Overview
Biological databases are the foundational infrastructure of bioinformatics. They collect, curate, organize, and distribute the vast quantities of molecular data generated by modern experimental biology. These databases range from small, specialized resources maintained by individual laboratories to massive international repositories such as GenBank, which contains billions of nucleotide sequences. The value of any single dataset is multiplied when it is deposited in a public database, because it becomes discoverable, reusable, and integrable with other data types. Data standards, controlled vocabularies, and cross-referencing between databases enable this interoperability.
Key Concepts
Primary databases store original experimental data, raw nucleotide sequences, protein sequences, and three-dimensional structures, with minimal interpretation. Secondary databases contain curated, annotated, or derived information built by analyzing primary data. Sequence databases (GenBank, UniProt) store nucleotide and protein sequences. Structure databases (PDB) archive macromolecular coordinates. Functional databases (GO, KEGG, Reactome) describe biological processes, pathways, and molecular functions. Cross-references connect records across databases, enabling integrated queries.
Applications
Biological databases support virtually every area of molecular biology. Researchers use them to retrieve sequences for DNA sequencing projects, search for homologous protein structures, and look up enzyme classification and nomenclature for metabolic reconstruction. Database integration is essential for systems biology, where data from multiple sources must be combined to build predictive models of cellular behavior.
Practical Protocol
Choosing the right database begins with defining your research question. If you need raw nucleotide sequences, start with a primary database like GenBank via the NCBI portal. For annotated reference genomes, switch to RefSeq. If your question concerns protein function, UniProtKB is the appropriate resource, while structural questions point to the PDB. For a typical genomics project, say, identifying novel genes in a bacterial genome, a common workflow proceeds as follows: first, query the Genome database at NCBI using the organism name to locate the assembly. Download the nucleotide sequences in GenBank or FASTA format. Submit these sequences to BLASTX against the UniProtKB/Swiss-Prot database to predict protein-coding regions based on homology. Map the resulting annotations to GO terms using InterProScan to infer molecular functions and biological processes. Finally, use KEGG Mapper to place the annotated genes onto metabolic pathway maps, revealing the organism’s metabolic capabilities. This integrated approach transforms raw sequence data into testable biological hypotheses. The same cross-database strategy applies to human genomics: starting with a list of differentially expressed genes from RNA-seq, a researcher can retrieve promoter sequences from UCSC, predict transcription factor binding sites using JASPAR, map protein products to Reactome pathways, and validate interactions via STRING, all by navigating between the appropriate databases linked through cross-references. Mastering this multi-database workflow is a core competency in modern bioinformatics, enabling researchers to generate robust, reproducible findings from the wealth of publicly available biological data.
