FlyBase:Overview

Introduction

Since its inception in 1992, FlyBase has provided an online repository of biological data about Drosophila species, focusing on the model organism D. melanogaster. Data in FlyBase are either curated manually from the primary research literature or are incorporated computationally from various sources, with the two input streams being integrated into a series of ‘report’ pages and other portals on the website. Many links to related data and resources at external databases are also incorporated. As a result, FlyBase serves as a nexus for all Drosophila-related information.

Much of the information in FlyBase is partitioned into 20 different data classes, each with an associated report on the website (Table 1). Traditional reports, such as those for genes, alleles or references, have been supplemented with several new ones in recent years to reflect the different reagents being used and new data types being produced. New or improved tools have also been developed in order to effectively search and analyze these novel data. A list of all current FlyBase tools, along with a brief description of their functionalities, is provided in Table 2 and specific use cases will be presented in context below.

This text provides an overview of the data content, organization and available tools within FlyBase. It is intended to serve as a basic introduction and guide for our diverse user base. Please refer to documentation on the FlyBase website and in cited publications for more details.

Table 1: Data classes and reports in FlyBase

Genetic/Genomic data	FB id prefix
Genes	FBgn
Transcripts	FBtr
Polypeptides	FBpp
Alleles	FBal
Aberrations	FBab
Balancers	FBba
Recombinant Constructs	FBtp
Insertions	FBti
Natural transposons	FBte
Sequence Features	FBsf
Physical Interactions	FBig
Reagents
Stocks	FBst
Strains	FBsr
Cell Lines	FBtc
Clones	FBcl
Integrated Data
Large Dataset Metadata	FBlc
Gene Groups	FBgg
Human Disease Models	FBhh
Other
References	FBrf
Images	FBim

Table 2: Tools in FlyBase

Tool name	Function/usage
Query Tools/Portals
QuickSearch	Simplified searches on various data classes
QueryBuilder	Advanced search on a field-by-field level for most data classes
Vocabularies	Search or browse all controlled vocabularies used to annotate records
CytoSearch	Search for genetic objects mapped via cytology-based data
RNA-Seq Profile	Find genes with specific patterns of expression across modENCODE developmental stage, tissue, treatment, or cell line RNA-Seq data
Interactions Browser	Explore genetic and physical interactions via static images
ImageBrowse	Browse through Drosophila images by organ system, life-cycle, tagma, or germ layer
Genomic/Map tools
BLAST	NCBI BLAST for finding nucleotide and protein sequences in Drosophila and dozens of related species
GBrowse	Graphical or tabular representation of the 12 sequenced Drosophila genomes
CytoSearch	Search for genetic objects mapped via sequence-based data
Feature Mapper	Search for sequence-mapped features that overlap a specific region or gene
RNA-Seq By Region	Evaluate the expression levels of exons, introns and/or intergenic regions from modENCODE developmental and tissue RNA-Seq data
Coordinates Converter	Convert coordinates between different genome releases (e.g. D. melanogaster R5 to R6)
Retrieve/Convert Tools
Batch Download	Bulk download of individual report fields, FASTA or XML files in a variety of formats
Coordinates Converter	(See Coordinates Converter above)
Upload/Convert IDs	Update lists of old IDs, convert lists (e.g. genes to alleles), or upload IDs into a hit list

The Homepage

The main purpose of the homepage is to provide an indication of, and intuitive access to, all available data and tools in FlyBase (Fig. 1). In addition, the homepage highlights new features within FlyBase and advertises topical issues of interest to the fly community.

Overview

The ‘Navigation Bar’ (NavBar) along the top of this (and every FlyBase) page incorporates drop-down menus containing links to key pages. For example, the ‘Tools’ menu lists all FlyBase tools, grouped by usage, and includes a ‘Tools Overview’ page to help users with search strategies. ‘Downloads’ contains links to all bulk data files that are available to download (see section 10.2), also with an overview page describing them. The ‘Links’ menu has direct links to major external sites, such as the Berkeley Drosophila Genome Project (BDGP) and modENCODE, along with a comprehensive list of ∼250 network and ∼75 reagent resources of interest to Drosophila researchers. A new ‘Community’ menu has been added to group features that facilitate interactions between researchers and FlyBase, such as Fast-Track Your Paper (FTYP) and the FlyBase Community Advisory Group (FCAG) (see section 11). The ‘About’ menu collates general information about the FlyBase database and consortium, and includes links to FlyBase release notes and to a listing of FlyBase-authored publications. The ubiquitous NavBar also features the ‘Jump to Gene’ box (see section 2.3).

Immediately below the NavBar on the homepage are prominent pictographs providing direct links to the most popular tools. Direct links to important community features are shown on the left-hand side of the homepage, including FTYP and the ‘FlyBase Newsletter’. Beneath these are links to general ‘News’ items and upcoming ‘Meetings’ and ‘Courses’. New or improved features within FlyBase are usually accompanied by an extended ‘Commentary’--an abbreviated teaser section is shown on the homepage, and all recent teasers cycle in order that several concurrent improvements can be brought to the attention of users. A link is also provided to view all current and previous Commentaries.

The FlyBase website is updated with new data (and often new features) about 6 times per year (see the ‘Release Schedule’ link under the ‘About’ menu of the NavBar). The header and footer of the homepage, and indeed every FlyBase page, state the version number and date of release of the FlyBase instance currently being viewed (Fig. 1). For example, ‘FB2015_04’ indicates the 4th release of 2015 and was used for compiling the data and screenshots in this chapter. It is important that users take note of this release number when using FlyBase data to direct experiments, and quote it when referring to FlyBase data in publications. Summary statistics for the current release and a record of general changes to FlyBase in each release are provided as ‘Release Notes’ and ‘New in this release’ pages, respectively, accessible via the NavBar ‘About’ menu. If needed, previous FlyBase releases (and accompanying ‘Release Notes’) are available via the ‘Archives’ menu of the NavBar.

QuickSearch

QuickSearch, located at the heart of the homepage, is the primary search tool on FlyBase and can be used to access all data types (Fig. 1). It has been significantly improved in recent years to make it as intuitive and flexible as possible (2). The ‘Simple’ tab provides a Google™-type functionality in that any text can be entered and a search is performed across the entire database for possible matches. Using the ‘Data Class’ tab restricts the search to the specified class and, optionally, just to symbols, names and IDs (rather than ‘All text’).

Other QuickSearch tabs offer dedicated search options for particular data classes or associated data. The ‘References’ tab allows field-specific searching of the comprehensive set of Drosophila-specific publications in the FlyBase bibliography (3). Similarly, the ‘Human Disease’ and ‘Gene Groups’ tabs facilitate searching of these particular classes of integrated data (see section 9). Rather than searching by data class, the remaining QuickSearch tabs allow searching for entities that share Expression, Phenotype, Protein Domain or Gene Ontology (GO) annotations.

If a query in QuickSearch, or any other FlyBase tool, results in multiple possible matches, then a tabular ‘hit list’ is produced (4). The hit list serves as a core list-management tool in FlyBase and can be manipulated in several ways. The entire list, or a selected subset, can be: sorted by the entries in any of the columns; analyzed or refined by criteria appropriate to the given data; converted to a related data class (e.g., alleles converted to genes); downloaded as a list; or exported to other FlyBase tools for further refinements or to download specified data.

Jump to Gene

‘Jump to Gene’ is a navigation tool located in the NavBar and is thus available from all FlyBase pages. It is useful for quickly moving between Gene Reports where the current FlyBase gene symbol or ID (FBgn number) is known. (Greek characters must be spelled out, e.g., ‘α-Est1’ as ‘alpha-Est1’.) In fact, ‘Jump to Gene’ also accepts current FlyBase IDs for all data classes, thus providing a direct route to other reports of interest. Note that input not recognized as a valid symbol or ID is treated as a gene synonym and a best match to a gene will be attempted. As the best match may not be to the intended gene, ‘Jump to Gene’ should not be used as a query tool – the ‘Simple’ or ‘Data Class’ tabs of QuickSearch should be used when the current FlyBase symbol or ID is not known for certain or when conducting a search.

The Gene Report

The Gene Report is the best place to start for users interested in a particular gene (2) (Fig. 2). In addition to listing gene-centric data (e.g., genomic location, expression data, orthologs), this report also summarizes data more fully described in separate reports (e.g., mutant alleles, phenotypes of those alleles, protein interaction data) and provides many links both within FlyBase and to external sites where additional information or details can be found. As such, Gene Reports act as hubs from which to explore all that is known about Drosophila genes.

The upper ‘always open’ part of the Gene Report has recently been reorganized to emphasize functional information, including a new section labeled ‘Families, Domains and Molecular Function’ (Fig. 2). This section highlights membership of the gene to any FlyBase ‘gene groups’ (see section 9.2) or UniProt protein families, any predicted UniProt or InterPro domains/motifs, and summarized molecular function data based on FlyBase GO annotations. At the top of the report, the ‘General Information’ section includes gene identifiers (FlyBase ID, symbol, name and CG number), the ‘Feature type’ (e.g., ‘protein coding gene’, ‘miRNA gene’ or ‘pseudogene’), and the ‘Gene Model Status’, which helps to distinguish genes currently localized to the sequenced genome from those whose gene model is incomplete/withdrawn or genes defined only by mutations. Below this is a section summarizing ‘Genomic Location’ information. It gives the cytogenetic position and sequence coordinates of the gene, alongside a graphical snapshot of its genomic location and convenient options to download FASTA files associated with the gene model.

The remaining data are organized into separate sections that are closed by default and can be opened by clicking on their title bars (Fig. 2). Several of these have been updated in recent years. For example, the ‘Gene Ontology (GO)’ section that displays the full set of GO annotations (5) now clearly distinguishes terms based on experimental evidence from those based on predictions/assertions. The ‘Summaries’ section that follows has also been revised and now contains additional entries including a description of the FlyBase gene group to which the gene belongs and a functional summary imported from UniProt. Further down the page, the ‘Orthologs’ section now contains three discrete subsections. The first, labeled ‘OrthoDB Orthologs’, displays the orthologs of the given gene as computed by OrthoDB (6), arranged into taxonomic groups. The second subsection, ‘Human Orthologs’ specifically highlights the orthologous human gene(s) (again inferred by OrthoDB). Here, links are provided to the corresponding Ensembl (7), HUGO Gene Nomenclature Committee (HGNC (8)) and Online Mendelian Inheritance in Man® (OMIM® (9)) gene and phenotype (disease) reports. Finally, the ‘External Data’ subsection features link-outs to species-specific orthologs determined by the integrative ortholog prediction tool, DIOPT (10). Note that link-outs such as these appear in several sections of the Gene Report and are consolidated in the ‘External Cross-references and Linkouts’ section towards the bottom of the report, facilitating navigation between databases.

Several other Gene Report sections have seen more significant updates, including expression data, physical interactions, and models of human disease. These are described separately below.

Querying genes and gene-related data: Gene Reports themselves are best found via the ‘Jump to Gene’ box on the NavBar or via the ‘Simple’ or the ‘Data Class (genes)’ tabs of QuickSearch. GO and Protein Domain annotations can be searched via their dedicated QuickSearch tabs. Strategies for searching for data in other sections or fields of the Gene Report are detailed below, or may be addressed using QueryBuilder (11) or the ‘Simple’ tab of QuickSearch.

Alleles and Phenotypes

The generation and study of mutants have been central to Drosophila research ever since its very beginnings (12) and remain a major component of FlyBase today. These data are summarized on the relevant Gene Report in the ‘Alleles and Phenotypes’ section (Fig. 2), and are described in full in separate Allele Reports (not shown). FlyBase makes extensive use of controlled vocabularies (CVs or ‘ontologies’) in recording allelic and phenotypic data (13,14). These are collections of related terms (e.g., ‘allele class’, ‘phenotypic class’, ‘fly anatomy’) arranged into parent-child hierarchies. They allow a single, defined term to be used across all FlyBase annotations, which in turn allows users to search with that term (or any of its synonyms) to retrieve all records annotated with it or its children. CV terms and associated annotations can be queried using the Vocabularies tool (11) (formerly TermLink), accessible via the pictograph on the homepage or the Tools menu of the NavBar.

Classical alleles, transgenic constructs and insertions

Classical alleles are defined as mutations that affect a gene at its endogenous locus. They are presented in Allele Reports and also appear in the ‘Classical Alleles’ subsection on the Gene Report (Fig. 2). They traditionally include point mutations, insertional mutations and intragenic deletions, though more recent additions include lesions induced by various recombination-mediated techniques and site-specific cleavage events. Where known, mutations are annotated with an ‘origin of mutation’ term (e.g., ‘ethyl methanesulfonate’ or ‘CRISPR/Cas9’), an ‘allele class’ term (e.g., ‘amorphic allele’) and details of their molecular lesion. If the mutation is caused by an insertion of a transposable element, then both an Allele Report and an Insertion Report are created to completely describe the lesion, with a prominent link forged between them.

The molecular details and uses of transgenic constructs appear in dedicated Recombinant Construct Reports. In addition, an associated Allele Report is created in such cases in order to properly and fully capture phenotypic data. That is, the Alleles data class in FlyBase comprises both ‘classical alleles’ and ‘alleles carried on transgenic constructs’, as indicated within the ‘Alleles and Phenotypes’ section of the Gene Report (Fig. 2). The allele ‘origin of mutation’ CV has been expanded to accommodate this convention and so includes terms such as ‘in vitro construct - RNAi’. The relationship between constructs and their associated alleles is clearly indicated and reciprocally linked in their respective report pages. Similarly, any specific insertions of a transgenic construct are captured in Insertion Reports and are reciprocally linked to their corresponding Recombinant Construct Report.

Recombinant Construct or Insertion Reports, as appropriate, are also made for reporters (e.g., lacZ or GFP) or binary drivers (e.g., GAL4). Again, an associated Allele Report is made in all these cases so that phenotypic (and expression) data dependent on their use can be stored and presented in a consistent manner across the database. Note that FlyBase uses a species prefix to distinguish genes originating from ‘foreign’ (non-D. melanogaster) species, and so these examples appear as alleles of ‘Ecol\lacZ’, ‘Avic\GFP’ and ‘Scer\GAL4’ on the website.

Transgenic techniques and resources are constantly expanding (15). FlyBase responds to the former by devising suitable curation strategies and/or revising CVs as appropriate. One response to the latter is the creation of the Large Dataset Metadata Report that collates the metadata and membership of large-scale collections of constructs, insertions, etc. (see section 9.1). The report page of each member contains basic descriptive information about the collection, together with a link to the respective metadata report.

Querying alleles, constructs and insertions: Specific alleles, transgenic constructs or insertions are best searched via the ‘Data Class’ tab of QuickSearch. As mentioned, the Vocabularies tool is useful to find instances of particular allele classes or mutagenic techniques, which can then be refined further using other FlyBase tools. CytoSearch, FeatureMapper or GBrowse (see section 7) are the preferred methods if you are looking for reagents that are mapped to specific genomic regions.

Phenotypes

Phenotypic data are attached to alleles or allele combinations using terms from the ‘phenotypic class’ and ‘fly anatomy’ CVs. The phenotype CV comprises ∼190 terms that are commonly used to describe Drosophila phenotypes, such as ‘lethal’, ‘sterile’, ‘homeotic’ or ‘Minute’ (13). The anatomy CV is much larger, comprising >8,800 terms that can be used to comprehensively describe Drosophila anatomy (14). Both types of CV term can be refined through the use of ‘qualifier’ terms that restrict the meaning of the term to a specific developmental stage, sex or other experimental/genetic condition – these appear after a ‘pipe’ symbol on the website, for example ‘small body | larval stage’. Any additional genotypic components that are necessary for the given phenotype, such as alleles in trans or GAL4 drivers, are included in the phenotype annotation and appear with the prefix ‘with’ on the website. CV-based phenotypic annotations are often supplemented with free text clarifications or extra details. All phenotypic statements in FlyBase are curated from the published literature.

CV-based phenotype annotations are shown with their associated allele in a table on the Gene Report in the ‘Summary of Allele Phenotypes’ subsection (Fig. 2). Clicking on an individual allele takes you to the corresponding Allele Report that additionally contains any free text description of the phenotype together with the source reference(s).

Querying phenotypes: The dedicated ‘Phenotype’ tab in QuickSearch facilitates searching of alleles by phenotypic class and/or anatomy terms, with an option to refine the search through the use of qualifiers. The Vocabularies tool offers a browsable view of the same data, while QueryBuilder can be used to compose more complex combinatorial queries.

Disease model annotations

Drosophila alleles or allele combinations that generate phenotypes stated to be models of human disease are additionally annotated using appropriate terms from the Disease Ontology (16,17). As for regular phenotypes, disease model annotations may be associated with either classical alleles (of fly genes orthologous to human ‘disease genes’) or with transgenic alleles (where disease-causing forms of human genes, or their Drosophila orthologs, are expressed via transgenic constructs). The criterion for disease annotation in FlyBase is that the phenotype must recapitulate some aspect of the disease pathology, though this can range from anatomical defects or behavioral abnormalities to cellular or molecular changes. Additional alleles may be described as modifying a disease model, either ‘ameliorating’ or ‘exacerbating’ it, if genetic interactions are observed. This information is presented in tabulated form in both the Gene Report and Allele Report in the section titled ‘Human Disease Model Data’ as well as in the new Human Disease Model Report (see section 9.3).

Querying disease model annotations: The ‘Human Disease’ tab of QuickSearch and the Vocabularies tool can both be used to find a Disease Ontology term and view records annotated with it.

Expresson Data

Separate Gene Expression Reports have been retired from FlyBase in favor of integrating these data into the Gene Report in a dedicated ‘Expression Data’ section (Fig. 2). Expression data may derive from either ‘low-’ or ‘high-throughput’ studies.

Low-throughput expression data

Traditional expression assays, such as in situ hybridization or immunolocalization to embryos or tissues, and stage- or tissue-specific Northern blots, are presented in the first three subsections of the ‘Expression Data’ section: ‘Transcript Expression’, ‘Polypeptide Expression’, and ‘Expression Deduced from Reporters’. These are data from the published literature, captured in a highly controlled format using the FlyBase anatomy and developmental stage CVs. For nuanced aspects of an expression pattern additional free-text descriptions are provided. Embryonic transcript expression data include data from the BDGP (18), which comprises in situ hybridizations using cDNA probes for over 7,000 genes, plus descriptions conforming to the FlyBase anatomy CV. The actual in situ images can be viewed by following the ‘BDGP expression data’ link in the subsection ‘External Data and Images’ (see section 5.3).

Querying low-throughput expression data: The use of hierarchical CVs allows expression data captured at a very detailed level to be queried using more general terms. The QuickSearch ‘Expression’ tab provides a dedicated interface for this type of query. It also allows combinatorial queries, typically to specify both stage and tissue. The initial hit list returned by this query is of endogenous genes for which the expression pattern is observed; reporter constructs or insertions can be retrieved by selecting one of the alternative result options at the top of the initial hit list.

High-throughput expression data

Within the ‘Expression Data’ section of the Gene Report, the ‘High-Throughput Expression Data’ subsection includes expression plots of high-throughput mRNA RNA-Seq data from modENCODE (19) and mRNA microarray data from FlyAtlas (20). FlyBase has produced quantitative views of these data in different stages, tissues, and cell culture types presented as bar graphs (Fig. 3). For the RNA-Seq data, RPKM counts (21) (reads per kilobase per million reads) have been calculated, averaged over the exonic extents of the gene. A further subsection, ‘Expression Clusters’, provides links to datasets consisting of genes possessing similar mRNA RNA-Seq expression dynamics, as determined by modENCODE (22,19).

Querying high-throughput expression data: Several new FlyBase tools that use the calculated RPKM data have been developed. The versatile ‘RNA-Seq Profile’ tool allows retrieval of genes with user-defined RNA-Seq expression patterns and levels (values are binned); this tool can be accessed from a pictograph on the homepage, from the Tools menu of the NavBar, or from the QuickSearch ‘Expression’ tab. A section of the QuickSearch ‘Expression’ tab also provides options to ‘Search for similarly expressed genes’, using the modENCODE RNA-Seq datasets. The ‘RNA-Seq by Region’ tool returns the average RPKM over a specified genomic region and also offers the option of a gene-specific query that returns an exon-by-exon RPKM count; this tool can be accessed from the Tools menu or from the ‘High-Throughput Expression Data’ subsection of the Gene Report.

External link-outs

The final ‘Expression Data’ subsection on the Gene Report is ‘External Data and Images’, with gene-specific links to other databases that include expression data for Drosophila. There are links to the original BDGP in situ data, as well as FlyExpress (23) analyses that use the BDGP data and allow a ‘Find Similar Patterns’ option. The FlyAtlas link provides the underlying microarray data used for the bar graphs described above. For the FlyExpress and SliceSeq (24) databases, sample images are shown.

Interactions

Both genetic and physical interaction data are presented in FlyBase. The former are primarily recorded in Allele Reports, while the latter are given in dedicated Physical Interaction Reports; both are summarized in the ‘Interactions and Pathways’ section of the relevant Gene Report page. Both types of interaction can be viewed either as a graphical ‘network diagram’ provided by esyN (25) or within the FlyBase Interactions Browser tool that includes additional viewing and configuration options (4). The ‘External Data’ subsection of the ‘Interactions and Pathways’ part of the Gene Report provides link-outs to relevant pages at third-party interaction databases, including BioGRID (26), DroID (27) and InterologFinder (28).

Genetic interactions

Genetic interaction data are recorded at the allele level using phenotypic class and anatomy CV terms (and optional qualifiers), similar to phenotype annotations but with the addition of terms such as ‘enhanced by’ or ‘suppressor of’ to indicate the nature of the interaction, together with the interacting allele. (Negative results, e.g., ‘not enhanced by’, are also captured.) In addition to enhancer/suppressor-type interactions, synthetic phenotypes that are present in a mutant combination but absent in single mutant conditions are also captured. All these interaction statements appear in the ‘Interactions’ section of the Allele Report, alongside free text clarifications where necessary. All genetic interaction statements in FlyBase are curated from the published literature.

The allele-level genetic interaction statements are used to compute a gene-level summary of these data, and this appears as a table within the ‘Summary of Genetic Interactions’ subsection in the Gene Report. This table shows the interacting genes, the nature of the interaction (limited to enhancer/suppressor-type interactions) and the reference(s) supporting the interaction. These gene-level interactions are used to power the esyN network diagram.

Querying genetic interactions: Both allele-level and gene-level genetic interaction data can be queried directly via the Interactions Browser tool. More specific and/or combinatorial searches may be conducted using QueryBuilder.

Physical interactions

The Physical Interaction Report displays pairwise physical interaction data for gene products, either protein-protein or RNA-protein. Each report includes the experimental assays used (e.g., co-immunoprecipitation, peptide mass fingerprinting), the role of each protein in an assay (e.g., bait or prey; whether a tagged or endogenous protein was used), the esyN network diagram, and a link to the Interactions Browser tool. With the goal of producing a set of high-confidence pairwise interactions, our current focus is on smaller-scale physical interaction data curated from the literature, which usually include multiple types of support for a described interaction. High-throughput interaction datasets are curated only when the authors take care to filter out false positives--criteria for curation may include: multiple negative control purifications, accounting for protein abundance in assessing the likelihood that a purified factor is a contaminant, a calculation of the confidence level, and an explicit cut-off to separate high confidence and lower confidence interactions. Examples are the DPiM dataset (29), the Hippo Pathway Interactome (30) and the ECIA extracellular interactome (31).

Within the ‘Summary of Physical Interactions’ subsection of the Gene Report, all pair-wise physical interactions involving that gene product are presented in tabulated form, with assays used, attributed publications, and links to the corresponding Physical Interaction Report.

Querying physical interactions: The ‘Simple’ or the ‘Data Class (physical interactions)’ tabs of QuickSearch can be used to find interactions involving a given gene or to search for assay terms present in the Physical Interaction Report.

Genomic Data

Genomic data in FlyBase comprise gene model annotations (i.e., the exon-intron structure and transcription and translation start/termination sites of genes) and any other sequence-based features that can be mapped to specific genomic coordinates, whether endogenous (e.g., regulatory regions, origins of replication) or describing a lesion/reagent (e.g., insertion sites, RNAi amplicons). All these data are viewable through the FlyBase implementation of GBrowse (32) (Fig. 4). Many are also associated with discrete Sequence Feature Reports and are searchable through the FeatureMapper tool (11) (Fig. 5). Note that FlyBase currently uses release 6 of the sequenced D. melanogaster genome (33) – the Coordinates Converter tool, accessible from the Tools menu of the NavBar, can be used to convert data from release 3, 4 or 5 coordinates (32).

Gene model annotations in FlyBase

For D. melanogaster, FlyBase has produced manually annotated gene models for over a decade (34). Since 2010, RNA-Seq data (35,19) and new transcription start site data (33,36) have supported many major changes in the gene model annotations for this species. This prompted a comprehensive review of all existing gene models and the annotation of several thousand new genes, primarily long non-coding RNA genes (34). Transcript and protein data are tabulated in the ‘Gene Model and Products’ section of the Gene Report (Fig. 2), with links therein to more detailed reports. D. melanogaster gene models continue to be updated regularly based on new high-throughput and literature-based data. An updated gene model set is submitted to GenBank approximately once a year and serves as the NCBI RefSeq set for this species.

For eight of the other sequenced Drosophila species (D. ananassae, D. erecta, D. pseudoobscura pseudoobscura, D. simulans, D. yakuba, D. mojavensis, D. virilis and D. willistoni), the long-standing CAF1-generated gene model annotations (37) have been replaced recently by sets generated by NCBI as part of their GNOMON annotation pipeline (38). Gene model annotations for three other species (D. grimshawi, D. persimilis and D. sechellia) have not been updated owing to poor genome assembly quality or to lack of RNA-Seq data, which provides the primary basis for robust annotation by the GNOMON pipeline.

Querying gene model data: Gene models can be searched directly in GBrowse (section 7.2) or via the CytoSearch or FeatureMapper tools (both accessible via the Tools menu of the NavBar). The FlyBase BLAST tool allows sequence-based queries against annotated transcripts or proteins from the twelve Drosophila species mentioned above. For D. melanogaster, Sequence Ontology terms and controlled comments have been used extensively to describe gene models and transcripts (34,39). These enable queries for exceptional cases, such as all genes with dicistronic transcripts or all transcripts annotated with non-canonical translation start sites, using QueryBuilder or the ‘Simple’ tab of QuickSearch. ‘Transcripts’ and ‘Polypeptides’ are data class options in QuickSearch and QueryBuilder, thus allowing class-specific and field-specific queries of gene products, respectively.

GBrowse

GBrowse, a genome annotation viewer that is part of the Generic Model Organism Database (GMOD) tool suite (40), is used by FlyBase to show gene models and supporting data, such as cDNAs, ESTs, RNA-Seq data, transcription start sites, gene predictions, and aligned proteins (32) (Fig. 4). In addition, this versatile tool allows representation of many other types of sequenced-based data and reagents--essentially anything that maps to the genome can be represented on GBrowse. By using the ‘Select Tracks’ option, the user can choose to view mapped genetic variants such as mutational lesions, transgenic insertions, aberration extents, and aberration breakpoints; regions carried on transgenic constructs such as rescue fragments and RNAi reagents; microarray oligonucleotides and RNAi amplicons; or high-throughput mapping of transcription-factor binding sites, insulator elements, and RNA-editing sites. When zoomed in to a range of 100–200bp, the tracks indicating forward/reverse translation and ‘DNA/GC content’ switch to the nucleotide or protein sequence. The current version of this tool is GBrowse2 (41), which allows rapid customization options: for example, a selected track can be moved by simply dragging the track title bar vertically, and tracks can be closed, opened or removed using the icons in the track title bar. Moreover, navigation within a genomic region has been facilitated by limited smooth-track panning (side-to-side sliding) and by a function that allows the user to lasso a smaller region and zoom in.

RNA-Seq expression data (19,42) are particularly informative when viewed in GBrowse. In 2010, FlyBase debuted a new topographical presentation of these data for GBrowse that allows visual assessment of many RNA-Seq tracks at once (Fig. 4). By clicking on the wrench/spanner icon in the track title, the presentation can be changed from log2 to linear, and from tilted to vertical; specific tracks corresponding to different tissues and/or development stages can be shown or hidden. RNA-Seq exon junction data (19,35), presented in a separate track on GBrowse, are extremely useful for judging alternative splicing and isoform-specific expression.

GBrowse supports a number of download options, accessible from the drop-down menu on the upper right of the page, including a FASTA file of the sequence shown and an HTML table view or a GFF file of all the mapped genes and features selected. The sequence of a lassoed genomic region can also be viewed and downloaded.

GBrowse can be accessed from one of the pictograph buttons at the top of the homepage or via the Tools menu on the NavBar. In addition, there is a link to the appropriate genomic region in GBrowse on the reports for every localized gene and mapped sequence feature. A genomic BLAST hit obtained using the FlyBase BLAST tool also includes a link to the relevant region in GBrowse.

Sequence features and other genomic data tracks

‘Sequence features’ are defined as regions of DNA/RNA that can be mapped to the genome sequence and to which a discrete function can usually be ascribed. They include endogenous regions such as enhancers, insulators, transcription factor binding sites, transcription start sites and origins of replication, as well as experimental reagents that map to the genome, such as RNAi reagents and putative enhancer element constructs. Sequence features appear within discrete tracks on GBrowse and are associated with dedicated report pages. (Note that certain GBrowse tracks, including point mutations, transgenic insertions and aberration extents, are not classed as ‘sequence features’ and are instead associated with specific Allele, Insertion or Aberration Reports.) Most sequence features currently in FlyBase were generated in response to the modENCODE project (22) and similar large-scale experiments (43,44).

The Sequence Feature Report is flexible, in order to accommodate many different types of genome-associated data. The typical report includes a link to the Large Dataset Metadata Report (see section 9.1) to which it belongs, the sequence itself and its genomic location, a genome snapshot showing the alignment of that feature alongside other sequence features included within that region, and links to any relevant external websites/databases. Clicking on the ‘GBrowse’ link near the top of the page goes to a full genome view of the respective region in GBrowse.

Querying sequence features and other genomic data tracks: Limited querying can be performed within GBrowse itself by specifying a ‘Landmark or Region’ and selecting particular tracks for display. A better approach is to use FeatureMapper (Fig. 5), which provides an intuitive interface for retrieval of specified genome features in one or more genomic regions, with results presented in a convenient table that includes an option to export to a hit list where possible. The CytoSearch tool allows retrieval of genes, aberrations, and transgenic insertions mapped to the genomic sequence. Sequence features are also included as a specific option in the ‘Data Class’ tab of QuickSearch.

Reagents

There are several ways to find reagents associated with a specific gene or genomic region. The ‘Stocks and Reagents’ section of the Gene Report is a good place to start. Here, subsections list publicly available fly stocks, genomic and cDNA clones, cell-based RNAi reagents and antibodies described in the published literature (Fig. 2). Other reagents are best found by searching a genomic region of interest using GBrowse, FeatureMapper or CytoSearch. For example, the GMR (45) and VDRC (44) putative enhancer collections are not associated with specific genes, while some classes of transgenic insertions are not listed in the Gene Report. Moreover, a visual representation of the location of a sequence-based reagent relative to the gene of interest is often informative when planning experiments.

Stocks

Stock Reports display the stock list genotype and the source collection, together with the stock number hyperlinked to the specific record at the appropriate stock center to facilitate ordering. There are links to Stock Reports from other appropriate reports (primarily alleles, aberrations, transgenic constructs and insertions) throughout FlyBase. The Bloomington Drosophila Stock Center is the most widely represented source, though many others are included--a complete list can be found in the ‘Links’ menu on the NavBar.

Querying stocks: Stocks can be searched specifically by selecting ‘stocks’ in the ‘Data Class’ tab of QuickSearch.

Strains

FlyBase Strain Reports contain data about wild type strains such as ‘Oregon-R’, significant mutant strains such as ‘iso-1’ (the D. melanogaster strain sequenced by the BDGP (33)), as well as the 200 or so inbred lines generated by the Drosophila Genetics Reference Panel (46). The reports include information on the origin and history of the strain alongside any known genetic or phenotypic components (e.g., the ‘iso-1’ strain harbors several mutations). Where relevant, links are also provided to Large Dataset Metadata Reports (section 9.1) that describe strain collections, and to Stock Reports to facilitate ordering. (Note that stocks are instances of strains in theory, but they are effectively distinct in time and place and may have characteristics that differ from the strains from which they descended.)

Querying strains: Strains can be searched using the QuickSearch ‘Simple’ tab.

Cell lines

Cell Line Reports display data obtained from the Drosophila Genomics Resource Center (DGRC) on cell lines, such as ‘Kc167’ or ‘S2R+’. The reports include the source and development stage of each line, its sex and karyotype (where known), and any parental or descendent lines. A link back to the DGRC is also provided for additional data and ordering information.

Querying cell lines: Cell lines can be searched specifically by selecting ‘cell lines’ in the ‘Data Class’ tab of QuickSearch.

cDNAs

cDNAs are shown in GBrowse and appear in the ‘Stocks and Reagents’ section of the Gene Report of the aligned gene(s). Links from GBrowse go to the GenBank report; links from the Gene Report go to the FlyBase Clone Report. The Clone Report includes the sequence, links to GenBank, and fields for ‘Known Problems’ and ‘FlyBase assessment’. Examples of known problems are clones that are chimeric or that contain genomic DNA or transposon sequences. The FlyBase assessment field displays a note if the clone has been replaced, for example “Caution: This cDNA clone replaced by FI01005”. There is also a link to the DGRC where clones are available from that resource.

Querying cDNAs: cDNA clones can be searched specifically by selecting ‘clones’ in the ‘Data Class’ tab of QuickSearch. FeatureMapper should be used to find cDNAs associated with a specific gene or genomic region.

Integrated Reports

As the amount of Drosophila data and resources increase in FlyBase, it has become both necessary and useful to organize and integrate related data into discrete sets or collections. This has multiple benefits, including the ability to associate metadata across a range of related entities, and to present related data to users in new ways that aid comprehension. To date, FlyBase has developed three types of such integrated reports.

Large Dataset Metadata

Large Dataset Metadata Reports, previously named Library/Collection Reports, provide information on large datasets and reagent collections that apply to the set as a whole. Examples of datasets are the protein interaction network defined by the Drosophila Protein interaction Mapping (DPiM) project (29), the set of RAMPAGE transcription start sites (36), and datasets generated by the modENCODE project (22). Examples of collections are the set of dsRNA amplicons used for RNAi-knockdown assays in cell culture by the Drosophila RNAi Screening Center (49), the set of defined X-chromosome duplications made by the Bloomington Stock Center (50), and several large construct and insertion collections. Metadata describing cDNA libraries are also captured in this format. The Large Dataset Metadata Report includes the type of dataset or collection, a brief description of the set, a summary of the experimental details, and a link to download all the associated features. Links to external data repositories and reagent sources are provided where relevant. The ‘Description’ field of the dataset report is propagated to each member report; reciprocal links are provided.

Querying large dataset metadata: The ‘Simple’ or the ‘Data Class (large dataset metadata)’ tabs of QuickSearch can be used to find datasets and collections of interest.

Gene Groups

Gene Group Reports have been introduced to allow easy access to, and analysis of, related sets of D. melanogaster genes and their associated data (47). Examples of gene groups include members of a gene family (Actins, Wnts…), subunits of a protein complex (proteasome, ribosome…), or other functional groupings (protein kinases, Ubiquitin E3 ligases…). All gene groups in FlyBase are based on published literature and the basis for the membership of each group is clearly attributed. The main feature of these reports is a ‘Members’ table that lists the genes comprising the group, arranged into a series of subgroups where appropriate. Buttons are provided to facilitate the downloading of associated data (phenotypes, expression data, protein interactions etc.) using Batch Download (section 10.2), or to further refine or analyze the gene set by exporting it to a standard hit list. Also shown are links to equivalent gene groups for other organisms, including nematodes (WormBase (48)) and humans (HGNC (8)). To aid navigation, the ‘Families, Domains and Molecular Function’ section of the Gene Report contains a link to any associated gene group(s) (Fig. 2).

Querying gene groups: Gene groups can be retrieved by entering the symbol/name of a group or any member gene in the ‘Gene Groups’ tab of QuickSearch. This tab also includes a link to a browsable list of all current gene groups in FlyBase.

Human Disease Models

Human Disease Model Reports provide a less specialized entry point into FlyBase for researchers interested in Drosophila models of human disease (17). Data from numerous outside sources, including OMIM, and from recent reviews are presented in a general ‘Disease Summary’ section, followed by information on orthology between a human gene implicated in the disease and the related Drosophila gene(s). For many diseases, multiple causative genes have been implicated; OMIM describes these as different disease subtypes and groups them into ‘phenotypic series’. In the Human Disease Model Report, such a phenotypic series of subtypes is presented in a table titled ‘Related Diseases’, which includes links to other relevant Human Disease Model Reports and provides a quick view of which disease subtypes have been modeled in flies.

The major portion of the disease report is devoted to ‘Experimental Findings’ in Drosophila, focusing on disease-related implications and results. Descriptions of specific experiments are meant to be generally accessible, with links to Allele Reports with more detailed information. Results may include data using both fly genes and human genes introduced into flies. The ‘Experimental Findings’ section initiates with a FlyBase-authored summary that presents a concise review, including phenotypes, interactions, and suitability of the model for drug assays; in addition, new findings and emerging mechanistic themes are highlighted. At the end of this section, a link to the FlyBase Disease Wiki is provided; comments and contributions from users are encouraged, especially those with expertise in the specific disease model. The last sections of the report draw relevant data from other sections of FlyBase, including physical interaction data for the orthologous Drosophila gene(s), a table of genetic reagents and stocks useful for investigations of human disease, and a table of Disease Ontology-based annotations of alleles used for that disease model (see section 4.3).

There are links to relevant Human Disease Model Reports in the ‘Human Disease Model Data’ section of Gene Reports (Fig. 2). Note that many such links are found in FlyBase Gene Reports for human genes (e.g., Hsap\SNCA and Hsap\TARDBP).

Querying Human Disease Model Reports: These reports can be found by using the ‘Human Disease’ or ‘Simple’ tabs of QuickSearch, or by searching the Disease Ontology within the Vocabularies tool.

Bulk Data Analysis and Downloads

Users increasingly want to be able to process data in bulk. They may have generated a hit list of genes (or any other data class) within FlyBase, or have a list of IDs from elsewhere to upload, and wish to analyze/refine this list or obtain associated data. Alternatively, users may wish to directly obtain bulk data files corresponding to a particular data type for processing off-line.

Uploading and analyzing data

A list of IDs (e.g., gene symbols or CG numbers, allele or insertion symbols, FlyBase identifiers) can be pasted or uploaded into the Upload/Convert IDs tool (Fig. 6; accessed via the Tools menu on the NavBar). This tool will then validate the list, updating any obsolete IDs to the current version where possible, and generate a ‘Conversion report’ clearly indicating if any of the submitted IDs failed verification. The user can choose to correct these cases, or ignore them before proceeding to convert the list into a standard FlyBase hit list (see section 2.2). This list can then be further analyzed/refined before being exported or downloaded as required.

Downloading data

Batch Download is a powerful tool for generating customized output files in various formats for most data types in FlyBase (11). Users may arrive at Batch Download via a hit list (as described above), by navigating to it from the Tools menu of the NavBar, or by clicking on its pictograph on the homepage. If the first, then the input list will be pre-filled (Fig. 6); otherwise the user can paste in or upload a list of symbols or IDs directly. Depending on the nature of the input and the desired outcome, the output format can then be specified as ‘FASTA Sequence’ (with the option to further specify introns, UTRs, CDS etc.), ‘Database Format’ (XML), or as ‘Field Data’ (with output options of an HTML table, a tab-separated value (tsv) file, or in the same format as the precomputed files described below). If the ‘Field Data’ option is selected, the user can then specify any combination of data fields (appropriate to the given data class) from a page styled in the same format as a standard FlyBase report page.

Bulk files of FlyBase data can be downloaded using our FTP site (ftp://ftp.flybase.org/releases/) or the ‘Downloads’ menu of the NavBar on the website (see the ‘Overview’ page under the Downloads menu for more details). ‘Precomputed files’ contain particular slices of FlyBase data that users or collaborators have requested over the years or are otherwise difficult to obtain in bulk (Table 3). Notable recent additions include D. melanogaster unique protein isoforms, RPKM gene expression values, gene groups, and physical interactions. Also included are several useful correspondence tables and the ontology files used in FlyBase (Table 3). In addition, Chado XML (database format) files are provided for all FlyBase data classes and comprehensive sets of FASTA, GFF and GTF files are available for the twelve originally sequenced and annotated Drosophila species (see section 7.1). The FASTA files comprise many different cuts of genomic data, including annotation categories such as small RNA classes and pseudogenes, components of gene model annotations such as exons, introns, UTRs and predicted translations, as well as other genome features such as transposons and intergenic sequences. As described above, Batch Download can also be used to obtain specified subsets of data in precomputed file, Chado XML or FASTA format by selecting the appropriate output options.

Most bulk files are regenerated for every release of FlyBase. Those corresponding to the current or previous (archived) versions of FlyBase are found under the appropriate submenus/subfolders on the web/FTP site. The release version used for a particular file is indicated in the file name and in the header lines of the file itself.

The FlyBase Community

FlyBase engages with our user community through multiple approaches. The primary method for users to get in touch with FlyBase about any matter remains our ‘Contact FlyBase’ page, accessible via the ‘Help’ menu on the NavBar or the link in the footer of any FlyBase page. All other community resources are grouped under the ‘Community’ menu of the NavBar and/or are found on the homepage.

If a user wants to specifically alert us to a Drosophila publication or data therein to be added to FlyBase, then the ‘Fast Track Your Paper’ (FTYP) tool should be used (51). This tool allows the user to indicate the key genes studied and flag data types present in a paper. The resulting gene-to-publication links are submitted directly to the FlyBase database while the data type information is used to prioritize the paper for more detailed curation. We actively solicit FTYP submissions using our ‘EmailAuthor’ pipeline, whereby the corresponding author of a Drosophila publication is automatically sent an email that includes a link to a personalized FTYP form (51). Approximately 50% of authors respond to this request, thereby reducing by half the amount of manual triaging to be done by FlyBase curators.

Our recently launched ‘FlyBase Community Advisory Group’ (FCAG) is a worldwide group of over 500 volunteers (lab heads, postdocs, students, technicians) who use FlyBase for a range of purposes. We contact this group up to six times per year with a survey on a variety of subjects to get feedback about how data collection, presentation and searching on FlyBase can be improved. By consulting this relatively large, diverse group of researchers, we hope to implement changes to FlyBase that are helpful for the greatest number of people.

Users may also help improve FlyBase by contributing to the Human Disease Wiki (described in section 9.3) or the FlyGene Wiki. There is a link to the latter at the top and within the ‘Summaries’ section of each Gene Report. This is pre-seeded with the automatically-generated FlyBase summary and users are encouraged to modify or add to this text to build up a more complete and readable summary of each gene’s main features and functions.

The FlyBase Forum is a Google™ Group that provides an alternative, more open platform for users to interact both with FlyBase and with each other. The forum has two areas: one for general questions and discussions about FlyBase and Drosophila protocols etc., and the other for relevant job postings.

Users are made aware of new or changed features in FlyBase through any of several means. First, there are the ‘News’ and ‘Commentary’ sections of the FlyBase homepage (Fig. 1). Second, users can sign up to receive an occasional Newsletter via email by clicking the link on the homepage. The Newsletter contains release announcements, significant website updates, and other important Drosophila community news. Third, to obtain more frequent updates, users can follow FlyBase on Twitter™ by clicking on the icon in the footer of any FlyBase page. Fourth, users can choose to subscribe to any FlyBase record (a specific gene, transgenic construct, reference etc.) and receive automatic updates through a feed reader by clicking the icon in the ‘Recent Updates’ section of any report page. Finally, users have the opportunity to see and hear about FlyBase updates in person at the Annual Drosophila Research Conference in the USA and the biennial European Drosophila Research Conference, where FlyBase representatives give presentations and are available to answer questions. Previous conference presentations and pamphlets can be obtained via the ‘FlyBase Guides’ link under the ‘Help’ menu in the NavBar.