Search

For nearly 25 years, FlyBase (flybase.org) has provided a freely available online database of biological information about Drosophila species, focusing on the model organism D. melanogaster. The need for a centralized, integrated view of Drosophila research has never been greater as advances in genomic, proteomic, and high-throughput technologies add to the quantity and diversity of available data and resources.FlyBase has taken several approaches to respond to these changes in the research landscape. Novel report pages have been generated for new reagent types and physical interaction data; Drosophila models of human disease are now represented and showcased in dedicated Human Disease Model Reports; other integrated reports have been established that bring together related genes, datasets, or reagents; Gene Reports have been revised to improve access to new data types and to highlight functional data; links to external sites have been organized and expanded; and new tools have been developed to display and interrogate all these data, including improved batch processing and bulk file availability. In addition, several new community initiatives have served to enhance interactions between researchers and FlyBase, resulting in direct user contributions and improved feedback.This chapter provides an overview of the data content, organization, and available tools within FlyBase, focusing on recent improvements. We hope it serves as a guide for our diverse user base, enabling efficient and effective exploration of the database and thereby accelerating research discoveries.

FlyBase (http://flybase.org) is the primary resource for molecular and genetic information on the Drosophilidae. The database serves researchers of diverse backgrounds and interests, and offers several different query tools to provide efficient access to the data available and facilitate the discovery of significant relationships within the database. Recently, FlyBase has developed Interactions Browser and enhanced GBrowse, which are graphical query tools, and made improvements to the search tools QuickSearch and QueryBuilder. Furthermore, these search tools have been integrated with Batch Download and new analysis tools through a more flexible search results list, providing powerful ways of exploring the data in FlyBase.

The Biological General Repository for Interaction Datasets (BioGRID) database (http://www.thebiogrid.org) was developed to house and distribute collections of protein and genetic interactions from major model organism species. BioGRID currently contains over 198 000 interactions from six different species, as derived from both high-throughput studies and conventional focused studies. Through comprehensive curation efforts, BioGRID now includes a virtually complete set of interactions reported to date in the primary literature for both the budding yeast Saccharomyces cerevisiae and the fission yeast Schizosaccharomyces pombe. A number of new features have been added to the BioGRID including an improved user interface to display interactions based on different attributes, a mirror site and a dedicated interaction management system to coordinate curation across different locations. The BioGRID provides interaction data with monthly updates to Saccharomyces Genome Database, Flybase and Entrez Gene. Source code for the BioGRID and the linked Osprey network visualization system is now freely available without restriction.

Motivation: A few years ago, FlyBase undertook to design a new database schema to store Drosophila data. It would fully integrate genomic sequence and annotation data with bibliographic, genetic, phenotypic and molecular data from the literature representing a distillation of the first 100 years of research on this major animal model system. In developing this new integrated schema, FlyBase also made a commitment to ensure that its design was generic, extensible and available as open source, so that it could be employed as the core schema of any model organism data repository, thereby avoiding redundant software development and potentially increasing interoperability. Our question was whether we could create a relational database schema that would be successfully reused. Results: Chado is a relational database schema now being used to manage biological knowledge for a wide variety of organisms, from human to pathogens, especially the classes of information that directly or indirectly can be associated with genome sequences or the primary RNA and protein products encoded by a genome. Biological databases that conform to this schema can interoperate with one another, and with application software from the Generic Model Organism Database (GMOD) toolkit. Chado is distinctive because its design is driven by ontologies. The use of ontologies (or controlled vocabularies) is ubiquitous across the schema, as they are used as a means of typing entities. The Chado schema is partitioned into integrated subschemas (modules), each encapsulating a different biological domain, and each described using representations in appropriate ontologies. To illustrate this methodology, we describe here the Chado modules used for describing genomic sequences. Availability: GMOD is a collaboration of several model organism database groups, including FlyBase, to develop a set of open-source software for managing model organism data. The Chado schema is freely distributed under the terms of the Artistic License ( http://www.opensource.org/licenses/artistic-license.php) from GMOD (www.gmod.org). Contact:cjm@fruitfly.org or emmert@morgan.harvard.edu.

Background A critical topic of insertional mutagenesis experiments performed on model organisms is mapping the hits of artificial transposons (ATs) at nucleotide level accuracy. Mapping errors may occur when sequencing artifacts or mutations as single nucleotide polymorphisms (SNPs) and small indels are present very close to the junction between a genomic sequence and a transposon inverted repeat (TIR). Another particular item of insertional mutagenesis is mapping of the transposon self-insertions and, to our best knowledge, there is no publicly available mapping tool designed to analyze such molecular events. Results We developed Genome ARTIST, a pairwise gapped aligner tool which works out both issues by means of an original, robust mapping strategy. Genome ARTIST is not designed to use next-generation sequencing (NGS) data but to analyze ATs insertions obtained in small to medium-scale mutagenesis experiments. Genome ARTIST employs a heuristic approach to find DNA sequence similarities and harnesses a multi-step implementation of a Smith-Waterman adapted algorithm to compute the mapping alignments. The experience is enhanced by easily customizable parameters and a user-friendly interface that describes the genomic landscape surrounding the insertion. Genome ARTIST is functional with many genomes of bacteria and eukaryotes available in Ensembl and GenBank repositories. Our tool specifically harnesses the sequence annotation data provided by FlyBase for Drosophila melanogaster (the fruit fly), which enables mapping of insertions relative to various genomic features such as natural transposons. Genome ARTIST was tested against other alignment tools using relevant query sequences derived from the D. melanogaster and Mus musculus (mouse) genomes. Real and simulated query sequences were also comparatively inquired, revealing that Genome ARTIST is a very robust solution for mapping transposon insertions. Conclusions Genome ARTIST is a stand-alone user-friendly application, designed for high-accuracy mapping of transposon insertions and self-insertions. The tool is also useful for routine aligning assessments like detection of SNPs or checking the specificity of primers and probes. Genome ARTIST is an open source software and is available for download at www.genomeartist.ro and at GitHub (https://github.com/genomeartist/genomeartist ). Electronic supplementary material The online version of this article (doi:10.1186/s13100-016-0061-0) contains supplementary material, which is available to authorized users.

FlyBase (http://flybase.org/) is a database of genetic and genomic data on the model organism Drosophila melanogaster and the entire insect family Drosophilidae. The FlyBase Consortium curates, annotates, integrates and maintains a wide variety of data within this domain. Access to the data is provided through graphical and textual user interfaces tailored to particular types of data. FlyBase data types include maps at the cytological, genetic and sequence levels, genes and alleles including their products, functions, expression patterns, mutant phenotypes and genetic interactions as well as aberrant chromosomes, annotated genomes, genetic stock collections, transposons, transgene constructs and insertions, anatomy and images, bibliographic data, and community contact information.

FlyBase (http://flybase.org) is the primary repository of genetic and molecular data of the insect family Drosophilidae. For the most extensively studied species, Drosophila melanogaster, a wide range of data are presented in integrated formats. Data types include mutant phenotypes, molecular characterization of mutant alleles and aberrations, cytological maps, wild-type expression patterns, anatomical images, transgenic constructs and insertions, sequence-level gene models and molecular classification of gene product functions. There is a growing body of data for other Drosophila species; this is expected to increase dramatically over the next year, with the completion of draft-quality genomic sequences of an additional 11 Drosophila species.

Transcription factors (TFs) are essential regulators of gene expression, and mutated TF genes have been shown to cause numerous human genetic diseases. Yet to date, no single, comprehensive database of human TFs exists. In this work, we describe the collection of an essentially complete set of TF genes from one depiction of the human ORFeome, and the design of a microarray to interrogate their expression. Taking 1468 known TFs from TRANSFAC, InterPro, and FlyBase, we used this seed set to search the ScriptSure human transcriptome database for additional genes. ScriptSure's genome-anchored transcript clusters allowed us to work with a nonredundant high-quality representation of the human transcriptome. We used a high-stringency similarity search by using BLASTN, and a protein motif search of the human ORFeome by using hidden Markov models of DNA-binding domains known to occur exclusively or primarily in TFs. Four hundred ninety-four additional TF genes were identified in the overlap between the two searches, bringing our estimate of the total number of human TFs to 1962. Zinc finger genes are by far the most abundant family (762 members), followed by homeobox (199 members) and basic helix-loop-helix genes (117 members). We designed a microarray of 50-mer oligonucleotide probes targeted to a unique region of the coding sequence of each gene. We have successfully used this microarray to interrogate TF gene expression in species as diverse as chickens and mice, as well as in humans.

The functional classification of genes on a genome-wide scale is now in its infancy, and we make a first attempt to assess existing methods and identify sources of error. To this end, we compared two independent efforts for associating proteins with functions, one implemented by FlyBase and the other by PANTHER at Celera Genomics. Both methods make inferences based on sequence similarity and the available experimental evidence. However, they differ considerably in methodology and process. Overall, assuming that the systematic error across the two methods is relatively small, we find the protein-to-function association error rate of both the FlyBase and PANTHER methods to be

The Gene Ontology Consortium (Gene Ontology Consortium 2000) provides a standardized vocabulary for the annotation of gene attributes, which fall into the three general categories of molecular function, biological process, and cellular component. Organism-specific databases such as FlyBase (FlyBase Consortium 2002), Saccharomyces Genome Database (SGD; Cherry et al. 1998), Mouse Genome Database (MGD; Blake et al. 2002), and WormBase (Stein et al. 2001), have codeveloped this vocabulary, and have used it to annotate genes with the attributes that the biomedical literature asserts that they hold. These databases are incomplete because there are genes whose attributes are not yet all known, and because there is literature that has not yet been digested by the database curators. In such cases it is useful to have a prediction of whether a gene has a certain attribute. Such predictions can help to make the databases more complete (and consequently more useful to researchers) by directing curators toward literature that they may have overlooked. Also, predictions that are not presently supported by the literature provide new hypotheses that may be tested experimentally. A variety of approaches for predicting Gene Ontology (GO) attributes have been attempted. Natural language processing was used in Raychaudhuri et al. (2002) to automate the curator's task of extracting gene–attribute associations from literature abstracts. Others have assigned attributes to genes on the basis of microarray data (Hvidsten et al. 2001) or protein folds (Schug et al. 2002). These approaches are especially valuable for assigning attributes to genes with otherwise unknown function. But once some attributes of a gene are known, statistical patterns among the annotations themselves can be useful for predicting additional attributes. In this paper, we model the probabilistic relationships between the GO annotations using two approaches, one based on decision trees and the other based on Bayesian networks. We assess the models using cross-validation on the SGD and FlyBase databases. We also manually assess 100 of those gene–attribute associations that the models indicate are likely to hold but that have not been annotated in the databases.

Release 6, the latest reference genome assembly of the fruit fly Drosophila melanogaster, was released by the Berkeley Drosophila Genome Project in 2014; it replaces their previous Release 5 genome assembly, which had been the reference genome assembly for over 7 years. With the enormous amount of information now attached to the D. melanogaster genome in public repositories and individual laboratories, the replacement of the previous assembly by the new one is a major event requiring careful migration of annotations and genome-anchored data to the new, improved assembly. In this report, we describe the attributes of the new Release 6 reference genome assembly, the migration of FlyBase genome annotations to this new assembly, how genome features on this new assembly can be viewed in FlyBase (http://flybase.org) and how users can convert coordinates for their own data to the corresponding Release 6 coordinates.

With the advancement of new high throughput sequencing technologies, there has been an increase in the number of genome sequencing projects worldwide, which has yielded complete genome sequences of human, animals and plants. Subsequently, several labs have focused on genome annotation, consisting of assigning functions to gene products, mostly using Gene Ontology (GO) terms. As a consequence, there is an increased heterogeneity in annotations across genomes due to different approaches used by different pipelines to infer these annotations and also due to the nature of the GO structure itself. This makes a curator's task difficult, even if they adhere to the established guidelines for assessing these protein annotations. Here we develop a genome-scale approach for integrating GO annotations from different pipelines using semantic similarity measures. We used this approach to identify inconsistencies and similarities in functional annotations between orthologs of human and Drosophila melanogaster, to assess the quality of GO annotations derived from InterPro2GO mappings compared to manually annotated GO annotations for the Drosophila melanogaster proteome from a FlyBase dataset and human, and to filter GO annotation data for these proteomes. Results obtained indicate that an efficient integration of GO annotations eliminates redundancy up to 27.08 and 22.32% in the Drosophila melanogaster and human GO annotation datasets, respectively. Furthermore, we identified lack of and missing annotations for some orthologs, and annotation mismatches between InterPro2GO and manual pipelines in these two proteomes, thus requiring further curation. This simplifies and facilitates tasks of curators in assessing protein annotations, reduces redundancy and eliminates inconsistencies in large annotation datasets for ease of comparative functional genomics.

Phenotypic analysis combined with molecular genetics is a powerful tool for mapping gene function onto the genome. Phenotypic data are, by their nature, descriptive, and as varied as the range of mutant phenotypes that can be presented by the organism under study. This paper discusses the mechanisms FlyBase has implemented to systematise published phenotypic data about Drosophila, and provides an introduction to the query tools available for the mining of the data. Though FlyBase is specific to Drosophila, the issues faced in devising protocols for capturing, storing and reporting data are the same issues faced by any database with an interest in using phenotypic data to maximise the potential of genomic analysis.

The distinctive black phenotype of ebony mutants has made it one of the most widely used phenotypic markers in Drosophila genetics. Without doubt, ebony showcases the fruits of the fly community's labours to annotate gene function. As of this writing, FlyBase lists 142 references, 1277 fly stocks, 15 phenotypes and 44 alleles. In addition to its namesake pigmentation phenotype, ebony mutants affect other traits, including phototaxis and courtship. With phenotypic consequences of ebony variants readily apparent in the laboratory, does natural selection also see them in the wild? In this issue of Molecular Ecology, Pool & Aquadro investigate this question and found signs of natural selection on the ebony gene that appear to have resulted from selection for darker pigmentation at higher elevations in sub-Saharan populations of Drosophila melanogaster. Such findings from population genomic analysis of wild-derived strains should be included in gene annotations to provide a more holistic view of a gene's function. The evolutionary annotation of ebony added by Pool & Aquadro substantiates that pigmentation can be adaptive and implicates elevation as an important selective factor. This is important progress because the selective factors seem to differ between populations and species. In addition, the study raises issues to consider when extrapolating from selection at the molecular level to selection at the phenotypic level.

International audience; GIF‐DB (Gene Interactions in the Fly Database) is a new WWW database (http://www‐biol.univ‐mrs.fr/ approximately lgpd/GIFTS\ₕome\ₚage. html ) describing gene molecular interactions involved in the process of embryonic pattern formation in the flyDrosophila melanogaster. The detailed information is distributed in specific lines arranged into an EMBL‐ (or SWISS‐PROT‐) like format. GIF‐DB achieves a high level of integration with other databases such as FlyBase, EMBL and SWISS‐PROT through numerous hyperlinks. The original concept of interaction databases examplified by GIF‐DB could be extended to other biological subjects and organisms so as to study gene regulatory networks in an evolutionary perspective.

FlyBase is a database of genetic and molecular data concerning Drosophila. FlyBase is maintained as a relational database (in Sybase) and is made available as html documents and flat files. The scope of FlyBase includes: genes, alleles (and phenotypes), aberrations, transposons, pointers to sequence data, clones, stock lists, Drosophila workers and bibliographic references. The Encyclopedia of Drosophila is a joint effort between FlyBase and the Berkeley Drosophila Genome Project which integrates FlyBase data with those from the BDGP.

Flytrap is a web-enabled relational database of transposable element insertions in Drosophila melanogaster. A green fluorescent protein (GFP) artificial exon carried by a transposable P-element is mobilized and inserted into a host gene intron creating a GFP fusion protein. The sequence of the tagged gene is determined by sequencing inverse-PCR products derived from genomic DNA. Flytrap contains two principle data types: micrographs of protein localization and a cellular component ontology, based on rules derived from the Gene Ontology consortium (http://www.geneontology.org), describing protein localization. Flytrap also has links to gene information contained in Flybase (http:// flybase.bio.indiana.edu). The system is designed to accept submissions of micrographs and descriptions from any type of tissue (e.g. wing imaginal disk, ovary) and at any stage of development. Insertion lines can be searched using a number of queries, including Berkeley Drosophila Genome Project (BDGP) numbers and protein localization. In addition, Flytrap provides online order forms linked to each insertion line so that users may request any line generated from this project. Flytrap may be accessed from the homepage at http://flytrap.med. yale.edu.

Identifying and selecting the most appropriate animal model is a challenge, but the U.S. Federal Government’s National Institutes of Health, has supported several efforts to make the job of finding disease models easier. These efforts began with support for databases about specific species, and have graduated to support for discipline-specific studies and cross-species resources. This article tracks LAMHDI, the initiative to Link Animal Models to Human DIsease ( www.lamhdi.org ), which brings together data about five species: flies, mice, rats, yeast, and zebrafish, and allows scientists to search across these data, principally by disease and gene. This chapter explains how LAMHDI was conceived, and where LAMHDI hopes to go in moving into phenotypical as well as genotypical data through networks and visualization.

For over a hundred years, visible mutations in genes of the fruitfly Drosophila melanogaster have been used to track chromosomal regions and gain insight into basic principles of biology (Morgan, 1910). These so-called “marker” mutations, many coming from the very early years of fly genetics, remain an omnipresent and irreplaceable part of the toolkit used in daily fly pushing activities in modern Drosophila laboratories. Generations of Drosophilists have learned about marker mutations from detailed descriptions and wonderfully instructive drawings in a series of printed catalogues, starting first with the “Dutch Book” (Morgan et al., 1925) and culminating with the famous Red Book (last updated in 1992) (Lindsley and Zimm, 1992). Much of this accumulated information is now accessible through FlyBase, however, all of these classic laboratory reference books are now out of print. Chyb and Gompel's recently published Atlas of Drosophila Morphology: Wild-type and Classical Mutants (Academic Press) provides a timely and much-needed twenty-first century update to the classic line of mutant reference handbooks for Drosophila researchers. Instead of providing a comprehensive catalogue, the Atlas focuses on 70 of the most common marker mutations still in use today at an entirely new level of resolution made possible though the power of digital photography. The technical and aesthetic perfection of the images in the Atlas was achieved by using a custom-built fly photo studio equipped with light-emitting diodes, taking multiple images at different focal planes and assembling them together in imaging software. These images achieve a clarity and crispness that authentically reproduces the direct view of flies under the dissection microscope, far surpassing the utility (and, arguably, also the beauty) of previous black-and-white line drawings, and provide researchers with an entirely new dimension of illustration—color—that is so important for many eye and body markers, yet so difficult to describe in words. Collectively, the images in this book show that Drosophila are truly, in the authors' words, “photogenic flies.” Like the famous “Learning to Fly” poster by Childress, Behringer and Halder (Wiley) that hangs in so many fly labs, and the recently issued photographs of marker mutants provided through FlyBase, the Atlas organizes marker mutations by anatomical structure, e.g. bristle, wing, other appendages, eye, and others. This organization by structure, which is further enhanced through summary photo plates, provides the didactic dimension of comparability that will be appreciated particularly by fly novices. Notably, the images are accompanied by extremely helpful information such as FlyBase entry, phenotypic variation, and similarities to other marker mutations. The Atlas is formatted so that easy-to-read textual information on the left page complements annotated images on the right, and it is spiral-bound so that it lays open flat unaided—all properties that enhance its practical value next to the microscope. The Atlas is also available in an e-book format for those more digitally inclined. But the Atlas has far more to offer. It provides exquisite written and photographic descriptions of the wild-type morphology including bristles and appendages, as well as tips such as sexing of larvae or pupae, or the subtle parameters used for pupal staging. This section, however, could have benefitted from images of the larval mouth hook anatomy used to determine larval stages (Ashburner et al., 2005). A further very useful section describes the markers and chromosomal organization of the major balancer chromosomes, with simple pictograms that illustrate their cytology and photos of flies carrying combinations of markers for a selection of balancers. The closing section provides beautiful images of the adult morphology and distinguishing genitalia of the nine species in the melanogaster group. This echoes the inclusion of wild-type D. melanogaster strains in earlier catalogs (Bridges and Brehme, 1944; Lindsley and Grell, 1968) and highlights how this book feels as much like a field guide as it does like a laboratory handbook. These images of other Drosophila species are unfortunately not accompanied by the kind of textual descriptions that made the other parts of the book so accessible to novices and should be considered in future editions. Like its predecessors, Chyb and Gompel's Atlas is a gift to the fly community that will undoubtedly become a standard addition to Drosophila labs around the world. As part of the modern cannon of Drosophila practical reference works (Greenspan, 2004; Ashburner et al., 2005; Markow and O'Grady, 2005; Roote and Prokop, 2013), the Atlas will serve an important role in helping new and experienced students of the fly to better understand the intricacies of this complex animal. But perhaps most importantly, this book provides a bridge between the classical and modern era of Drosophila research, allowing core knowledge about the fly to be transmitted more easily to new generations of researchers and help maintain Drosophila at the forefront of biological science into the future. The phrase “check in the Atlas!” will undoubtedly echo in the many Drosophila labs around the world for years to come.

Mitochondria are organelles present in the cytoplasm of most eukaryotic cells; although they have their own DNA, the majority of the proteins necessary for a functional mitochondrion are coded by the nuclear DNA and only after transcription and translation they are imported in the mitochondrion as proteins. The primary role of the mitochondrion is electron transport and oxidative phosphorylation. Although it has been studied for a long time, the interest of researchers in mitochondria is still alive thanks to the discovery of mitochondrial role in apoptosis, aging and cancer. Aim of the MitoDrome database is to annotate the Drosophila melanogaster nuclear genes coding for mitochondrial proteins in order to contribute to the functional characterization of nuclear genes coding for mitochondrial proteins and to knowledge of gene diseases related to mitochondrial dysfunctions. Indeed D. melanogaster is one of the most studied organisms and a model for the Human genome. Data are derived from the comparison of Human mitochondrial proteins versus the Drosophila genome, ESTs and cDNA sequence data available in the FlyBase database. Links from the MitoDrome entries to the related homologous entries available in MitoNuC will be soon imple-mented. The MitoDrome database is available at http://bighost.area.ba.cnr.it/BIG/MitoDrome. Data are organised in a flat-file format and can be retrieved using the SRS system.

FlyBase is the model organism database for Drosophila genetic and genomic information. Over the last 20 years, FlyBase has had to adapt and change to keep abreast of advances in biology and database design. We are continually looking for ways to improve curation efficiency and efficacy. Genetic literature curation focuses on the extraction of genetic entities (e.g. genes, mutant alleles, transgenic constructs) and their associated phenotypes and Gene Ontology terms from the published literature. Over 2000 Drosophila research articles are now published every year. These articles are becoming ever more data-rich and there is a growing need for text mining to shoulder some of the burden of paper triage and data extraction. In this article, we describe our curation workflow, along with some of the problems and bottlenecks therein, and highlight the opportunities for text mining. We do so in the hope of encouraging the BioCreative community to help us to develop effective methods to mine this torrent of information. Database URL: http://flybase.org

RNA-Seq is a powerful tool for the annotation of genomes, in particular for the identification of isoforms and UTRs. Nevertheless, several software tools exist and no standard strategy to obtain a reliable annotation is yet established. We tested different combinations of the most commonly used reference-based alignment tools (TopHat, GSNAP) in combination with two frequently used reference-based assemblers (Cufflinks, Scripture) and evaluated the potential of RNA-Seq to improve the annotation of Drosophila pseudoobscura. While GSNAP maps a higher proportion of reads, TopHat resulted in a more accurate annotation when used in combination with Cufflinks. Scripture had the lowest sensitivity. Interestingly, after subsampling to the same coverage for GSNAP and TopHat, we find that both mappers have similar performance, implying that the advantage of TopHat is mainly an artifact of the lower coverage. Overall, we observed a low concordance among the different approaches tested both at junction and isoform levels. Using data from both sexes of two adult strains of D. pseudoobscura we detected alternative splicing for about 30% of the FlyBase multiple-exon genes. Moreover, we extended the boundaries for 6523 genes (about 40%). We annotated 669 new genes, 45% of them with splicing evidence. Most of the new genes are located on unassembled contigs, reflecting their incomplete annotation. Finally, we identified 99 additional new genes that are not represented in the current genome contigs of D. pseudoobscura, probably due to location in genomic regions that are difficult to assemble (e.g. heterochromatic regions).

Much of the data within Model Organism Databases (MODs) comes from manual curation of the primary research literature. Given limited funding and an increasing density of published material, a significant challenge facing all MODs is how to efficiently and effectively prioritize the most relevant research papers for detailed curation. Here, we report recent improvements to the triaging process used by FlyBase. We describe an automated method to directly e-mail corresponding authors of new papers, requesting that they list the genes studied and indicate (‘flag’) the types of data described in the paper using an online tool. Based on the author-assigned flags, papers are then prioritized for detailed curation and channelled to appropriate curator teams for full data extraction. The overall response rate has been 44% and the flagging of data types by authors is sufficiently accurate for effective prioritization of papers. In summary, we have established a sustainable community curation program, with the result that FlyBase curators now spend less time triaging and can devote more effort to the specialized task of detailed data extraction. Database URL: http://flybase.org/

Background The enormous potential of natural variation for the functional characterization of genes has been neglected for a long time. Only since recently, functional geneticists are starting to account for natural variation in their analyses. With the new sequencing technologies it has become feasible to collect sequence information for multiple individuals on a genomic scale. In particular sequencing pooled DNA samples has been shown to provide a cost-effective approach for characterizing variation in natural populations. While a range of software tools have been developed for mapping these reads onto a reference genome and extracting SNPs, linking this information to population genetic estimators and functional information still poses a major challenge to many researchers. Results We developed PoPoolation DB a user-friendly integrated database. Popoolation DB links variation in natural populations with functional information, allowing a wide range of researchers to take advantage of population genetic data. PoPoolation DB provides the user with population genetic parameters (Watterson's θ or Tajima's π), Tajima's D, SNPs, allele frequencies and indels in regions of interest. The database can be queried by gene name, chromosomal position, or a user-provided query sequence or GTF file. We anticipate that PoPoolation DB will be a highly versatile tool for functional geneticists as well as evolutionary biologists. Conclusions PoPoolation DB, available at http://www.popoolation.at/pgt, provides an integrated platform for researchers to investigate natural polymorphism and associated functional annotations from UCSC and Flybase genome browsers, population genetic estimators and RNA-seq information.

Genomic annotations describing structural and functional features of genes and gene products through controlled terminologies and ontologies are extremely valuable, especially for computational analyses aimed at inferring new biomedical knowledge, which rely on available annotations. Yet, they are incomplete, especially for recently studied genomes, and only some of available annotations represent highly reliable human curated information. In order to help and speedup the time-consuming curation process and improve available annotations, computational methods able to provide prioritized lists of predicted annotations are paramount. Starting from a previous work on automatic prediction of Gene Ontology annotations based on singular value decomposition (SVD) of gene-to-term annotation matrix, here we propose a novel prediction algorithm that incorporates gene clustering based on gene functional similarity computed on Gene Ontology annotations. We tested both prediction methods performing k-fold cross-validation on two organism genomes, Saccharomyces cerevisiae (SGD) and Drosophila melanogaster (FlyBase). Results demonstrate effectiveness of our approach.

In collaboration with the journal GENETICS, we've developed and launched a pipeline by which interactive full-text HTML/PDF journal articles are published with named entities linked to corresponding resource pages in "WormBase":http://www.wormbase.org/ (WB). Our interactive articles allow a reader to click on over ten different data type objects (gene, protein, transgene, etc.) and be directed to the relevant webpage. This seamless connection from the article to summaries of data types promotes a deeper level of understanding for the naïve reader, and incisive evaluation for the sophisticated reader. Further, this collaboration allows us to identify and collect information before the publication of the article. The pipeline uses automated recognition scripts to identify entities that already exist in the database and a self-reporting form we created at WB that is sent to the author by GENETICS for submitting entities that do not already exist in our database. We include a manual quality control step to make sure ambiguous links are corrected, and that all new entities have been reported and linked properly. The automated entity recognition scripts allows us to potentially link any object found in a database as well as to expand this pipeline to other databases. We have already adapted this pipeline for linking Saccharomyces cerevisiae GENETICS articles to the "Saccharomyces Genome Database":http://www.yeastgenome.org/ (SGD) and are currently expanding this pipeline for linking genes in Drosophila articles to "FlyBase":http://flybase.org/. By integrating journals and databases, we are integrating the major modes of communication in the biological sciences, which will undoubtedly increase the pace of discovery.

The body wall musculature of a Drosophila larva is composed of an intricate pattern of 30 segmentally repeated muscle fibers in each abdominal hemisegment. Each muscle fiber has unique spatial and behavioral characteristics that include its location, orientation, epidermal attachment, size and pattern of innervation. Many, if not all, of these properties are dictated by founder cells, which determine the muscle pattern and seed the fusion process. Myofibers are then derived from fusion between a specific founder cell and several fusion competent myoblasts (FCMs) fusing with as few as 3-5 FCMs in the small muscles on the most ventral side of the embryo and as many as 30 FCMs in the larger muscles on the dorsal side of the embryo. The focus of the present review is the formation of the larval muscles in the developing embryo, summarizing the major issues and players in this process. We have attempted to emphasize experimentally-validated details of the mechanism of myoblast fusion and distinguish these from the theoretically possible details that have not yet been confirmed experimentally. We also direct the interested reader to other recent reviews that discuss myoblast fusion in Drosophila, each with their own perspective on the process [1-4]. With apologies, we use gene nomenclature as specified by Flybase (http://flybase.org) but provide Table 1 with alternative names and references.

We propose an algorithm that predicts potentially missing Gene Ontology annotations, in order to speed up the time-consuming annotation curation process. The proposed method extends a previous work based on the singular value decomposition of the gene-term annotation matrix and incorporates gene clustering, based on gene functional similarity computed by means of the Gene Ontology annotations. We tested the prediction method by performing K-fold cross-validation on the genomes of two organisms, Saccharomyces cerevisiae (SGD) and Drosophila melanogaster (FlyBase).

Protein-trap strains of Drosophila melanogaster provide a very useful tool for examining the 3D-expression patterns of proteins and purification of protein complexes. Here we present BrainTrap, available at http://fruitfly.inf.ed.ac.uk/braintrap, an online database of 3D confocal datasets showing reporter gene expression and protein localization in the adult brain of Drosophila. Full size images throughout the volume of the entire brain can be viewed interactively in a web browser. The database includes searchable annotations linked to the FlyBase Drosophila anatomy ontology. Anatomical search criteria can be specified using automatic completion and a hierarchical browser for the ontology. The provenance of all annotation is retained and the location where the annotator made the conclusion can be highlighted. Database URL: http://fruitfly.inf.ed.ac.uk/braintrap

The Gene Ontology (GO) is a well-established, structured vocabulary that has been successfully used for 10 years in the annotation of proteins. GO terms, created in consultation with the biology community, are used to replace the multiple nomenclatures used by scientific databases that can hamper data integration. Currently GO consists of more than 26,500 terms distributed over three ontologies that describe the molecular function, biological process and subcellular location of a protein in a generic cell.The Gene Ontology Annotation (GOA) database ("http://www.ebi.ac.uk/GOA":http://www.ebi.ac.uk/GOA) aims to provide high-quality manual and electronic GO annotations to proteins within the UniProt Knowledgebase (UniProtKB). By annotating all ‘known’ proteins with GO terms and transferring this knowledge to highly similar ‘unknown’ proteins, GOA offers a valuable contribution to the understanding of all proteomes.As well as generating manual annotation, made by extracting experimental evidence from full text peer-reviewed publications, GOA produces electronic annotation by making large-scale assignments of GO terms to proteins using computational methods. To date we have six electronic annotation methods including; InterPro2GO, Swiss-Prot Keyword2GO and the projection of annotations between orthologous species using Ensembl Compara.GOA provides annotated entries for over 180,000 species and is the largest and most comprehensive open-source contributor of annotations to the GO Consortium annotation effort. In addition, by integrating GO annotations from model organism groups (e.g. FlyBase, GeneDB, MGI, RGD, SGD and TAIR), GOA ensures the dataset remains a key reference. GOA prioritises the annotation of the human proteome and provides this annotation to the GO Consortium’s Reference Genome project.GOA produces monthly releases of annotations to the human, mouse, rat, zebrafish, cow, chicken and Arabidopsis proteomes as well as a file for the multiple species within UniProtKB. The GOA dataset can be queried through a user-friendly web interface via our QuickGO browser("http://www.ebi.ac.uk/QuickGO":http://www.ebi.ac.uk/QuickGO) or downloaded in a parsable format via the EBI("ftp://ftp.ebi.ac.uk/pub/databases/GO/goa":ftp://ftp.ebi.ac.uk/pub/databases/GO/goa) and GO FTP sites. The GOA dataset has increasingly been integrated into tools that aid in the analysis of large datasets resulting from high-throughput experiments thus assisting researchers in biological interpretation of their results.

The amount of biological data has risen exponentially over the last decade. Along with this rise, the number and types of bioinformatics resources has risen such that the sheer number of bioinformatics resources is overwhelming. For these resources to attain their full potential, they must be efficiently and extensively utilized. But in such a plethora of resources, how does a researcher new to a field find the resources that will meet their needs. Once a resource is found, how does the researcher quickly learn to utilize that resource fully? There are resource lists such as those provided by the journal Nucleic Acid Research (NAR), BioMed Central and the Univ. of Pittsburgh Health Sciences Library, and most resources include their own documentation. But lists and site documentation don’t always cast a wide enough net to catch all users. For a resource to truly maximize their user community, it often takes multiple different outreach approaches. OpenHelix specializes in providing customized outreach services to bioinformatics resources, including those featured on this poster. Based on the conclusions from our Phase I SBIR grant, which tested the efficiency of several methods for training researchers on the use of genomic resources, OpenHelix ("www.openhelix.com":http://www.openhelix.com) has developed, and currently provides up-to-date online training materials on a large number of bioinformatics resources, covering major providers and research areas. Through a Phase II SBIR grant from NHGRI and other funding, we are developing a search portal for online bioinformatics resources. With the search portal, researchers will be able to find the bioinformatics and genomics online databases and resources most relevant to their needs. The search portal will contain an index of hundreds of the most popular and powerful resources, as well as the content of over 100 OpenHelix tutorials. Using various ranking techniques, the portal will be able to provide more relevant results than a simple keyword search. Additionally, as part of an extension of our Phase II grant, we are providing a sponsored suite of Model Organism Database trainings that include GBrowse, FlyBase, MGI, RGD, SGD, WormBase and ZFIN. We are also developing, and will be offering, sponsored live trainings on resources at NCBI as a way of filling the void left by the demise of the NCBI Field Guide outreach programs. Conclusion: OpenHelix aims to increase community curation by helping resources increase their user community

Complete functional annotation of genomes is a powerful tool for researchers; however, such annotation is a time-consuming task limited by the availability of experimental data. The function of genes for which there is no experimental data can often be predicted via comparison to related, annotated sequences of known function. We describe here the Reference Genome project, an effort from the Gene Ontology (GO) Consortium to fully annotate twelve genomes to rigorous standards: human, plus eleven organisms that are important models in biomedical research, including mouse, fly, zebrafish, yeast and E. coli. To achieve this, we examine existing experimentally based annotations in a phylogenetic context in order to infer the function(s) of ancestral proteins and propagate these annotations to their descendants. This endeavor faces many difficult challenges, such as: the determination and provision of reference protein sets for each genome; the identification of gene families for curation; the application of consistent best practices for annotation; the development of methodologies for evaluating progress towards our goal; and the development of software tools to support this effort. Annotated genomes are greatly valuable to the research community and will provide the basis for using sequence similarity to annotate further genomes. An overview of the project as well as links to all resources described below can be found at "http://geneontology.org/GO.refgenome.shtml":http://geneontology.org/GO.refgenome.shtmlThis work is supported by NHGRI grant HG002273 (Gene Ontology Consortium) and NIGMS GM081084-01A1 (Phylogenetic tree building and annotation software development). Pascale Gaudet for the Reference Genome Group of the Gene Ontology Consortium. The Reference Genome project is overseen by Pascale Gaudet (dictyBase), Rex Chisholm (dictyBase) and includes these representatives from the curatorial staff: Tanya Berardini (TAIR), Emily Dimmer (GOA), Stacia R. Engel (SGD), Petra Fey (dictyBase), David P. Hill (MGI), Doug Howe (ZFIN), Jim Hu (EcoliWiki), Rachael Huntley (GOA), Varsha K. Khodiyar (UCL), Ranjana Kishore (WormBase), Donghui Li (TAIR), Ruth C. Lovering (UCL), Fiona McCarthy (AgBase), Li Ni (MGI), Victoria Petri (RGD), Deborah A. Siegel (EcoliWiki), Susan Tweedie (FlyBase), Kimberly Van Auken (WormBase), and Valerie Wood (GeneDB)—as well as the following computational staff representatives: Siddhartha Basu (dictyBase), Seth Carbon (BBOP), Mary Dolan (MGI), and Christopher J. Mungall (BBOP)—those establishing the protein families to be annotated: Kara Dolinski (PPOD), Michael S. Livstone (PPOD), and Paul Thomas (PANTHER)—and, the four PIs of the GO Consortium: Michael Ashburner (FlyBase), Judith A. Blake (MGI), J. Michael Cherry (SGD), and Suzanna E. Lewis (BBOP).

As research in the biological sciences continues to advance at a rapid pace, it is increasingly important that the data be captured, standardized, organized, and made accessible to the scientific community. This is the job of a biocurator. Here we describe the process of biocuration from our perspective as FlyBase curators.

Background The availability of 12 fully sequenced Drosophila species genomes provides an excellent opportunity to explore the evolutionary mechanism, structure and function of gene families in Drosophila. Currently, several important resources, such as FlyBase, FlyMine and DroSpeGe, have been devoted to integrating genetic, genomic, and functional data of Drosophila into a well-organized form. However, all of these resources are gene-centric and lack the information of the gene families in Drosophila. Description FlyPhy is a comprehensive phylogenomic analysis platform devoted to analyzing the genes and gene families in Drosophila. Genes were classified into families using a graph-based Markov Clustering algorithm and extensively annotated by a number of bioinformatic tools, such as basic sequence features, functional category, gene ontology terms, domain organization and sequence homolog to other databases. FlyPhy provides a simple and user-friendly web interface to allow users to browse and retrieve the information at multiple levels. An outstanding feature of the FlyPhy is that all the retrieved results can be added to a workset for further data manipulation. For the data stored in the workset, multiple sequence alignment, phylogenetic tree construction and visualization can be easily performed to investigate the sequence variation of each given family and to explore its evolutionary mechanism. Conclusion With the above functionalities, FlyPhy will be a useful resource and convenient platform for the Drosophila research community. The FlyPhy is available at http://bioinformatics.zj.cn/fly/.

Background Despite increasing interest in applying Natural Language Processing (NLP) to biomedical text, whether this technology can facilitate tasks such as database curation remains unclear. Results PaperBrowser is the first NLP-powered interface that was developed under a user-centered approach to improve the way in which FlyBase curators navigate an article. In this paper, we first discuss how observing curators at work informed the design and evaluation of PaperBrowser. Then, we present how we appraise PaperBrowser's navigational functionalities in a user-based study using a text highlighting task and evaluation criteria of Human-Computer Interaction. Our results show that PaperBrowser reduces the amount of interactions between two highlighting events and therefore improves navigational efficiency by about 58% compared to the navigational mechanism that was previously available to the curators. Moreover, PaperBrowser is shown to provide curators with enhanced navigational utility by over 74% irrespective of the different ways in which they highlight text in the article. Conclusion We show that state-of-the-art performance in certain NLP tasks such as Named Entity Recognition and Anaphora Resolution can be combined with the navigational functionalities of PaperBrowser to support curation quite successfully.

FlyBase ( http://flybase.org ) is the primary database of integrated genetic and genomic data about the Drosophilidae, of which Drosophila melanogaster is the most extensively studied species. Information in FlyBase originates from a variety of sources ranging from large-scale genome projects to the primary research literature. Data-types include sequence-level gene models, molecular classification of gene product functions, mutant phenotypes, mutant lesions and chromosome aberrations, gene expression patterns, transgene insertions, and anatomical images. Query tools allow interrogation of FlyBase through DNA or protein sequence, by gene or mutant name, or through terms from the several ontologies used to capture functional, phenotypic, and anatomical data. Links between FlyBase and external databases provide extensive opportunity for extending exploration into other model organism databases and resources of biological and molecular information. This review will introduce the FlyBase web server and query tools.

Web data integration has been an active research topic for the last few years. Bioinformatics is a huge domain wealth of data. By reverse engineering web data, in this paper we report a Java based implementation of biologically relevant application of web data integration consisting of few sites such as Flybase, YMF and STAMP. We have developed a tool that can automate the finding of transcription factor binding sites for arbitrary gene regulatory networks. We employed vertical (i.e link) and horizontal (i.e combine) integration for the whole process. As a running example and proof of concept of our approach, we report here a real example motivated by the biological domain of Drosophila Melanogaster model organism database. For a given set of genes, we could find a set of sequences (800~1500 bp) from Flybase site, then we feed those into YMF site to get a motif of 6{~}8 bp. The result motifs were fed into STAMP site and by combining result from Fly curated by Bergman and by TRANSFAC, we were able to get a sample correct result for a Bicoid gene, namely bin. The lesson learned from the exercise can be applied in generic dynamic web form based data integration application like e-commerce, comparison shopping etc.

BackgroundThe skills used by winged insects to explore their environment are strongly dependent upon the integration of neurosensory information comprising visual, acoustic and olfactory signals. The neuronal architecture of the wing contains a vast array of different sensors which might convey information to the brain in order to guide the trajectories during flight. In Drosophila, the wing sensory cells are either chemoreceptors or mechanoreceptors and some of these sensors have as yet unknown functions. The axons of these two functionally distinct types of neurons are entangled, generating a single nerve. This simple and accessible coincidental signaling circuitry in Drosophila constitutes an excellent model system to investigate the developmental variability in relation to natural behavioral polymorphisms.Methodology/principal findingsA fluorescent marker was generated in neurons at all stages of the Drosophila life cycle using a highly efficient and controlled genetic recombination system that can be induced in dividing precursor cells (MARCM system, flybase web site). It allows fluorescent signals in axons only when the neuroblasts and/or neuronal cell precursors like SOP (sensory organ precursors) undergo division during the precedent steps. We first show that a robust neurogenesis continues in the wing after the adults emerge from the pupae followed by an extensive axonal growth. Arguments are presented to suggest that this wing neurogenesis in the newborn adult flies was influenced by genetic determinants such as the frequency dependent for gene and by environmental cues such as population density.ConclusionsWe demonstrate that the neuronal architecture in the adult Drosophila wing is unfinished when the flies emerge from their pupae. This unexpected developmental step might be crucial for generating non-heritable variants and phenotypic plasticity. This might therefore constitute an advantage in an unstable ecological system and explain much regarding the ability of Drosophila to robustly adapt to their environment.

We proposed a new measure (SimPLD) for calculating the semantic similarity of terms in gene ontology (GO) based on the depth of least common ancestor (LCA) of two terms and the path length between them in GO hierarchy. The similarity between genes is computed based on this measure when it is applied to the GO-terms related to those genes. The method is based on the average of SimPLD between the GO terms annotated for both genes in a given gene pair. We evaluated the proposed method with a series of experiments on large groups of genes and proteins from two genomes: Saccharomyces database (SGD) and Drosophila Melanogaster (FlyBase); and one dataset of Human-Yeast protein pairs. The experimental results proved that the method has fairly impressive agreement with Blast sequence similarity. Therefore SimPLD can be used as an automated tool for determining the similarity between genes and proteins.

AphidBase aims to (i) store recently acquired genomic resources on aphids and (ii) compare them to other insect resources as functional annotation tools. For that, the Drosophila melanogaster genome has been loaded in the database using the GMOD open source software for a comparison with the 17 069 pea aphid unique transcripts (contigs) and the 13 639 gene transcripts of the Anopheles gambiae. Links to FlyBase and A.gambiae Entrez databases allow a rapid characterization of the putative functions of the aphid sequences. Text mining of the D.melanogaster literature was performed to construct a network of co-cited gene or protein names, which should facilitate functional annotation of aphid homolog sequences. AphidBase represents one of the first genomic databases for a hemipteran insect. Availability: http://w3.rennes.inra.fr/AphidBase Contact: denis.tagu@rennes.inra.fr, jean-pierre.gauthier@rennes.infra.fr

Background Proteins control and mediate many biological activities of cells by interacting with other protein partners. This work presents a statistical model to predict protein interaction networks of Drosophila melanogaster based on insight into domain interactions. Results Three high-throughput yeast two-hybrid experiments and the collection in FlyBase were used as our starting datasets. The co-occurrences of domains in these interactive events are converted into a probability score of domain-domain interaction. These scores are used to infer putative interaction among all available open reading frames (ORFs) of fruit fly. Additionally, the likelihood function is used to estimate all potential protein-protein interactions. All parameters are successfully iterated and MLE is obtained for each pair of domains. Additionally, the maximized likelihood reaches its converged criteria and maintains the probability stable. The hybrid model achieves a high specificity with a loss of sensitivity, suggesting that the model may possess major features of protein-protein interactions. Several putative interactions predicted by the proposed hybrid model are supported by literatures, while experimental data with a low probability score indicate an uncertain reliability and require further proof of interaction. Fly-DPI is the online database used to present this work. It is an integrated proteomics tool with comprehensive protein annotation information from major databases as well as an effective means of predicting protein-protein interactions. As a novel search strategy, the ping-pong search is a naïve path map between two chosen proteins based on pre-computed shortest paths. Adopting effective filtering strategies will facilitate researchers in depicting the bird's eye view of the network of interest. Fly-DPI can be accessed at http://flydpi.nhri.org.tw. Conclusion This work provides two reference systems, statistical and biological, to evaluate the reliability of protein interaction. First, the hybrid model statistically estimates both experimental and predicted protein interaction relationships. Second, the biological information for filtering and annotation itself is a strong indicator for the reliability of protein-protein interaction. The space-temporal or stage-specific expression patterns of genes are also critical for identifying proteins involved in a particular situation.

FlyBase (http://flybase.org/) is the primary database of genetic and genomic data for the insect family Drosophilidae. Historically, Drosophila melanogaster has been the most extensively studied species in this family, but recent determination of the genomic sequences of an additional 11 Drosophila species opens up new avenues of research for other Drosophila species. This extensive sequence resource, encompassing species with well-defined phylogenetic relationships, provides a model system for comparative genomic analyses. FlyBase has developed tools to facilitate access to and navigation through this invaluable new data collection.

Biologists often need to find information about genes whose function is not described in the genome databases. Currently they must try to search disparate biomedical literature to locate relevant articles, and spend considerable efforts reading the retrieved articles in order to locate the most relevant knowledge about the gene. We describe our software, the first that automatically generates gene summaries from biomedical literature. We present a two-stage summarization method, which involves first retrieving relevant articles and then extracting the most informative sentences from the retrieved articles to generate a structured gene summary. The generated summary explicitly covers multiple aspects of a gene, such as the sequence information, mutant phenotypes, and molecular interaction with other genes. We propose several heuristic approaches to improve the accuracy in both stages. The proposed methods are evaluated using 10 randomly chosen genes from FlyBase and a subset of Medline abstracts about Drosophila. The results show that the precision of the top selected sentences in the 6 aspects is typically about 50-70%, and the generated summaries are quite informative, indicating that our approaches are effective in automatically summarizing literature information about genes. The generated summaries not only are directly useful to biologists but also serve as useful entry points to enable them to quickly digest the retrieved literature articles.

In this small and charming book, Won for All, Michael Ashburner gives us a glittering account of the sequencing of the Drosophila genome by a public–private partnership between government-funded laboratories and Celera Genomics. He portrays both the working life and the good life of science, with neat character sketches set off by Lewis Miller's excellent portraits. Michael's flair for detail and inveterate name-dropping, albeit of restaurants rather than people, lends itself nicely to re-creating the time and place of key events in this collaboration. The original fast-paced manuscript, which I liked so well when I first saw a draft in 2001, has been updated and provided with extensive footnotes that inform without interrupting the narrative. Technical background is given in two excellent postscripts: a fly primer from Scott Hawley, and an overview of fly functional genomics from Ethan Bier. Michael and I have been orbiting around Cambridge (the original one, as he puts it) for many years. Our paths intersect occasionally and pleasantly, but never so forcibly as over genomics—in keeping with Maurice Wilkins' remark to Horace Judson that: “DNA, you know, is Midas' gold. Everyone who touches it goes mad” [ 1]. It has indeed proved so in the age of sequencing genomes. One of the emergent themes of the book is competition between animal model organisms, which was intense from the 1960s to the 1990s. Flies, worms, molluscs, mice, and men: each had its own set of dedicated followers, and each had strengths and weaknesses in the common goal of understanding the biology of animals and humans in particular. At a slightly greater distance, there was competition with micro-organisms and plants for understanding life in general. Compared with the nematode worm, on which I worked, the fly was the doyen, with half a century of intensive work behind it, 10-fold more researchers, a rich morphology of wings and legs and bristles, and a huge range of sophisticated genetic techniques. The worm was a shy newcomer, kicked into excitement by Sydney Brenner, who chose it for its small size and rapid growth. Never mind the fly's sophistication; for those of us who came early the worm was a sandpit to play in: anything we did was novel because so little had been done before. With the advent of molecular cloning in the 1970s, developmental biology entered a new and thrilling phase. No longer were experiments scored only by visible outcomes; now the aim was to isolate and study the genes controlling these tangible phenomena. There were triumphs: for example, homeobox genes controlling development in the fly, and cell-death genes found in the worm. And now the fly displayed anew one of its long-known attributes: the polytene chromosomes in the salivary glands. Composed of multiple aligned copies of the fly's DNA, they provide an extraordinary pocket map of its genome, gratis. The polytenes were so advantageous to gene hunters that attempts were made to find a genetic modification of the worm that would produce something comparable, to no avail. However, this initial disadvantage soon turned to the advantage of the worm researchers, because we were forced to construct our own physical map from cloned fragments of worm DNA. Clumsy and incomplete at first, the map grew into a full coverage of the genome. We arrayed our clones on membranes and proudly announced our artificial polytenes. And the arrays were more than polytenes: not only a map, but actual fragments that could be shipped around the world and used directly in people's experiments. More was to come. The physical map allowed the worm an early start on genome sequencing, long before automation was sufficiently advanced to tackle the whole thing. Our sequence was publicly released as we went along, and soon, as researchers isolated individual genes from human and other organisms, they more often found a match to worm than anything else. People were drawn to the worm, and it prospered, increasing pressure on the fly community to start their own genome project. I had nothing to do with the fly scientifically, but its genome made a big impact on me through the Human Genome Project. Michael disarmingly suggests that it's a mystery why Celera would sequence the fly for no financial return. The reason, of course, was that it provided a doable and very public advertisement for the human genome sequence, which was Celera's real business. It meant that Gerry Rubin, as leader of the fly sequencing project, could be persuaded to speak in support of the company at US congressional hearings that discussed appropriations for the public sequencing effort. Celera's fly sequence was indeed “won for all,” but for the human it would be a different story, as those who searched for public–private partnership discovered in December 1999 [ 2]. And where human sequence went, there went everything. Had the human genome sequencing project gone private, the public databases, which are one of the most important foundations of modern biology, would have withered and been replaced by a corporate entity under the control not of its users but of its shareholders. Gerry was the fly hero, doing what he had to do to get his organism sequenced. But despite Scott Hawley's opening words, Michael is a hero for us all. Celera initially posted the fly sequence with copyright restrictions, against all previous agreements, and it was Michael who ensured that the restrictions were removed. Later, he was prominent in supporting free release of the human sequence. Today, in his development of FlyBase and Gene Ontology, and in his championing of open-access publication, he continues as a giant of public-domain science. The marvellous outcome is that the worm, the fly, the human, and hundreds of other genomes have all been won for all. The rivalries have been transformed into a new wholeness of biology, united through its common history written in the language of the genomes.

The maintenance of biomedical digital libraries (including organism databases and protein databases) involves analysis of a large number of documents. Much work is done manually: curators study large numbers of biomedical documents while updating and annotating organism databases such as MGI (mouse genome informatics) and Flybase (a database of the fruit-fly genome). We summarize the annotation process in organism databases, and describe some of the roles played by the gene ontology and by document databases such as PubMed. Efforts are ongoing to automate parts of the annotation process. Biomedical text mining contests, such as the TREC Genomics Track (Hersh et al., 2004, 2005), define annotation subtasks, and provide training and test data. So far, these efforts have focused on the analysis of the text content of documents. We are investigating the analysis of figures in biomedical documents; the information derived from figure analysis may later be combined with the information derived from text analysis. We present an algorithm for using figures in document triage; triage involves determining which documents are relevant to a given annotation task. In our triage algorithm, we segment figures into subfigures and classify the subfigures as graphical, gel, fluorescence microscopy, and other microscopy. A secondary classification into subcategories is performed by clustering, using clusters created from the subfigures in the labeled training data. The classifications of all subfigures in a document are combined to form a document descriptor. The document descriptor is then classified using a naive Bayes classifier, as either relevant or irrelevant to the given annotation task.

This paper demonstrates how Drosophila gene name recognition and anaphoric linking of gene names and their products can be achieved using existing information in FlyBase and the Sequence Ontology. Extending an extant approach to gene name recognition we achieved a F-score of 0.8559, and we report a preliminary experiment using a baseline anaphora resolution algorithm. We also present guidelines for annotation of gene mentions in texts and outline how the resulting system is used to aid FlyBase curation.

The representation of biological knowledge in databases is a necessity for modern biomedical research. Historically, there has been very little collaboration or coordination between different database providers. Although many grandiose schemes for the "integration" of biological databases have been proposed over the years, none have been practical to the point of implementation. Yet the need for integration remains, as many biologists, both those at the bench and those who analyse data computationally, wish to integrate data from a diversity of sources. The Gene Ontology Consortium (GOC) began, some seven years ago, to develop a resource that could be used by both the model organism databases (e.g. FlyBase, WormBase, Mouse Genome Database, The Arabidopsis Information Resource) and the large "horizontal" databases (e.g. UniProt, GeneDB, TIGR Gene Index) as a standard for the annotation of gene products. The GOC now maintains several structured controlled vocabularies for the annotation of gene products. The first three of these are used for the annotation of gene products with respect to these domains: their molecular function, their cellular location and the biological processes in which they are involved. This database of nearly 18,000 terms is now used for the annotation of the gene products of all of the major experimental eukaryotes and many prokaryotes. The philosophy of the GOC is now being extended to cover further domains of biological knowledge. Under the umbrella of "obo" (open biological ontologies) structured controlled vocabularies are now available, or are being developed, for sequence annotation, anatomies and development, cells and tissues, mouse pathology and experimental treatments. In this talk I will discuss how the concept of ontologies can be used for the intelligent design of database schema, and for the development of common tools for data exchange. I will also discuss some of the major limitations of the current models of data representation used by the GO Consortium, and proposals that will make the design of ontologies for shared use both more flexible and powerful.

The Gene Ontology (GO) is a controlled vocabulary widely used for the annotation of gene products. GO is organized in three hierarchies for molecular functions, cellular components, and biological processes but no relations are provided among terms across hierarchies. The objective of this study is to investigate three non-lexical approaches to identifying such associative relations in GO and compare them among themselves and to lexical approaches. The three approaches are: computing similarity in a vector space model, statistical analysis of co-occurrence of GO terms in annotation databases, and association rule mining. Five annotation databases (FlyBase, the Human subset of GOA, MGI, SGD, and WormBase) are used in this study. A total of 7,665 associations were identified by at least one of the three non-lexical approaches. Of these, 12% were identified by more than one approach. While there are almost 6,000 lexical relations among GO terms, only 203 associations were identified by both non-lexical and lexical approaches. The associations identified in this study could serve as the starting point for adding associative relations across hierarchies to GO, but would require manual curation. The application to quality assurance of annotation databases is also discussed.