Biological databases and visualisation

A biological database is an organised, searchable collection of biological data, linked to where it came from. It may hold gene sequences, disease-causing mutations, how common certain DNA variants are, protein sequences, 3-D structures and interactions.

Think of it as building a Google for molecules, or a library catalogue: you can find a particular book without searching every shelf. In the same way, you can search sequences and structures far too small to see. As genome data grows, this is essential.

A good database is easy to use, well documented, supported, cross-referenced (linked to other databases), free of errors and kept up to date, since new data and mutations keep being found (just as your phone needs regular updates). Databases are either relational (data in linked tables, searched using SQL) or non-relational (NoSQL, for large, unstructured data). They are run by software called a database management system (DBMS).

Major databases:

• GenBank: DNA sequences.
• PDB: 3-D structures of proteins and other molecules.
• UniProt: protein sequences and functions.
• PubMed: research papers.
• KEGG: biological pathways, drugs and diseases.
• OMIM: human genes and genetic disorders.

Data must also be seen to be understood. Visualisation tools: the UCSC Genome Browser and IGV (genomes); KEGG, Biocarta and Reactome (pathways); CIRCOS (data in a circular layout); and Excel, R and D3.js (charts and graphics).