceCharacterRequest" <wsdl:part name="in2" type="soapenc:string"/> wssem:modelReference="ProPreO#peptide_sequence"> </wsdl:message> <wsdl:part name="in0" type="soapenc:string"/> <wsdl:message name="replaceCharacterResponse"> <wsdl:part name="in1" type="soapenc:string"/> <wsdl:part name="replaceCharacterReturn" type="soapenc:string"/> <wsdl:part name="in2" type="soapenc:string"/> </wsdl:message> </wsdl:message> 13 WSDL ModifyDB WSDL-S ModifyDB data Description of a sequence Web Service using: Web Service Description peptide_sequence Language Concepts defined in process Ontology ProPreO process Ontology Biological UDDI (BUDDI) WS Registry for Proteomics and Glycomics There are no current registries that use semantic classification of Web Services in glycoproteomics BUDDI classification based on proteomics and glycomics classification – part of integrated glycoproteomics Web Portal called Stargate NGP to be published in BUDDI Can enable other systems such as myGrid to use NGP Web Services to build a glycomics workbench 14 Conclusions As part of NCRR Integrated Technology Resource for Biomedical Glycomics, we implemented a Semantic Web Process for high throughput glycomics in open, web-centric environment Large domain specific ontologies with process (ProPreO) and domain (GlycO) knowledge concepts was used to describe and classify Web Services – at Semantic level Used proposed Semantic Web Service specification (WSDL-S) to add semantics to Web Service description Biological UDDI (BUDDI) – part of Stargate is being developed as a single-window resource to discover and publish Web Services in glycoproteomics domain 15 Resources NCRR (Integrated Technology Resource for Biomedical Glycomics): http://cell.ccrc.uga.edu/world/glycomics/glycomics.php Bioinformatics core of Glycomics project: http://lsdis.cs.uga.edu/projects/glycomics/ ProPreO process Ontology: http://lsdis.cs.uga.edu/projects/glycomics/propreo/ GlycO domain Ontology: http://lsdis.cs.uga.edu/projects/glycomics/glyco/ Stargate – GlycoProteomics Web Portal: http://128.192.9.86/stargate WSDL-S: joint UGA-IBM technical note http://lsdis.cs.uga.edu/library/download/WSDL-S-V1.pdf 16 Acknowledgement Special Thanks: James Atwood (CCRC, UGA) Meenakshi Nagarajan (LSDIS Lab, UGA) Blake Hunter (LSDIS Lab, UGA) 17 Extra Slides: Stargate subsystems – a bit of detail BUDDI – BioUDDI is envisioned as the ‘yellow pages’ for all WS in life sciences The classification of WS uses biological taxonomy Open resource for the worldwide community of life sciences research Format Converter – Enables conversion of two available representation formats into a xml-based representation IUPAC to LINUCS to GLYDE (a xml-based representation) Web Service Generator – Enables existing java application to be exposed as Web Services Generates required files from a java application to allow deployment as a Web Service Enable the newly generated Web Service to be published on BioUDDI 18 Extra Slides: Stargate subsystems – a bit of detail Group Forum – Members of the research group use it to foster a sense of community Schedule meetings, discuss issues, collaborate on papers… Post papers for peer reviews, publications on relevant topic Stargate Search – is an integrated unit of the Stargate Enables search for research publication within the research group Enables search on the internet Login – Allows restrictions on accessibility of selected parts of Stargate 19 Extra Slides: The take home message… Forum Internet Search 20 Web Service Generator BUDDI