Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists

DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better... Nucleic Acids Research, 2007, Vol. 35, Web Server issue W169–W175 doi:10.1093/nar/gkm415 DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists 1 1 1 1 2 Da Wei Huang , Brad T. Sherman , Qina Tan , Joseph Kir , David Liu , 2 5 2 3 David Bryant , Yongjian Guo , Robert Stephens , Michael W. Baseler , 4 1, H. Clifford Lane and Richard A. Lempicki * 1 2 3 Laboratory of Immunopathogenesis and Bioinformatics, Advanced Biomedical Computing Center, Clinical Services Program, SAIC-Frederick, Inc., National Cancer Institute at Frederick, MD 21702, USA, Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, 20892, USA, Bioinformatics and Scientific IT Program, NIAID Office of Technology Information Systems, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, 20892, USA Received January 22, 2007; Revised April 14, 2007; Accepted May 6, 2007 ABSTRACT bio-pathways and more. With DAVID (http://david. niaid.nih.gov), investigators gain more power to All tools in the DAVID Bioinformatics Resources interpret the biological mechanisms associated aim to provide functional interpretation of large lists with large gene lists. of genes derived from genomic studies. The newly updated DAVID Bioinformatics Resources consists of the DAVID Knowledgebase and five integrated, INTRODUCTION web-based functional annotation tool suites: In the post-genomic era, biological interpretation of the DAVID Gene Functional Classification Tool, the large gene lists derived from high-throughput experiments, DAVID Functional Annotation Tool, the DAVID Gene such as genes from microarray experiments, is a challen- ID Conversion Tool, the DAVID Gene Name Viewer ging task. The first version of DAVID (the Database for and the DAVID NIAID Pathogen Genome Browser. Annotation, Visualization and Integration Discovery), The expanded DAVID Knowledgebase now inte- released in 2003 (1,2), as well as a number of other similar grates almost all major and well-known public publicly available high-throughput functional annotation bioinformatics resources centralized by the DAVID tools (3–23), partially address the challenge by system- Gene Concept, a single-linkage method to agglom- atically mapping a large number of interesting genes in a list to associated Gene Ontology (GO) terms (10), and erate tens of millions of diverse gene/protein then statistically highlighting the most over-represented identifiers and annotation terms from a variety of (enriched) GO terms out of a list of hundreds or public bioinformatics databases. For any uploaded thousands of terms. This increases the likelihood that gene list, the DAVID Resources now provides not the investigator will identify the biological processes only the typical gene-term enrichment analysis, most pertinent to the biological phenomena under but also new tools and functions that allow users study (19). While this tool is extremely useful and has to condense large gene lists into gene functional been cited in hundreds of publications during the past groups, convert between gene/protein identifiers, three years, the development of other effective data mining visualize many-genes-to-many-terms relationships, algorithms, as additional components to the DAVID cluster redundant and heterogeneous terms into Bioinformatics Resources, will improve the power of groups, search for interesting and related genes or investigators to analyze their gene lists from different terms, dynamically view genes from their lists on biological angles. The newly added contents, functions *To whom correspondence should be addressed. Tel: +1-301-846-7114; Fax: 301-846-7672; Email: [email protected] The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. 2007 The Author(s) This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. W170 Nucleic Acids Research, 2007, Vol. 35, Web Server issue Table 1. Over 22 types of gene identifiers integrated by the DAVID and tool suites in the DAVID Bioinformatics Resources Gene Concept within the DAVID Knowledgebase intend to address several issues that other tools have not been able to extensively address: (i) to dramatically Gene ID Type Total ID Unique Cluster expand the biological information coverage in the DAVID Knowledgebase by comprehensively integrating AFFY_ID 2254679 845117 more than 20 types of major gene/protein identifiers ENTREZ_GENE_ID 1734858 1602339 GENPEPT_ACCESSION 4065385 2511637 and more than 40 well-known functional annotation GENBANK_ACCESSION 16828735 2409120 categories from dozens of public databases; (ii) to address GENEBANK_ID 20291282 2358084 the enriched and redundant relationships among many- PIR_ACCESSION 282281 258079 genes-to-many-terms (i.e. one gene could associate with PIR_ID 308092 266645 PIR_NREF_ID 3355759 2677404 many different, redundant terms and one term could REFSEQ_GENOMIC 1866800 1552597 associate with many genes) by developing a set of REFSEQ_MRNA 645831 561447 novel algorithms, such as the DAVID Gene Functional REFSEQ_PROTEIN 1644632 1373467 Classification Tool, the Functional Annotation Clustering REFSEQ_RNA 1364 852 UNIGENE 161138 158938 Tool, the Linear Searching Tool, the Fuzzy Gene-Term UNIPROT_ACCESSION 2864344 2097488 Heat Map Viewer, etc.; (iii) to dynamically visualize genes UNIPROT_ID 2789453 2096712 from a users list within the most relevant KEGG and UNIREF100_ID 2552342 2088692 BioCarta pathways with the DAVID Pathway Viewer; OFFICIAL_GENE_SYMBOL 1693151 1600906 FLYBASE_ID 27109 26642 (iv) to allow users to create and use customized gene HAMAP_ID 63925 63822 backgrounds for typical gene-term enrichment analysis HSSP_ID 265000 258750 utilizing the improved computational power and (v) to TIGR_ID 120117 111699 facilitate efficient communication and experience WORMBASE_ID 43675 21243 RGD_ID 25230 25060 exchange within the scientific community by moderating NOT SURE ALL IDs the DAVID Forum. This article summarizes the key DAVID components Any of the gene identifier types above can be cross-mapped to the and tool suites in the newly released DAVID DAVID Knowledgebase. ‘Not Sure’ is a new ID type specifically designed for the DAVID web site. For a given ‘not sure’ ID, all Bioinformatics Resources, highlighting new or expanded possible matching IDs will be systematically scanned across the entire analytic features that provide investigators with additional DAVID collection. means to explore and extract biological meaning from large gene lists that users input to the system (Supplementary File 1). For in-depth algorithm informa- tion, appropriate references and supplementary materials DAVID Knowledgebase are provided. A highly integrated gene-annotation database with comprehensive data coverage is essential for the success of any high-throughput annotation algorithms. Due to the FEATURES AND FUNCTIONALITIES complex and distributed nature of biological research, our current biological knowledge is distributed among many Computational Infrastructure redundant annotation databases maintained by indepen- The aim of the DAVID software design is to provide dent groups. One gene could have several different users with the simplest usability and fastest exploration identifiers within one or more database(s). Similarly, the speed through better internal software engineering biological terms associated with different gene identifiers practices. Therefore, the DAVID Bioinformatics Tools, for the same gene could be collected in different levels as web-based applications on a Tomcat web server across different databases. Due to these issues, most high- in a Linux machine (4-CPU for 3.5 GHz speed, 8 GB throughput annotation tools rely on one, or at most a few, memory), requires no configuration and installation in resource(s), which limits the analytic comprehensiveness the client’s computers. Java is the primary language and the level of throughput. The DAVID Knowledgebase used for all of the server side components of the is now built around the ‘DAVID Gene Concept’, a single calculation engines and the Java Server Page (JSP) web linkage method to agglomerate tens of millions of interfaces, in a full object-oriented fashion. In-memory gene/protein identifiers from a variety of public genomic Java data objects holding all genes-to-annotation resources (Table 1), including NCBI, PIR and UniProt information up to 2.5 GB in size were developed to (24–27), into broader secondary gene clusters, called greatly increase the data IO speed compared to that the DAVID Gene Concept (Figure 1, and more tech- through typical relational databases (e.g. Oracle). The nical details at http://david.abcc.ncifcrf.gov/helps/ Java Remote Method Invocation (RMI), a distributed knowledgebase/DAVID_gene.html), Grouping these computing technique, is also used to take advantage gene identifiers improves cross-referencing capability, of multiple computing resources. A set of automated allowing more than 40 categories of publicly available programs monitors many aspects of the web services in functional annotation to be comprehensively assigned to order to maximize the performance and minimize and centralized by the DAVID Gene Concept (Table 2, the down time period. see Supplementary File 2 for a complete list of annotation Nucleic Acids Research, 2007, Vol. 35, Web Server issue W171 NREF(NF00095014) DAVID Gene (2858470) PIR_ID : I38140 UniRef100_Q16825 GenPept : CAA56042 UniRef100_Q16825 Swissprot : PTN21_HUMAN RefSeq : NP_008970 Swissprot:PTN21_HUMAN Uniprot : Q16825 Uniprot : Q16825 Entrez Gene (11099) UniRef100_Q8WX29 Uniprot : Q16825 Genepept :: CAA56042 Swissprot : Q8WX29_HUMAN Refseq : NP_008970 Uniprot : Q8wx29 Uniprot : Q16825 NREF : NF00095014 Genepept :: CAD19000 PIR_ID : I38140 UniRef100_Q8WX29 NREF(NF00828766) Uniprot : Q16825 Swissport:Q8WX29_HUMAN Uniprot : Q8WX29 Genpept : CAA56042 Uniprot : Q8wx29 Genpept : CAD19000 RefSeq : NP_008970 NREF : NF00828766 Genpept : CAD19000 Entrez Gene : 11099 Figure 1. A DAVID gene constructed by a single linkage algorithm. Two UniRef100 clusters, two NRef 100 clusters and one Entrez Gene cluster were systematically found sharing one or more protein identifiers with each other. The single-linkage rule can further iteratively agglomerate them as a whole into one DAVID gene. Thus, for this particular example of tyrosine-protein phosphatase non-receptor type 21 (PTPN21), the resulting DAVID gene is able to collect and integrate all gene/protein identifiers more comprehensively than each original gene cluster. Table 2. The wide-range collection of heterogeneous functional annotations in the DAVID Knowledgebase Ontology (440 million records) Protein Domain/Family (415 millions) Sequence Features (421 millions) GO_BIOLOGICAL PROCESS BLOCKS_ID ALIAS_GENE_SYMBOL GO_MOLECULAR FUNCTION COG_KOG_NAME CHROMOSOME GO_CELLULAR COMPONENT INTERPRO_NAME CYTOBAND PANTHER_BIOLOGICAL PROCESS PDB_ID GENE_NAME PANTHER_MOLECULAR FUNCTION PFAM_NAME GENE_SYMBOL COG_KOG_ONTOLOGY PIR_ALN HOMOLOGOUS_GENE P-P Interaction (44 millions) PIR_HOMOLOGY_DOMAIN ENTREZ_GENE_SUMMARY BIND PIR_SUPERFAMILY_NAME OMIM_ID DIP PRINTS_NAME PIR_SUMMARY MINT PRODOM_NAME PROTEIN_MW NCICB_CAPATHWAY PROSITE_NAME REFSEQ_PRODUCT TRANSFAC_ID SCOP_ID SEQUENCE_LENGTH HIV_INTERACTION SMART_NAME SP_COMMENT HIV_INTERACTION_CATEGORY TIGRFAMS_NAME Functional Category (46.9 millions) HPRD_INTERACTION PANTHER_SUBFAMILY PIR_SEQ_FEATURE REACTOME_INTERACTION PANTHER_FAMILY SP_COMMENT_TYPE Disease Association (9,000) Pathways (450 000) SP_PIR_KEYWORDS GENETIC_ASSOCIATION_DB BioCarta UP_SEQ_FEATURE OMIM_DISEASE KEGG_PATHWAY Gene Tissue Expression (41.0 million) Literature (42.8 millions) PANTHER_PATHWAY GNF Microarray GENERIF_SUMMARY PID UNIGENE EST PUBMED_ID BBID CGAP SAGE HIV_INTERACTION_PUBMED_ID KEGG_REACTION CGAP EST Over 60 functional categories from dozens of independent public sources (databases) (see Supplementary File 2 for a complete list) are collected and integrated in the DAVID Knowledgebase. sources and more technical details at http://david.abcc. enhances the comprehensiveness of high-throughput gene ncifcrf.gov/helps/knowledgebase/DAVID_gene.html). To functional analysis by overlapping multiple biological the best of our knowledge, this annotation coverage far aspects together. It also provides a solid foundation for exceeds that of the original DAVID database and those the further development of more advanced high through- currently used by other similar high-throughput annota- put analytic algorithms that may be added to the DAVID tion tools. The DAVID knowledgebase not only increases Bioinformatics Resources. More importantly, the entire the accessibility to a wide range of heterogeneous DAVID Knowledgebase, in simple pair-wise text format annotation data in one centralized location, but also files containing a broad, highly integrated annotation W172 Nucleic Acids Research, 2007, Vol. 35, Web Server issue data collection, is freely available to the public (http:// DAVID Gene Functional Classification Tool Suite david.abcc.ncifcrf.gov/knowledgebase), which will benefit The DAVID Gene Functional Classification Tool (http:// various high-throughput data mining projects by other david.abcc.ncifcrf.gov/gene2gene.jsp) is a completely new research groups. The DAVID Knowledgebase is expected component in the DAVID Bioinformatics Resources. to be updated more frequently in the near future than The tool provides a novel way to functionally analyze a its current annual update. large number of genes in a high-throughput fashion by classifying them into gene groups based on their annota- tion term co-occurrence. This is accomplished and DAVID Functional Annotation Tool Suite visualized by a set of new fuzzy classification algorithms, including a kappa statistics measurement of gene–gene This tool suite (http://david.abcc.ncifcrf.gov/summary. functional relationship, a fuzzy multi-linkage partitioning jsp), introduced in the first version of DAVID, mainly method and a fuzzy genes-terms heat map visualization, provides typical batch annotation and gene-GO term etc. (manuscript submitted, and more details at http:// enrichment analysis to highlight the most relevant GO david.abcc.ncifcrf.gov/manuscripts/fuzzy_cluster/). The terms associated with a given gene list (2). The new version power of the tool is that it allows users to simultaneously of the tool keeps the same enrichment analytic algorithm view the rich and redundant internal relationship of but with extended annotation content coverage, increasing functionally related genes and their annotation terms from only GO in the original version of DAVID to within biological modules. Investigators are able to currently over 40 annotation categories, including GO functionally analyze their gene list in a highly related terms, protein–protein interactions, protein functional many-genes-to-many-terms network context instead of domains, disease associations, bio-pathways, sequence a one-term-to-many-genes or a one-gene-to-many-terms general features, homologies, gene functional summaries, view in the typical gene-annotation enrichment analysis. gene tissue expressions, literatures, etc. (Table 2). The improved annotation coverage alone provides investiga- DAVID Gene ID Conversion Tool Suite tors with much more power to analyze their genes using A significant number of different types of gene/protein many different biological aspects in a single space. Flexible identifiers, not mutually mapped to each other across options are provided to display results in an individual three independent resources, NCBI, PIR and UniProt annotation chart report or a combined chart report. (25,26,28), are now maximally integrated in the DAVID In addition to pre-built gene population backgrounds Knowledgebase (Figure 1, more details at http://david. (e.g. Affy U133) used in gene-annotation enrichment abcc.ncifcrf.gov/helps/knowledgebase/ analysis, with its improved computational power, the new DAVID_gene.html), whose scope is more expansive than tool accepts user-defined population gene list, an option one system only. Even though the DAVID rarely found in other similar web-based, high-throughput Knowledgebase is used primarily for improvement of annotation tools. This feature was added in order to more annotation terms integration and coverage, such compre- specifically meet the users’ requirements for the best hensive gene identifier coverage and cross-referencing analytical results. capability could itself be very useful for researchers to The DAVID Functional Annotation Clustering is convert their gene/protein identifiers from one type to a newly added feature (manuscript submitted, and another among over 20 major types of identifier systems more details at http://david.abcc.ncifcrf.gov/manuscripts/ (Table 1). Thus, with the newly introduced DAVID fuzzy_cluster/) to the DAVID Functional Annotation Gene ID Conversion Tool (http://david.abcc.ncifcrf.gov/ Tool. This function uses a novel algorithm to measure conversion.jsp), interesting genes derived from one identi- relationships among the annotation terms based on the fier system can be quickly translated to other gene degrees of their co-association genes to group the similar, identifier types preferred by a given annotation resource. redundant and heterogeneous annotation contents from In addition, the DAVID Gene ID Conversion Tool the same or different resources into annotation groups. provides a ‘not sure’ type for ambiguous gene identifiers, This reduces the burden of associating similar redundant whereby the tool can systematically suggest the potential terms and makes the biological interpretation more type(s). For instance, a user has a gene ID ‘3558’ without focused in a group level (Figure 2). The tool also provides ID type information. DAVID Gene ID Conversion Tool a look at the internal relationships among the clustered will scan all possibilities across all gene ID systems terms. The clustered format is able to give a more collected in the DAVID Knowledgebase. Two choices insightful view about the relationships of annotations will be suggested, i.e. ‘3558’ could be an Entrez Gene ID compared to the traditional un-clustered term report, over for IL2 (human) or a Genbank ID for CNA1 (yeast). which similar annotation terms may be spread among Thus, the user can make a decision based on above hundreds, if not thousands, of other terms. In addition, information. to take full advantage of the well-known KEGG and BioCarta pathways, the new DAVID Pathway Viewer, DAVID Gene Name Batch Viewer another feature of the DAVID Functional Annotation Tool, can display genes from a user’s list on pathway After obtaining a list of interesting genes, probably the maps to facilitate biological interpretation in a network first question researchers will ask is ‘What are the names context. of my genes?’ Even though it is a simple question, Nucleic Acids Research, 2007, Vol. 35, Web Server issue W173 Figure 2. An HTML report from the Functional Annotation Clustering. The annotation cluster 1 in the example shows that GO term cytokine activity, KEGG pathway cytokine–cytokine receptor interaction, and GO term receptor binding, etc. are grouped together. Thus, the different biological aspects regarding a relevant biology can be explored at the same time. most high-throughput annotation tools do not answer it in co-occurrence of annotations between genes (more details a straightforward way. The new DAVID Gene Name at http://david.abcc.ncifcrf.gov/helps/linear_search.html). Batch Viewer is designed to simply list the gene names for all given genes. In addition, hyperlinks are provided on DAVID NIAID Pathogen Browser each gene entry, allowing users to explore in depth other functional information around the gene. Thus, this tool The National Institute of Allergy and Infectious Diseases provides users with a first glance and initial ideas about (NIAID) has defined category A, B and C priority their interesting genes before proceeding to analysis by pathogens (http://www3.niaid.nih.gov/Biodefense/bandc_ other more comprehensive analytic tool. Moreover, priority.htm), which have subsequently become important hyperlinks, labeled as ‘RT’, are provided for each gene in biodefense research funding, attracting broad interest in order to search other functionally related genes in user’s from the research community. Since the organisms listed gene list or the entire genome. The search is based on in these categories may not be familiar to researchers who W174 Nucleic Acids Research, 2007, Vol. 35, Web Server issue ++ Highly Applied + Relevant Initial glance of major biological functions associated with my gene list ++ ++ ++ + + Which biological terms/functions are specifically enriched in my gene list? ++ ++ View the genes in my list on related biological pathways ++ ++ Which diseases are associated with my gene list? ++ ++ Which protein functional domains are associated with my gene list ? ++ ++ Which other genes frequently interact with the genes in my list? ++ ++ How to group the highly redundant annotations into group? ++ What are the major gene functional groups in my gene list? ++ ++ View related annotation and related genes on a single graphic view ++ What are other functionally similar genes in genome, but not in my list? + + ++ ++ What are other annotations functionally similar to my interesting one? ++ ++ What are the gene names in my list? +++ How to convert my gene IDs to other type of IDs? +++ How to directly link to DAVID functions ? ++ How can I download DAVID data for in-house study? ++ + + ++ Figure 3. A roadmap to choose appropriate DAVID functions and tools. have recently joined the emerging field, the DAVID CONCLUSION NIAID Pathogen Browser (http://david.abcc.ncifcrf.gov/ The newly released DAVID Bioinformatics Resources are GB.jsp) is provided as a quick starting point for them to an expanded version of the original DAVID. It provides a search the most relevant genes in the organisms by set of powerful, novel tools that researchers can use to biological key words of interests. A large list of genes explore their large gene lists in depth from many different retrieved from the search could be further transferred to biological angles (Figure 3) in order to extract associated the DAVID Bioinformatics Resources for in-depth func- biological meanings to the greatest extent possible. The tional analysis with any of the previously mentioned tools. advanced data collection in the DAVID Knowledgebase Although the tool is still in its early stage, it may help not only creates a solid annotation data foundation for researchers gain understanding of the genes related to a the various DAVID analytic tools, but also is freely priority pathogen of interest. More development is available to the public in a simple pair-wise text format ongoing to extend the searching scope to all available to promote the development of novel annotation algo- genomes and annotations collected in DAVID rithms and techniques within the scientific community. knowledgebase. The DAVID Bioinformatics Resources are accessible at http://david.niaid.nih.gov. DAVID API Services SUPPLEMENTARY DATA DAVID API services (http://david.abcc.ncifcrf.gov/api/) are newly added features that allow users to directly Supplementary Data are available at NAR Online. pass gene list to various DAVID tools via a set of pre- defined URLs instead of DAVID submission forms. ACKNOWLEDGEMENTS Thus, DAVID tools can easily serve as part of the analytic pipeline in other bioinformatics web sites. They The authors are grateful to the referees and editors for can also be used in bioinformatics scripts to automate their constructive comments. Thanks goes to Melaku functional annotation for large number of gene lists, Gedil, Ping Ren, and Jun Yang in the LIB group for which are too many to be accomplished by the manual biological discussions. We also thank Bill Wilton and procedures. Mike Tartakovsky for information technology and Functional Annotation Chart Functional Annotation Clustering Functional Annotation Table Gene Functional Classification Gene Name Batch Viewer Gene ID Conversion Tool DAVID Knowledge base DAVID API Nucleic Acids Research, 2007, Vol. 35, Web Server issue W175 16. Zhang,B., Schmoyer,D., Kirov,S. and Snoddy,J. (2004) GOTree network support. This research was supported in whole by Machine (GOTM): a web-based platform for interpreting sets of the National Institute of Allergy and Infectious Disease. interesting genes using Gene Ontology hierarchies. BMC This project has been funded in whole with federal funds Bioinformatics, 5, 16. from the National Cancer Institute, National Institutes of 17. Zeeberg,B.R., Qin,H., Narasimhan,S., Sunshine,M., Cao,H., Kane,D.W., Reimers,M., Stephens,R.M., Bryant,D. et al. (2005) Health, under contract N01- CO-12400. The content of High-Throughput GoMiner, an ‘industrial-strength’ integrative gene this publication does not necessarily reflect the views or ontology tool for interpretation of multiple-microarray experiments, policies of the Department of Health and Human Services, with application to studies of Common Variable Immune Deficiency nor does mention of trade names, commercial products, (CVID). BMC Bioinformatics, 6, 168. 18. Ben-Shaul,Y., Bergman,H. and Soreq,H. (2005) Identifying subtle or organizations imply endorsement by the U.S. interrelated changes in functional gene categories using continuous Government. Funding to pay the Open Access publication measures of gene expression. Bioinformatics, 21, 1129–1137. charges for this article was provided by the same source as 19. Khatri,P. and Draghici,S. (2005) Ontological analysis of gene above. expression data: current tools, limitations, and open problems. Bioinformatics, 21, 3587–3595. 20. Robinson,P.N., Wollstein,A., Bohme,U. and Beattie,B. (2004) Conflict of interest statement. None declared. Ontologizing gene-expression microarray data: characterizing clusters with Gene Ontology. Bioinformatics, 20, 979–981. 21. Draghici,S., Khatri,P., Bhavsar,P., Shah,A., Krawetz,S.A. and REFERENCES Tainsky,M.A. (2003) Onto-Tools, the toolkit of the modern biologist: Onto-Express, Onto-Compare, Onto-Design and 1. Hosack,D.A., Dennis,G.Jr., Sherman,B.T., Lane,H.C. and Onto-Translate. Nucleic Acids Res., 31, 3775–3781. Lempicki,R.A. (2003) Identifying biological themes within lists 22. Khatri,P., Bhavsar,P., Bawa,G. and Draghici,S. (2004) of genes with EASE. Genome Biol., 4, R70. Onto-Tools: an ensemble of web-accessible, ontology-based 2. Dennis,G.Jr., Sherman,B.T., Hosack,D.A., Yang,J., Gao,W., tools for the functional design and interpretation of high- Lane,H.C. and Lempicki,R.A. (2003) DAVID: Database for throughput gene expression experiments. Nucleic Acids Res., 32, Annotation, Visualization, and Integrated Discovery. Genome Biol., W449–W456. 4, P3. 23. Khatri,P., Sellamuthu,S., Malhotra,P., Amin,K., Done,A. and 3. Maere,S., Heymans,K. and Kuiper,M. (2005) BiNGO: a Cytoscape Draghici,S. (2005) Recent additions and improvements to the plugin to assess overrepresentation of gene ontology categories in Onto-Tools. Nucleic Acids Res., 33, W762–W765. biological networks. Bioinformatics, 21, 3448–3449. 24. Benson,D.A., Karsch-Mizrachi,I., Lipman,D.J., Ostell,J. and 4. Berriz,G.F., King,O.D., Bryant,B., Sander,C. and Roth,F.P. (2003) Wheeler,D.L. (2006) GenBank. Nucleic Acids Res., 34, D16–D20. Characterizing gene sets with FuncAssociate. Bioinformatics, 19, 25. Apweiler,R., Bairoch,A., Wu,C.H., Barker,W.C., Boeckmann,B., 2502–2504. Ferro,S., Gasteiger,E., Huang,H., Lopez,R. et al. (2004) UniProt: 5. Bluthgen,N., Brand,K., Cajavec,B., Swat,M., Herzel,H. and the Universal Protein knowledgebase. Nucleic Acids Res., 32, Beule,D. (2005) Biological profiling of gene groups utilizing Gene D115–D119. Ontology. Genome Inform., 16, 106–115. 26. Wu,C.H., Apweiler,R., Bairoch,A., Natale,D.A., Barker,W.C., 6. Shah,N.H. and Fedoroff,N.V. (2004) CLENCH: a program for Boeckmann,B., Ferro,S., Gasteiger,E., Huang,H. et al. (2006) The calculating Cluster ENriCHment using the Gene Ontology. Universal Protein Resource (UniProt): an expanding universe of Bioinformatics, 20, 1196–1197. protein information. Nucleic Acids Res., 34, D187–D191. 7. Masseroli,M., Galati,O. and Pinciroli,F. (2005) GFINDer: genetic 27. Wu,C.H., Yeh,L.S., Huang,H., Arminski,L., Castro-Alvear,J., disease and phenotype location statistical analysis and mining of Chen,Y., Hu,Z., Kourtesis,P., Ledley,R.S. et al. (2003) The protein dynamically annotated gene lists. Nucleic Acids Res., 33, information resource. Nucleic Acids Res., 31, 345–347. W717–W723. 28. Maglott,D., Ostell,J., Pruitt,K.D. and Tatusova,T. (2005) Entrez 8. Liu,H., Hu,Z.Z. and Wu,C.H. (2005) DynGO: a tool for visualizing Gene: gene-centered information at NCBI. Nucleic Acids Res., 33, and mining of Gene Ontology and its associations. BMC D54–D58. Bioinformatics, 6, 201. 9. Al-Shahrour,F., Diaz-Uriarte,R. and Dopazo,J. (2004) FatiGO: a APPENDIX: URLs TO ACCESS MAJOR web tool for finding significant associations of Gene Ontology terms with groups of genes. Bioinformatics, 20, 578–580. COMPONENTS IN DAVID 10. Ashburner,M., Ball,C.A., Blake,J.A., Botstein,D., Butler,H., DAVID Home Page: http://david.niaid.nih.gov or Cherry,J.M., Davis,A.P., Dolinski,K., Dwight,S.S. et al. (2000) Gene ontology: tool for the unification of biology. The Gene http://david.abcc.ncifcrf.gov Ontology Consortium. Nat. Genet., 25, 25–29. DAVID Knowledgebase Download: http://david.abcc. 11. Lee,J.S., Katari,G. and Sachidanandam,R. (2005) GObar: a gene ncifcrf.gov/knowledgbase ontology based analysis and visualization tool for gene sets. DAVID Functional Annotation Tool Suite: http://david. BMC Bioinformatics, 6, 189. abcc.ncifcrf.gov/summary.jsp 12. Castillo-Davis,C.I. and Hartl,D.L. (2003) GeneMerge—post- genomic analysis, data mining, and hypothesis testing. DAVID Gene Functional Classification Tool Suite: Bioinformatics, 19, 891–892. http://david.abcc.ncifcrf.gov/gene2gene.jsp 13. Beissbarth,T. and Speed,T.P. (2004) GOstat: find statistically DAVID Gene ID Conversion Tool: http://david.abcc. overrepresented Gene Ontologies within a group of genes. ncifcrf.gov/conversion.jsp Bioinformatics, 20, 1464–1465. DAVID Gene Name Batch Viewer: http://david.abcc. 14. Zhong,S., Storch,K.F., Lipan,O., Kao,M.C., Weitz,C.J. and Wong,W.H. (2004) GoSurfer: A graphical interactive tool for ncifcrf.gov/list.jsp comparative analysis of large gene sets in Gene Ontologytrade DAVID NIAID Pathogen Browser Tool: http://david. mark Space. Appl. Bioinformatics, 3, 261–264. abcc.ncifcrf.gov/GB.jsp 15. Martin,D., Brun,C., Remy,E., Mouren,P., Thieffry,D. and Jacq,B. DAVID API Services: http://david.abcc.ncifcrf.gov/api (2004) GOToolBox: functional analysis of gene datasets based on DAVID Forum: http://david.abcc.ncifcrf.gov/forum Gene Ontology. Genome Biol., 5, R101. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists

Download PDF
7 pages

Loading next page...
 
/lp/oxford-university-press/david-bioinformatics-resources-expanded-annotation-database-and-novel-8nE04UDYqE

References (28)

Publisher
Oxford University Press
Copyright
© Published by Oxford University Press.
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/gkm415
pmid
17576678
Publisher site
See Article on Publisher Site

Abstract

Nucleic Acids Research, 2007, Vol. 35, Web Server issue W169–W175 doi:10.1093/nar/gkm415 DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists 1 1 1 1 2 Da Wei Huang , Brad T. Sherman , Qina Tan , Joseph Kir , David Liu , 2 5 2 3 David Bryant , Yongjian Guo , Robert Stephens , Michael W. Baseler , 4 1, H. Clifford Lane and Richard A. Lempicki * 1 2 3 Laboratory of Immunopathogenesis and Bioinformatics, Advanced Biomedical Computing Center, Clinical Services Program, SAIC-Frederick, Inc., National Cancer Institute at Frederick, MD 21702, USA, Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, 20892, USA, Bioinformatics and Scientific IT Program, NIAID Office of Technology Information Systems, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, 20892, USA Received January 22, 2007; Revised April 14, 2007; Accepted May 6, 2007 ABSTRACT bio-pathways and more. With DAVID (http://david. niaid.nih.gov), investigators gain more power to All tools in the DAVID Bioinformatics Resources interpret the biological mechanisms associated aim to provide functional interpretation of large lists with large gene lists. of genes derived from genomic studies. The newly updated DAVID Bioinformatics Resources consists of the DAVID Knowledgebase and five integrated, INTRODUCTION web-based functional annotation tool suites: In the post-genomic era, biological interpretation of the DAVID Gene Functional Classification Tool, the large gene lists derived from high-throughput experiments, DAVID Functional Annotation Tool, the DAVID Gene such as genes from microarray experiments, is a challen- ID Conversion Tool, the DAVID Gene Name Viewer ging task. The first version of DAVID (the Database for and the DAVID NIAID Pathogen Genome Browser. Annotation, Visualization and Integration Discovery), The expanded DAVID Knowledgebase now inte- released in 2003 (1,2), as well as a number of other similar grates almost all major and well-known public publicly available high-throughput functional annotation bioinformatics resources centralized by the DAVID tools (3–23), partially address the challenge by system- Gene Concept, a single-linkage method to agglom- atically mapping a large number of interesting genes in a list to associated Gene Ontology (GO) terms (10), and erate tens of millions of diverse gene/protein then statistically highlighting the most over-represented identifiers and annotation terms from a variety of (enriched) GO terms out of a list of hundreds or public bioinformatics databases. For any uploaded thousands of terms. This increases the likelihood that gene list, the DAVID Resources now provides not the investigator will identify the biological processes only the typical gene-term enrichment analysis, most pertinent to the biological phenomena under but also new tools and functions that allow users study (19). While this tool is extremely useful and has to condense large gene lists into gene functional been cited in hundreds of publications during the past groups, convert between gene/protein identifiers, three years, the development of other effective data mining visualize many-genes-to-many-terms relationships, algorithms, as additional components to the DAVID cluster redundant and heterogeneous terms into Bioinformatics Resources, will improve the power of groups, search for interesting and related genes or investigators to analyze their gene lists from different terms, dynamically view genes from their lists on biological angles. The newly added contents, functions *To whom correspondence should be addressed. Tel: +1-301-846-7114; Fax: 301-846-7672; Email: [email protected] The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. 2007 The Author(s) This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. W170 Nucleic Acids Research, 2007, Vol. 35, Web Server issue Table 1. Over 22 types of gene identifiers integrated by the DAVID and tool suites in the DAVID Bioinformatics Resources Gene Concept within the DAVID Knowledgebase intend to address several issues that other tools have not been able to extensively address: (i) to dramatically Gene ID Type Total ID Unique Cluster expand the biological information coverage in the DAVID Knowledgebase by comprehensively integrating AFFY_ID 2254679 845117 more than 20 types of major gene/protein identifiers ENTREZ_GENE_ID 1734858 1602339 GENPEPT_ACCESSION 4065385 2511637 and more than 40 well-known functional annotation GENBANK_ACCESSION 16828735 2409120 categories from dozens of public databases; (ii) to address GENEBANK_ID 20291282 2358084 the enriched and redundant relationships among many- PIR_ACCESSION 282281 258079 genes-to-many-terms (i.e. one gene could associate with PIR_ID 308092 266645 PIR_NREF_ID 3355759 2677404 many different, redundant terms and one term could REFSEQ_GENOMIC 1866800 1552597 associate with many genes) by developing a set of REFSEQ_MRNA 645831 561447 novel algorithms, such as the DAVID Gene Functional REFSEQ_PROTEIN 1644632 1373467 Classification Tool, the Functional Annotation Clustering REFSEQ_RNA 1364 852 UNIGENE 161138 158938 Tool, the Linear Searching Tool, the Fuzzy Gene-Term UNIPROT_ACCESSION 2864344 2097488 Heat Map Viewer, etc.; (iii) to dynamically visualize genes UNIPROT_ID 2789453 2096712 from a users list within the most relevant KEGG and UNIREF100_ID 2552342 2088692 BioCarta pathways with the DAVID Pathway Viewer; OFFICIAL_GENE_SYMBOL 1693151 1600906 FLYBASE_ID 27109 26642 (iv) to allow users to create and use customized gene HAMAP_ID 63925 63822 backgrounds for typical gene-term enrichment analysis HSSP_ID 265000 258750 utilizing the improved computational power and (v) to TIGR_ID 120117 111699 facilitate efficient communication and experience WORMBASE_ID 43675 21243 RGD_ID 25230 25060 exchange within the scientific community by moderating NOT SURE ALL IDs the DAVID Forum. This article summarizes the key DAVID components Any of the gene identifier types above can be cross-mapped to the and tool suites in the newly released DAVID DAVID Knowledgebase. ‘Not Sure’ is a new ID type specifically designed for the DAVID web site. For a given ‘not sure’ ID, all Bioinformatics Resources, highlighting new or expanded possible matching IDs will be systematically scanned across the entire analytic features that provide investigators with additional DAVID collection. means to explore and extract biological meaning from large gene lists that users input to the system (Supplementary File 1). For in-depth algorithm informa- tion, appropriate references and supplementary materials DAVID Knowledgebase are provided. A highly integrated gene-annotation database with comprehensive data coverage is essential for the success of any high-throughput annotation algorithms. Due to the FEATURES AND FUNCTIONALITIES complex and distributed nature of biological research, our current biological knowledge is distributed among many Computational Infrastructure redundant annotation databases maintained by indepen- The aim of the DAVID software design is to provide dent groups. One gene could have several different users with the simplest usability and fastest exploration identifiers within one or more database(s). Similarly, the speed through better internal software engineering biological terms associated with different gene identifiers practices. Therefore, the DAVID Bioinformatics Tools, for the same gene could be collected in different levels as web-based applications on a Tomcat web server across different databases. Due to these issues, most high- in a Linux machine (4-CPU for 3.5 GHz speed, 8 GB throughput annotation tools rely on one, or at most a few, memory), requires no configuration and installation in resource(s), which limits the analytic comprehensiveness the client’s computers. Java is the primary language and the level of throughput. The DAVID Knowledgebase used for all of the server side components of the is now built around the ‘DAVID Gene Concept’, a single calculation engines and the Java Server Page (JSP) web linkage method to agglomerate tens of millions of interfaces, in a full object-oriented fashion. In-memory gene/protein identifiers from a variety of public genomic Java data objects holding all genes-to-annotation resources (Table 1), including NCBI, PIR and UniProt information up to 2.5 GB in size were developed to (24–27), into broader secondary gene clusters, called greatly increase the data IO speed compared to that the DAVID Gene Concept (Figure 1, and more tech- through typical relational databases (e.g. Oracle). The nical details at http://david.abcc.ncifcrf.gov/helps/ Java Remote Method Invocation (RMI), a distributed knowledgebase/DAVID_gene.html), Grouping these computing technique, is also used to take advantage gene identifiers improves cross-referencing capability, of multiple computing resources. A set of automated allowing more than 40 categories of publicly available programs monitors many aspects of the web services in functional annotation to be comprehensively assigned to order to maximize the performance and minimize and centralized by the DAVID Gene Concept (Table 2, the down time period. see Supplementary File 2 for a complete list of annotation Nucleic Acids Research, 2007, Vol. 35, Web Server issue W171 NREF(NF00095014) DAVID Gene (2858470) PIR_ID : I38140 UniRef100_Q16825 GenPept : CAA56042 UniRef100_Q16825 Swissprot : PTN21_HUMAN RefSeq : NP_008970 Swissprot:PTN21_HUMAN Uniprot : Q16825 Uniprot : Q16825 Entrez Gene (11099) UniRef100_Q8WX29 Uniprot : Q16825 Genepept :: CAA56042 Swissprot : Q8WX29_HUMAN Refseq : NP_008970 Uniprot : Q8wx29 Uniprot : Q16825 NREF : NF00095014 Genepept :: CAD19000 PIR_ID : I38140 UniRef100_Q8WX29 NREF(NF00828766) Uniprot : Q16825 Swissport:Q8WX29_HUMAN Uniprot : Q8WX29 Genpept : CAA56042 Uniprot : Q8wx29 Genpept : CAD19000 RefSeq : NP_008970 NREF : NF00828766 Genpept : CAD19000 Entrez Gene : 11099 Figure 1. A DAVID gene constructed by a single linkage algorithm. Two UniRef100 clusters, two NRef 100 clusters and one Entrez Gene cluster were systematically found sharing one or more protein identifiers with each other. The single-linkage rule can further iteratively agglomerate them as a whole into one DAVID gene. Thus, for this particular example of tyrosine-protein phosphatase non-receptor type 21 (PTPN21), the resulting DAVID gene is able to collect and integrate all gene/protein identifiers more comprehensively than each original gene cluster. Table 2. The wide-range collection of heterogeneous functional annotations in the DAVID Knowledgebase Ontology (440 million records) Protein Domain/Family (415 millions) Sequence Features (421 millions) GO_BIOLOGICAL PROCESS BLOCKS_ID ALIAS_GENE_SYMBOL GO_MOLECULAR FUNCTION COG_KOG_NAME CHROMOSOME GO_CELLULAR COMPONENT INTERPRO_NAME CYTOBAND PANTHER_BIOLOGICAL PROCESS PDB_ID GENE_NAME PANTHER_MOLECULAR FUNCTION PFAM_NAME GENE_SYMBOL COG_KOG_ONTOLOGY PIR_ALN HOMOLOGOUS_GENE P-P Interaction (44 millions) PIR_HOMOLOGY_DOMAIN ENTREZ_GENE_SUMMARY BIND PIR_SUPERFAMILY_NAME OMIM_ID DIP PRINTS_NAME PIR_SUMMARY MINT PRODOM_NAME PROTEIN_MW NCICB_CAPATHWAY PROSITE_NAME REFSEQ_PRODUCT TRANSFAC_ID SCOP_ID SEQUENCE_LENGTH HIV_INTERACTION SMART_NAME SP_COMMENT HIV_INTERACTION_CATEGORY TIGRFAMS_NAME Functional Category (46.9 millions) HPRD_INTERACTION PANTHER_SUBFAMILY PIR_SEQ_FEATURE REACTOME_INTERACTION PANTHER_FAMILY SP_COMMENT_TYPE Disease Association (9,000) Pathways (450 000) SP_PIR_KEYWORDS GENETIC_ASSOCIATION_DB BioCarta UP_SEQ_FEATURE OMIM_DISEASE KEGG_PATHWAY Gene Tissue Expression (41.0 million) Literature (42.8 millions) PANTHER_PATHWAY GNF Microarray GENERIF_SUMMARY PID UNIGENE EST PUBMED_ID BBID CGAP SAGE HIV_INTERACTION_PUBMED_ID KEGG_REACTION CGAP EST Over 60 functional categories from dozens of independent public sources (databases) (see Supplementary File 2 for a complete list) are collected and integrated in the DAVID Knowledgebase. sources and more technical details at http://david.abcc. enhances the comprehensiveness of high-throughput gene ncifcrf.gov/helps/knowledgebase/DAVID_gene.html). To functional analysis by overlapping multiple biological the best of our knowledge, this annotation coverage far aspects together. It also provides a solid foundation for exceeds that of the original DAVID database and those the further development of more advanced high through- currently used by other similar high-throughput annota- put analytic algorithms that may be added to the DAVID tion tools. The DAVID knowledgebase not only increases Bioinformatics Resources. More importantly, the entire the accessibility to a wide range of heterogeneous DAVID Knowledgebase, in simple pair-wise text format annotation data in one centralized location, but also files containing a broad, highly integrated annotation W172 Nucleic Acids Research, 2007, Vol. 35, Web Server issue data collection, is freely available to the public (http:// DAVID Gene Functional Classification Tool Suite david.abcc.ncifcrf.gov/knowledgebase), which will benefit The DAVID Gene Functional Classification Tool (http:// various high-throughput data mining projects by other david.abcc.ncifcrf.gov/gene2gene.jsp) is a completely new research groups. The DAVID Knowledgebase is expected component in the DAVID Bioinformatics Resources. to be updated more frequently in the near future than The tool provides a novel way to functionally analyze a its current annual update. large number of genes in a high-throughput fashion by classifying them into gene groups based on their annota- tion term co-occurrence. This is accomplished and DAVID Functional Annotation Tool Suite visualized by a set of new fuzzy classification algorithms, including a kappa statistics measurement of gene–gene This tool suite (http://david.abcc.ncifcrf.gov/summary. functional relationship, a fuzzy multi-linkage partitioning jsp), introduced in the first version of DAVID, mainly method and a fuzzy genes-terms heat map visualization, provides typical batch annotation and gene-GO term etc. (manuscript submitted, and more details at http:// enrichment analysis to highlight the most relevant GO david.abcc.ncifcrf.gov/manuscripts/fuzzy_cluster/). The terms associated with a given gene list (2). The new version power of the tool is that it allows users to simultaneously of the tool keeps the same enrichment analytic algorithm view the rich and redundant internal relationship of but with extended annotation content coverage, increasing functionally related genes and their annotation terms from only GO in the original version of DAVID to within biological modules. Investigators are able to currently over 40 annotation categories, including GO functionally analyze their gene list in a highly related terms, protein–protein interactions, protein functional many-genes-to-many-terms network context instead of domains, disease associations, bio-pathways, sequence a one-term-to-many-genes or a one-gene-to-many-terms general features, homologies, gene functional summaries, view in the typical gene-annotation enrichment analysis. gene tissue expressions, literatures, etc. (Table 2). The improved annotation coverage alone provides investiga- DAVID Gene ID Conversion Tool Suite tors with much more power to analyze their genes using A significant number of different types of gene/protein many different biological aspects in a single space. Flexible identifiers, not mutually mapped to each other across options are provided to display results in an individual three independent resources, NCBI, PIR and UniProt annotation chart report or a combined chart report. (25,26,28), are now maximally integrated in the DAVID In addition to pre-built gene population backgrounds Knowledgebase (Figure 1, more details at http://david. (e.g. Affy U133) used in gene-annotation enrichment abcc.ncifcrf.gov/helps/knowledgebase/ analysis, with its improved computational power, the new DAVID_gene.html), whose scope is more expansive than tool accepts user-defined population gene list, an option one system only. Even though the DAVID rarely found in other similar web-based, high-throughput Knowledgebase is used primarily for improvement of annotation tools. This feature was added in order to more annotation terms integration and coverage, such compre- specifically meet the users’ requirements for the best hensive gene identifier coverage and cross-referencing analytical results. capability could itself be very useful for researchers to The DAVID Functional Annotation Clustering is convert their gene/protein identifiers from one type to a newly added feature (manuscript submitted, and another among over 20 major types of identifier systems more details at http://david.abcc.ncifcrf.gov/manuscripts/ (Table 1). Thus, with the newly introduced DAVID fuzzy_cluster/) to the DAVID Functional Annotation Gene ID Conversion Tool (http://david.abcc.ncifcrf.gov/ Tool. This function uses a novel algorithm to measure conversion.jsp), interesting genes derived from one identi- relationships among the annotation terms based on the fier system can be quickly translated to other gene degrees of their co-association genes to group the similar, identifier types preferred by a given annotation resource. redundant and heterogeneous annotation contents from In addition, the DAVID Gene ID Conversion Tool the same or different resources into annotation groups. provides a ‘not sure’ type for ambiguous gene identifiers, This reduces the burden of associating similar redundant whereby the tool can systematically suggest the potential terms and makes the biological interpretation more type(s). For instance, a user has a gene ID ‘3558’ without focused in a group level (Figure 2). The tool also provides ID type information. DAVID Gene ID Conversion Tool a look at the internal relationships among the clustered will scan all possibilities across all gene ID systems terms. The clustered format is able to give a more collected in the DAVID Knowledgebase. Two choices insightful view about the relationships of annotations will be suggested, i.e. ‘3558’ could be an Entrez Gene ID compared to the traditional un-clustered term report, over for IL2 (human) or a Genbank ID for CNA1 (yeast). which similar annotation terms may be spread among Thus, the user can make a decision based on above hundreds, if not thousands, of other terms. In addition, information. to take full advantage of the well-known KEGG and BioCarta pathways, the new DAVID Pathway Viewer, DAVID Gene Name Batch Viewer another feature of the DAVID Functional Annotation Tool, can display genes from a user’s list on pathway After obtaining a list of interesting genes, probably the maps to facilitate biological interpretation in a network first question researchers will ask is ‘What are the names context. of my genes?’ Even though it is a simple question, Nucleic Acids Research, 2007, Vol. 35, Web Server issue W173 Figure 2. An HTML report from the Functional Annotation Clustering. The annotation cluster 1 in the example shows that GO term cytokine activity, KEGG pathway cytokine–cytokine receptor interaction, and GO term receptor binding, etc. are grouped together. Thus, the different biological aspects regarding a relevant biology can be explored at the same time. most high-throughput annotation tools do not answer it in co-occurrence of annotations between genes (more details a straightforward way. The new DAVID Gene Name at http://david.abcc.ncifcrf.gov/helps/linear_search.html). Batch Viewer is designed to simply list the gene names for all given genes. In addition, hyperlinks are provided on DAVID NIAID Pathogen Browser each gene entry, allowing users to explore in depth other functional information around the gene. Thus, this tool The National Institute of Allergy and Infectious Diseases provides users with a first glance and initial ideas about (NIAID) has defined category A, B and C priority their interesting genes before proceeding to analysis by pathogens (http://www3.niaid.nih.gov/Biodefense/bandc_ other more comprehensive analytic tool. Moreover, priority.htm), which have subsequently become important hyperlinks, labeled as ‘RT’, are provided for each gene in biodefense research funding, attracting broad interest in order to search other functionally related genes in user’s from the research community. Since the organisms listed gene list or the entire genome. The search is based on in these categories may not be familiar to researchers who W174 Nucleic Acids Research, 2007, Vol. 35, Web Server issue ++ Highly Applied + Relevant Initial glance of major biological functions associated with my gene list ++ ++ ++ + + Which biological terms/functions are specifically enriched in my gene list? ++ ++ View the genes in my list on related biological pathways ++ ++ Which diseases are associated with my gene list? ++ ++ Which protein functional domains are associated with my gene list ? ++ ++ Which other genes frequently interact with the genes in my list? ++ ++ How to group the highly redundant annotations into group? ++ What are the major gene functional groups in my gene list? ++ ++ View related annotation and related genes on a single graphic view ++ What are other functionally similar genes in genome, but not in my list? + + ++ ++ What are other annotations functionally similar to my interesting one? ++ ++ What are the gene names in my list? +++ How to convert my gene IDs to other type of IDs? +++ How to directly link to DAVID functions ? ++ How can I download DAVID data for in-house study? ++ + + ++ Figure 3. A roadmap to choose appropriate DAVID functions and tools. have recently joined the emerging field, the DAVID CONCLUSION NIAID Pathogen Browser (http://david.abcc.ncifcrf.gov/ The newly released DAVID Bioinformatics Resources are GB.jsp) is provided as a quick starting point for them to an expanded version of the original DAVID. It provides a search the most relevant genes in the organisms by set of powerful, novel tools that researchers can use to biological key words of interests. A large list of genes explore their large gene lists in depth from many different retrieved from the search could be further transferred to biological angles (Figure 3) in order to extract associated the DAVID Bioinformatics Resources for in-depth func- biological meanings to the greatest extent possible. The tional analysis with any of the previously mentioned tools. advanced data collection in the DAVID Knowledgebase Although the tool is still in its early stage, it may help not only creates a solid annotation data foundation for researchers gain understanding of the genes related to a the various DAVID analytic tools, but also is freely priority pathogen of interest. More development is available to the public in a simple pair-wise text format ongoing to extend the searching scope to all available to promote the development of novel annotation algo- genomes and annotations collected in DAVID rithms and techniques within the scientific community. knowledgebase. The DAVID Bioinformatics Resources are accessible at http://david.niaid.nih.gov. DAVID API Services SUPPLEMENTARY DATA DAVID API services (http://david.abcc.ncifcrf.gov/api/) are newly added features that allow users to directly Supplementary Data are available at NAR Online. pass gene list to various DAVID tools via a set of pre- defined URLs instead of DAVID submission forms. ACKNOWLEDGEMENTS Thus, DAVID tools can easily serve as part of the analytic pipeline in other bioinformatics web sites. They The authors are grateful to the referees and editors for can also be used in bioinformatics scripts to automate their constructive comments. Thanks goes to Melaku functional annotation for large number of gene lists, Gedil, Ping Ren, and Jun Yang in the LIB group for which are too many to be accomplished by the manual biological discussions. We also thank Bill Wilton and procedures. Mike Tartakovsky for information technology and Functional Annotation Chart Functional Annotation Clustering Functional Annotation Table Gene Functional Classification Gene Name Batch Viewer Gene ID Conversion Tool DAVID Knowledge base DAVID API Nucleic Acids Research, 2007, Vol. 35, Web Server issue W175 16. Zhang,B., Schmoyer,D., Kirov,S. and Snoddy,J. (2004) GOTree network support. This research was supported in whole by Machine (GOTM): a web-based platform for interpreting sets of the National Institute of Allergy and Infectious Disease. interesting genes using Gene Ontology hierarchies. BMC This project has been funded in whole with federal funds Bioinformatics, 5, 16. from the National Cancer Institute, National Institutes of 17. Zeeberg,B.R., Qin,H., Narasimhan,S., Sunshine,M., Cao,H., Kane,D.W., Reimers,M., Stephens,R.M., Bryant,D. et al. (2005) Health, under contract N01- CO-12400. The content of High-Throughput GoMiner, an ‘industrial-strength’ integrative gene this publication does not necessarily reflect the views or ontology tool for interpretation of multiple-microarray experiments, policies of the Department of Health and Human Services, with application to studies of Common Variable Immune Deficiency nor does mention of trade names, commercial products, (CVID). BMC Bioinformatics, 6, 168. 18. Ben-Shaul,Y., Bergman,H. and Soreq,H. (2005) Identifying subtle or organizations imply endorsement by the U.S. interrelated changes in functional gene categories using continuous Government. Funding to pay the Open Access publication measures of gene expression. Bioinformatics, 21, 1129–1137. charges for this article was provided by the same source as 19. Khatri,P. and Draghici,S. (2005) Ontological analysis of gene above. expression data: current tools, limitations, and open problems. Bioinformatics, 21, 3587–3595. 20. Robinson,P.N., Wollstein,A., Bohme,U. and Beattie,B. (2004) Conflict of interest statement. None declared. Ontologizing gene-expression microarray data: characterizing clusters with Gene Ontology. Bioinformatics, 20, 979–981. 21. Draghici,S., Khatri,P., Bhavsar,P., Shah,A., Krawetz,S.A. and REFERENCES Tainsky,M.A. (2003) Onto-Tools, the toolkit of the modern biologist: Onto-Express, Onto-Compare, Onto-Design and 1. Hosack,D.A., Dennis,G.Jr., Sherman,B.T., Lane,H.C. and Onto-Translate. Nucleic Acids Res., 31, 3775–3781. Lempicki,R.A. (2003) Identifying biological themes within lists 22. Khatri,P., Bhavsar,P., Bawa,G. and Draghici,S. (2004) of genes with EASE. Genome Biol., 4, R70. Onto-Tools: an ensemble of web-accessible, ontology-based 2. Dennis,G.Jr., Sherman,B.T., Hosack,D.A., Yang,J., Gao,W., tools for the functional design and interpretation of high- Lane,H.C. and Lempicki,R.A. (2003) DAVID: Database for throughput gene expression experiments. Nucleic Acids Res., 32, Annotation, Visualization, and Integrated Discovery. Genome Biol., W449–W456. 4, P3. 23. Khatri,P., Sellamuthu,S., Malhotra,P., Amin,K., Done,A. and 3. Maere,S., Heymans,K. and Kuiper,M. (2005) BiNGO: a Cytoscape Draghici,S. (2005) Recent additions and improvements to the plugin to assess overrepresentation of gene ontology categories in Onto-Tools. Nucleic Acids Res., 33, W762–W765. biological networks. Bioinformatics, 21, 3448–3449. 24. Benson,D.A., Karsch-Mizrachi,I., Lipman,D.J., Ostell,J. and 4. Berriz,G.F., King,O.D., Bryant,B., Sander,C. and Roth,F.P. (2003) Wheeler,D.L. (2006) GenBank. Nucleic Acids Res., 34, D16–D20. Characterizing gene sets with FuncAssociate. Bioinformatics, 19, 25. Apweiler,R., Bairoch,A., Wu,C.H., Barker,W.C., Boeckmann,B., 2502–2504. Ferro,S., Gasteiger,E., Huang,H., Lopez,R. et al. (2004) UniProt: 5. Bluthgen,N., Brand,K., Cajavec,B., Swat,M., Herzel,H. and the Universal Protein knowledgebase. Nucleic Acids Res., 32, Beule,D. (2005) Biological profiling of gene groups utilizing Gene D115–D119. Ontology. Genome Inform., 16, 106–115. 26. Wu,C.H., Apweiler,R., Bairoch,A., Natale,D.A., Barker,W.C., 6. Shah,N.H. and Fedoroff,N.V. (2004) CLENCH: a program for Boeckmann,B., Ferro,S., Gasteiger,E., Huang,H. et al. (2006) The calculating Cluster ENriCHment using the Gene Ontology. Universal Protein Resource (UniProt): an expanding universe of Bioinformatics, 20, 1196–1197. protein information. Nucleic Acids Res., 34, D187–D191. 7. Masseroli,M., Galati,O. and Pinciroli,F. (2005) GFINDer: genetic 27. Wu,C.H., Yeh,L.S., Huang,H., Arminski,L., Castro-Alvear,J., disease and phenotype location statistical analysis and mining of Chen,Y., Hu,Z., Kourtesis,P., Ledley,R.S. et al. (2003) The protein dynamically annotated gene lists. Nucleic Acids Res., 33, information resource. Nucleic Acids Res., 31, 345–347. W717–W723. 28. Maglott,D., Ostell,J., Pruitt,K.D. and Tatusova,T. (2005) Entrez 8. Liu,H., Hu,Z.Z. and Wu,C.H. (2005) DynGO: a tool for visualizing Gene: gene-centered information at NCBI. Nucleic Acids Res., 33, and mining of Gene Ontology and its associations. BMC D54–D58. Bioinformatics, 6, 201. 9. Al-Shahrour,F., Diaz-Uriarte,R. and Dopazo,J. (2004) FatiGO: a APPENDIX: URLs TO ACCESS MAJOR web tool for finding significant associations of Gene Ontology terms with groups of genes. Bioinformatics, 20, 578–580. COMPONENTS IN DAVID 10. Ashburner,M., Ball,C.A., Blake,J.A., Botstein,D., Butler,H., DAVID Home Page: http://david.niaid.nih.gov or Cherry,J.M., Davis,A.P., Dolinski,K., Dwight,S.S. et al. (2000) Gene ontology: tool for the unification of biology. The Gene http://david.abcc.ncifcrf.gov Ontology Consortium. Nat. Genet., 25, 25–29. DAVID Knowledgebase Download: http://david.abcc. 11. Lee,J.S., Katari,G. and Sachidanandam,R. (2005) GObar: a gene ncifcrf.gov/knowledgbase ontology based analysis and visualization tool for gene sets. DAVID Functional Annotation Tool Suite: http://david. BMC Bioinformatics, 6, 189. abcc.ncifcrf.gov/summary.jsp 12. Castillo-Davis,C.I. and Hartl,D.L. (2003) GeneMerge—post- genomic analysis, data mining, and hypothesis testing. DAVID Gene Functional Classification Tool Suite: Bioinformatics, 19, 891–892. http://david.abcc.ncifcrf.gov/gene2gene.jsp 13. Beissbarth,T. and Speed,T.P. (2004) GOstat: find statistically DAVID Gene ID Conversion Tool: http://david.abcc. overrepresented Gene Ontologies within a group of genes. ncifcrf.gov/conversion.jsp Bioinformatics, 20, 1464–1465. DAVID Gene Name Batch Viewer: http://david.abcc. 14. Zhong,S., Storch,K.F., Lipan,O., Kao,M.C., Weitz,C.J. and Wong,W.H. (2004) GoSurfer: A graphical interactive tool for ncifcrf.gov/list.jsp comparative analysis of large gene sets in Gene Ontologytrade DAVID NIAID Pathogen Browser Tool: http://david. mark Space. Appl. Bioinformatics, 3, 261–264. abcc.ncifcrf.gov/GB.jsp 15. Martin,D., Brun,C., Remy,E., Mouren,P., Thieffry,D. and Jacq,B. DAVID API Services: http://david.abcc.ncifcrf.gov/api (2004) GOToolBox: functional analysis of gene datasets based on DAVID Forum: http://david.abcc.ncifcrf.gov/forum Gene Ontology. Genome Biol., 5, R101.

Journal

Nucleic Acids Research – Oxford University Press

Published: Jul 1, 2007

There are no references for this article.