What is a dataset?
In GDI, a dataset is a structured collection of genomic information pertaining to human health, diseases, and research studies. A dataset can be a single file or a collection of files that provide comprehensive information about a specific research topic, disease area, or study cohort. For example, a genomic data for COVID-19 Viral Sequences can include dataset records describing patient data, virus samples, and sequencing results.
When you add a dataset to the GDI Data Catalogue, you provide two types of information that together give a complete picture of the genomic subject: metadata describes the dataset itself, and data resources are the actual genomic data files associated with the dataset.
Metadata
Metadata (displayed as Additional Info in the portal) are descriptive details about your genomic dataset that provide context and information about the dataset itself. It includes details pertaining to:
- Identification: Title, description, keywords, and unique identifiers
- Responsibility: Contact points, publisher, creator, and data steward information
- Access information: Rights, availability, licensing, and access restrictions
- Others: Other key information that allows users to locate and access the data.
Data resources
Data resources pertain to the actual genomic data files associated with a dataset. In GDI, you can upload the file or provide links to external data resources. GDI supports common genomic data file formats, including:
- VCF (Variant Call Format): For storing gene sequence variations
- FASTA/FASTQ: For storing raw sequence reads
- BAM/CRAM: For storing aligned sequence data
- CSV/TSV: For storing tabular data such as phenotypic data
Datasets in GDI can be organised in several ways—such as by organisation and groups—to help you manage data effectively, while making it easy for researchers to find relevant datasets.