Genome Database for Vaccinium: a knowledgebase for genetics, genomics, and breeding research

Abstract

The Genome Database for Vaccinium (GDV, https://www.vaccinium.org/) is a knowledgebase serving the global Vaccinium research community. GDV centralizes genetics, genomics, and breeding data for blueberry, cranberry, bilberry, and lingonberry. Launched in 2011, GDV transforms research data into an integrated resource through data curation, standardization, and value-added analyses that enable cross-study comparisons and knowledge synthesis. GDV houses diverse data types including germplasm records, genetic markers, linkage maps, QTL/GWAS results, traits, genome assemblies, gene annotations, and expression profiles. The database team perform comprehensive analyses including homology searches against protein databases, InterPro domain identification, and synteny analyses between Vaccinium genomes to identify orthologous relationships. These analyses facilitate knowledge transfer within and between species, enabling researchers to leverage findings across the Vaccinium genus. Users access data through intuitive search interfaces and specialized tools for genetic map visualization, genome browsing, metabolic pathway exploration, and sequence similarity searches. The integrated Breeding Information Management System (BIMS) enables breeders to manage private breeding programs as well as accessing public breeding data. Currently containing 46 genome assemblies, over 287,000 genetic markers, 4,328 QTL, and 3.3 million gene models, GDV continues expanding through collaborations with the VacCAP project and international research groups. The database evolves based on community needs and emerging data types, ensuring its continued relevance for advancing Vaccinium research and crop improvement efforts worldwide.

Introduction

Biological research has entered a big-data era driven by experiments conducted at “omics” scales, generating extensive datasets with thousands of usable data points. However, researchers often struggle to access and search these large datasets effectively, making crop-specific knowledgebases like the Genome Database for Vaccinium (GDV, https://www.vaccinium.org/) essential (Humann et al. 2023). As experimental results and genomic resources continue to grow, centralizing relevant data becomes crucial for translational science and crop improvement. The ability to link trait and marker data to specific genome positions enables researchers and breeders to pursue more efficient crop improvement strategies in this genomics-driven era.

GDV is 1 of 5 crop databases developed by the MainLab program at Washington State University that share common infrastructure, search interfaces, and tools as previously described (Jung et al. 2025). These knowledgebases utilize Tripal v3.3 (Spoor et al. 2019), which provides a Drupal v7.1 interface for data stored in the Chado v1.2 database schema (Jung et al. 2011). Among these resources, CottonGen (https://www.cottongen.org/) (Yu et al. 2021) and the Genome Database for Rosaceae (https://www.rosaceae.org/) (Jung et al. 2019) are the most comprehensive and established, with widespread adoption in their respective research communities. The Citrus Genome Database (https://www.citrusgenomedb.org/) and Pulse Crop Database (https://www.pulsedb.org/) continue expanding and strengthening connections within their communities. The strength of these 5 knowledgebases being managed by a single research program is that the analysis pipelines, data curation standards, and tools to search and view data are implemented across all the sites yet still allow the flexibility for customization of which tools and data are available to meet each crop community's needs. GDV focuses on data from only the Vaccinium species and supports a smaller research community than the other 4 databases. User accounts are not needed to access the public data on GDV and BIMS. However, user accounts are required to maintain BIMS breeding programs which contain private data. GDV is the only worldwide, Vaccinium-focused knowledgebase that integrates genetic, genomic, and breeding data into a single database enabling queries and visualization across data types.

The USDA NIFA and Vaccinium industry-supported VacCAP project (https://www.vacciniumcap.org/) has been instrumental in GDV's development over the past 6 yr. VacCAP has contributed 88% of the genomes, 64% of the markers, 13% of the maps, and 85% of the QTL currently in GDV, while providing valuable insights into Vaccinium community needs that guide database development priorities. These contributions from VacCAP, combined with data from other published studies, have established GDV as a comprehensive worldwide resource for Vaccinium researchers and breeders.

Data types and curation

GDV covers 4 major Vaccinium crops: blueberry, cranberry, bilberry, and lingonberry. Trained GDV curators process and integrate diverse genetic, genomic, and breeding data from peer-reviewed publications through manual curation ensuring data quality and consistency. As more genomes and datasets have been added, usage has steadily grown with pages accessed now exceeding 300,000 per year and users from over 120 countries (https://www.vaccinium.org/gdv_usage). By August 2025, more than 130 peer-reviewed publications have cited use of GDV. While GDV was created in 2011, most of the data growth has occurred in the last 5 yr (Table 1).

Table 1.

GDV data growth since 2017. The table summarizes the increase in data volume over time, including markers, maps, QTL, GWAS, genomes, and other curated datasets.

Year	Genomes	Genes	mRNA	Maps	Markers	QTL	GWAS
2017	0	0	1,157	10	4,253	26	NA
2018	1	36,840	37,997	10	4,253	26	NA
2019	2	165,399	166,577	25	15,979	403	NA
2020	4	245,382	253,280	31	89,951	403	NA
2021	8	466,354	476,613	34	136,212	465	NA
2022	42	3,193,094	3,235,659	38	245,127	816	NA
2023	42	3,193,094	3,235,659	39	247,457	836	308
2024	45	3,288,063	3,404,332	39	269,522	4,153	320

Genetic data

Genetic data types include germplasm records (572), genetic markers (287,835), genetic maps (39), quantitative trait loci (4,328), and traits (261) as detailed in Table 2. Notable marker collections include the blueberry 22K genotyping array and the 17K cranberry genotyping array (Clare et al. 2024, 2025), with genome positions of all probes stored in the database. All information from the same publication is linked by internal identification numbers, and most data types require minimum metadata including population information for genetic maps and trait measurement details and analysis statistics for QTL.

Table 2.

Data types in GDV and current count.

Data type	Count
Genetic data
Germplasm	572
Genetic markers	287,835
Genetic maps	39
QTL	4,328
Traits	261
Phenotypic measurements	1,381
Genomic data
GWAS loci	1,024
Blueberry genomes	28
Cranberry genomes	15
Bilberry genome	1
Lingonberry genomes	2
Genes	3,312,676
mRNA	3,428,085

Genomic data

Genomic data consists of genome sequences, predicted gene/mRNA models, and associated data such as gene expression and SNPs. The database contains 46 genome assemblies across all 4 crops and 3.3 million gene models (Table 2). Since 2023, GDV has added genome positions for trait loci identified via GWAS analyses and in 2025 adopted the genome naming convention proposed by the AgBioData Consortium (Cannon et al. 2025). We recently started to manually curate more detailed information about gene annotations from the literature. The new data originates from studies that do further analyses to confirm the predicted gene function. We have developed curation evidence codes for the types of experiments conducted to support the gene symbol and function, and this information is linked to the publication in each gene record. This information is displayed on the gene page and is also searchable through the Gene Search. The Vaccinium Pangenome Project data (Yocca et al. 2023) is available for download and viewing in JBrowse2 and the UCSC Genome Browser.

Tools and analytical capabilities

GDV integrates customizable search interfaces with specialized Tripal modules (Jung et al. 2017, 2021a, 2025) and established bioinformatics tools:

• MapViewer (Buble et al. 2019): Interactive visualization and comparison of genetic maps, linkage group comparison, and genome visualization with SNP markers from GWAS studies. Allows users to customize the view and colors of features on the linkage group and select which feature types are displayed. Enables users to export figures.

• JBrowse/JBrowse2 (Buels et al. 2016; Diesh et al. 2023): Genome browsing with enhanced track options including pair-wise synteny tracks and pangenome alignment visualization. Features can be located by coordinates or by name.

• BLAST (Camacho et al. 2009): Sequence similarity searches against curated databases. Users can query nucleotide or protein sequences against GDV genomes.

• PathwayCyc (Karp et al. 2021): Metabolic pathway exploration and visualization tool. Allows users to overlay RNA-seq or metabolomics data onto a cellular overview.

• Synteny Viewer (https://github.com/tripal/tripal_synview): Visualization of syntenic blocks and orthologs from pair-wise genome comparisons. The stored ortholog data can be searched and downloaded using GDV search tools.

• Expression Heatmap Tool (Chen et al. 2017): Displays transcript expression levels using analysis values from published datasets. Users can view and download the data from the gene page or the heatmap visualization page.

GDV data are collected from published manuscripts and the processing pipeline depends on the data type (Fig. 1). Genetic data such as markers, maps, QTL, and GWAS are manually curated, and relevant metadata are stored for each data type. Curators map relevant metadata with established ontologies and standardize the naming of maps, QTL, and markers within the database. For RNA-seq datasets, the biomaterial metadata is curated and uploaded along with the expression values (for example counts, FPKM) provided by the source publication. All data loading scripts include multiple error-checking to ensure that data names are unique and that required metadata are provided. Once loaded, these data types can be queried using MegaSearch (Jung et al. 2021a) or other Tripal visualization extensions.

Fig. 1.

Overview of GDV data curation and analysis pipelines. Most data are manually curated by trained personnel who enter data and metadata from published manuscripts into templates and link available ontology terms. Data are uploaded to the database and made accessible through search tools and visualization interfaces. Genomic data are reviewed by curators, checked for file format and quality using in-house scripts, analyzed with bioinformatic tools, and stored with functional annotation results. Genome files are formatted for compatibility with GDV tools. Icons from Microsoft Powerpoint.

Genome data are processed through a separate pipeline (Fig. 1). Genome assembly files are validated to ensure proper formatting and the presence of required features. Gene models are functionally annotated, and the resulting annotations are uploaded into GDV where they are accessible through MegaSearch and Tripal visualization tools. Genome files are also formatted and/or analyzed for compatibility with JBrowse, PathwayCyc, and BLAST tools.

Value-added analyses

GDV curators perform several in silico analyses that add value to the raw data as previously described (Jung et al. 2025). For predicted gene models, the database conducts homology analyses using BLAST against SwissProt and TrEMBL protein databases (The UniProt Consortium 2017) and uses InterProScan (Blum et al. 2021) to identify protein domains and GO terms (through InterPro2GO mappings) for upload into the database. Pair-wise genome synteny analyses between genomes using McScanX (Wang et al. 2012) identify ortholog/paralog relationships, with results searchable via MegaSearch and viewable through the Synteny Viewer. The versions of the software tools and the utilized databases are stored as metadata associated with the analysis results when added to the database.

For genetic data, markers that appear on multiple maps are linked to the same marker record, allowing users to view all map positions. QTL and GWAS data are linked to a curated trait ontology that incorporates Plant Trait Ontology terms (Cooper et al. 2018) and custom Vaccinium-specific terms. A specialized pipeline aligns PCR primer pairs to genomes using blastn-short settings, allowing historical QTL, as well as PCR-based markers, to be positioned on current genome assemblies. Data are retrieved by users in a variety of ways (Fig. 2). For the data stored within the database, the search interfaces are the most efficient way to perform targeted queries and select which information to retrieve in common file types (see examples below). Bulk download of data, such as genome assemblies, annotations, and additional files from publications, are available through the GDV Data Repository. Tools such as JBrowse2 and PathwayCyc also allow for data download.

Fig. 2.

Overview of user data access on GDV. Database content can be queried through customizable searches with options to view and download results. Authors can store and share additional manuscript-related files in the GDV Data Repository. Data are also available for download from tools on GDV that are separate from the Chado database.

Breeding Information Management System

GDV provides access to an embedded Breeding Information Management System (BIMS), a specialized web-based platform designed to support crop breeding programs (Jung et al. 2021b). BIMS allows breeders to manage comprehensive breeding datasets through secure user accounts, storing and organizing private breeding data including pedigrees, phenotypic and genotypic measurements, crossing records, and selection decisions. All data entered into BIMS are backed up daily to ensure security and data integrity.

The system features interoperability through compatibility with the Field Book Android App (Rife and Poland 2014) for efficient field data collection and complies with the Breeding API (BrAPI) standards (Selby et al. 2019) to facilitate data exchange between breeding programs and external applications.

Public Germplasm Resources on GDV

A key BIMS feature is the availability of phenotype data from the USDA Germplasm Resources Information Network (GRIN). Users can access and browse USDA-GRIN phenotype measurements for Vaccinium accessions from GRIN Global (https://npgsweb.ars-grin.gov/gringlobal/search), which contains extensive characterization data for germplasm maintained in the National Plant Germplasm System. This functionality allows breeders to evaluate potential parent materials based on documented performance traits, access historical phenotypic data spanning multiple years and locations, and identify germplasm with desirable characteristics for incorporation into breeding programs.

GDV collaborated with the Breeding Insight (https://breedinginsight.org/) team to develop a blueberry crop ontology (Hislop et al. 2024), with terms stored as trait descriptors available for BIMS users and published on the Crop Ontology website (https://cropontology.org/term/CO_371:ROOT). Breeders can select these standardized trait descriptors and utilize them in private breeding programs, facilitating more efficient data sharing when results become publicly available.

Fair compliance

GDV supports the FAIR (Findable, Accessible, Interoperable, Reusable) principles for scientific data (Wilkinson et al. 2016) and implements the requirements as follows:

• Findable: All curated data includes metadata from source publications, and the relational database structure links data to publication citations visible on detail pages and search results. Each data type has standardized metadata requirements for discoverability.

• Accessible: GDV provides open access to all data without user authentication requirements. Customizable search interfaces enable users to query and download data using only a web browser and internet connection.

• Interoperable: Data are available in standard community formats (GFF, FASTA, CSV) and metadata uses established ontologies including Plant Trait Ontology (Cooper et al. 2018), Crop Ontology (Shrestha et al. 2012), Sequence Ontology (Eilbeck et al. 2005), and Gene Ontology (Ashburner et al. 2000; The Gene Ontology Consortium et al. 2023) terms.

• Reusable: All data are freely accessible with clear publication attribution and licensing, enabling unrestricted reuse by the research community.

The power of GDV: data interconnectivity

The power of GDV relies on data curation and standardization that ensures high-quality, reusable datasets, comprehensive analyses performed by the GDV team that add value beyond the original publications, and seamless data interconnectivity facilitated by diverse search interfaces and tools for data exploration. The best way to demonstrate this capability is through examples of how interconnected data supports real research scenarios. These examples are also detailed on a GDV Tutorial page with step-by-step instructions (https://www.vaccinium.org/gdv_tutorials).

Example 1: Marker-to-gene discovery for brix improvement

Breeders seeking to improve soluble solids (brix) levels in blueberry often need PCR markers to screen populations and identify putatively related genes. To address this need, users could search for markers with primer sequences that are linked to the brix trait. GDV enables trait-based searches for any markers associated with traits via QTL and GWAS analyses.

Figure 3 demonstrates this workflow using the Marker Search page (https://www.vaccinium.org/tripal_megasearch?datatype=tripal_megasearch_marker). When the brix trait is selected and the “Has Marker Primer” option is checked, the system identifies 11 associated markers. The interface allows users to select desired data fields and either view results on the webpage or download them in CSV or TSV formats. The results show that while most markers were developed in cranberry, 1 marker was developed specifically in blueberry.

Fig. 3.

The marker search interface. The trait Brix has been selected, and marker primers are checked (lower arrows, right). Clicking the Refresh Count button (circled) shows 11 markers matching the parameters (upper left). Desired fields are selected under Downloadable Fields, and clicking View displays the results table.

Detailed marker analysis

Clicking on a specific marker (such as Pr031818821, mapped in both blueberry and cranberry, https://www.vaccinium.org/bio_data/7416) opens a detailed marker page containing comprehensive information (Fig. 4). This includes the following: marker type and primer sequences; links to source publications; genetic map positions with MapViewer access; and genome positions across 7 different assemblies with JBrowse links. Through JBrowse, users can explore nearby genes and access detailed gene information within GDV. The system displays marker, QTL, and GWAS positions with bidirectional links between JBrowse and GDV details.

Fig. 4.

The marker details page. Displays all information about a marker. The Overview section contains basic details including primer sequences. Alignments show genome positions with links to open locations in JBrowse. Map Positions summarizes genetic maps and locations, with links to MapViewer. All related publications are listed in the Publications section.

Comparative genomics integration

Figure 5 illustrates the power of synteny analysis in JBrowse2 (https://www.vaccinium.org/jbrowse2/?config=config.json&session=share-9nWY0O_QLK&password=w8Rbg), showing the QTL position associated with marker Pr031818821. Users can observe that this brix locus also associates with QTL for titratable acidity and fruit weight. The comparative view displays the Vaccinium corymbosum clone W85-20 P0 genome (Mengist et al. 2023) in the top tracks and the Vaccinium macrocarpon cv. Stevens genome (Diaz-Garcia et al. 2021) in the lower tracks. The center track shows synteny analysis results using minimap2 (Li 2018), which identifies sequence homology between genomes. For each genome, the display includes the following: annotated gene models (top track); orthologs identified through MCScan (Tang et al. 2024) analysis (second track); and QTL mapping to the region via marker Pr031818821 (third track). Although these QTL were originally mapped in blueberry, the conserved synteny between cranberry and blueberry genomes suggests this region could be of interest in cranberry breeding programs as well.

Fig. 5.

JBrowse tracks displaying syntenic relationships and mapped QTL. The top 3 tracks display the V. corymbosum W85-20 P0 genome, and the bottom 3 tracks display the V. macrocarpon Stevens genome. The center track shows sequence-based synteny (red, match; yellow, insertion; blue, deletion). For each genome, the first track displays gene models, the second shows MCScan orthologs, and the third shows traits mapped via QTL with the Pr031818821 marker.

Example 2: Gene function characterization

Evolution of data types

The data landscape in Vaccinium research has evolved significantly since GDV's inception. Initially, genetic data (maps, markers, QTL) dominated the database, followed by genome assemblies. Currently, the predominant data types include SNPs linked to traits via GWAS and detailed gene characterization through bioinformatics, RNA-seq, and functional assays including gene cloning and deletion studies.

GDV began incorporating GWAS data in 2023 and detailed experimental gene function data in 2025. The data from these deeper gene studies include evidence codes that help users identify the experimental approaches used for functional characterization, with this information integrated into gene pages and made searchable through the Gene Search.

DFR gene analysis workflow

Consider researchers interested in dihydroflavonol-4-reductase (DFR) due to its role in anthocyanin biosynthesis. Users can search for this gene using the Gene Search interface (https://www.vaccinium.org/tripal_megasearch?datatype=tripal_megasearch_gene) by entering “DFR” in the “Gene Symbol/Description from publications” section (Fig. 6). To narrow results to experimentally validated genes, users can select “Cloning” from the curation evidence dropdown menu.

Fig. 6.

Gene search interface. The interface has 4 expandable search sections. In this example, the “DFR” gene symbol and “Cloning” curation evidence type is selected (horizontal arrows). After selecting display fields, clicking View shows 2 matching results in a table.

The search returns 2 genes, with VcDFR11 showing both cloning and interaction evidence. Clicking on the VcDFR11 gene name (https://www.vaccinium.org/bio_data/116287) in the table opens the gene page with more details (Fig. 7). The gene symbol, VcDFR11, is displayed in the gene overview along with the gene name from the genome and other relevant information (Fig. 7). The DFR homology was identified via a BLAST analysis using the SwissProt database under the Homology section of the gene page and there is also a DFR domain identified under the InterPro section (Fig. 7). From the gene page, users can also see that in the Expression section there is data for this gene from 4 different RNA-seq studies and the data can be browsed and there are links to the studies (Fig. 8). A list of orthologs from other genomes is available under the Orthologs section (Fig. 8) with clickable links to the gene details for each. Ortholog data can also be queried using the Ortholog/Paralog Search (https://www.vaccinium.org/tripal_megasearch?datatype=tripal_megasearch_ortholog). This integrated approach demonstrates how GDV connects multiple data types and analytical results to provide comprehensive gene characterization, supporting both basic research and applied breeding applications.

Fig. 7.

Gene details page. The Gene Overview section displays basic information such as name, symbol, description, and genome. Additional data are accessible via the left pane. The Annotated Terms section lists ontology terms, and the Homology and InterPro sections display analysis results generated by GDV.

Fig. 8.

More data from the gene details page. The Expression section provides curated transcript data from published manuscripts. The Orthologs section lists orthologous genes identified across other genomes.

Future directions

GDV continues to evolve to meet the expanding needs of the Vaccinium research community. By early 2026, GDV will undergo a comprehensive infrastructure upgrade to the latest versions of Drupal and Tripal. Concurrent with these backend improvements, a redesigned homepage and enhanced search interfaces will be implemented to provide more intuitive navigation and streamlined user experience. Several new analytical capabilities are in development. These include addition of Gigwa to facilitate exploration of genotyping data (Sempéré et al. 2019 ), particularly valuable for the large-scale SNP datasets increasingly common in Vaccinium research, and more custom JBrowse2 tracks to display orthologs across multiple genomes alongside core annotation tracks for key genomes. The Expression Heatmap module will receive significant updates to enable more flexible searching and filtering of expression value and biomaterial metadata. Improvements will also be made to improve ontology term provenance (eg GO) as well as the display of this associated information on the website when viewing these terms as this is very useful information for users of GDV.

To support user adoption and training, MainLab Bioinformatics maintains a YouTube channel (https://www.youtube.com/channel/UC0S8SqxuMpfScejyE-VCtxA) featuring instructional videos for navigating the crop databases and using BIMS. The shared infrastructure across all 5 knowledgebases allows demonstrations from one database to serve as universal tutorials. Separate user guides for each website are currently being transitioned to a unified documentation system using Read the Docs (https://about.readthedocs.com/) to streamline maintenance and improve accessibility. This consolidated manual is under active development and already includes comprehensive JBrowse2 usage instructions (https://mainlab-usermanual.readthedocs.io/en/latest/). An updated BIMS user manual is also available (https://bims-userguide.readthedocs.io/en/latest/).

GDV development priorities will continue to be shaped by Vaccinium research community needs, emerging data types, and technological advances. Through ongoing collaboration with users and data producers, the goal is to ensure GDV remains an essential resource for advancing Vaccinium genetics, genomics, and breeding research. Like most biological databases, ensuring long-term sustainability for these critical community resources presents an ongoing challenge. In collaboration with the AgBioData consortium (https://www.agbiodata.org) (Harper et al. 2018; Clarke et al. 2023), we are actively working to develop sustainable funding models and operational strategies to address this fundamental issue facing the broader scientific database community.

Data availability

All genomic, genetic, and germplasm data curated in GDV is publicly accessible at https://www.vaccinium.org/. Private breeding trial data managed through BIMS are not publicly available and remain confidential within individual breeding programs.

Funding

This work is supported by the United States Department of Agriculture (USDA) via four sources. Three awards from the USDA National Institute of Food and Agriculture: Specialty Crop Research Initiative project award numbers 2022-51181-38449 and 2019-51181-30015 as well as National Research Support Project 10. This work is also supported by USDA Hatch project 1014919 at Washington State University.

Conflicts of interest. None declared.