sequence databases (GenBank) and formats
/ JEN-bank /
Science only compounds if discoveries are shared, and in genomics that sharing happens through enormous public repositories of sequence. When a lab determines a new gene or genome, it deposits the sequence in a database that anyone on Earth can search and download for free. That cumulative library is what makes a single BLAST search able to recognize your unknown gene — it is comparing against decades of everyone's deposited work.
GenBank, run by the US NIH (and mirrored by partner databases in Europe and Japan), is the best-known of these archives: a vast, growing collection of publicly available DNA and RNA sequences with their annotations and the publications behind them. To move all this data around, the field relies on a few plain-text formats. FASTA is the simplest: a one-line header beginning with the symbol > giving the sequence's name, followed by the raw letters, like '>my_gene' on one line and 'ATGGCGTTAC...' on the next. FASTQ extends this for sequencing reads by adding, for every base, a quality score saying how confident the machine was in that letter. Other formats like SAM/BAM record where reads aligned, and VCF records variants.
These resources and formats are the quiet infrastructure that makes modern biology interoperable — a sequence deposited in Tokyo can be analyzed in São Paulo an hour later. But the honesty caveat is important: public databases are only lightly curated, so they contain errors, contamination from other organisms, and out-of-date or wrong annotations that propagate when people copy them. A sequence's presence in GenBank means someone submitted it, not that it is correct, which is one more reason results need independent checking rather than blind trust in a downloaded label.
A FASTA record looks like '>BRCA1 partial sequence' on one line and 'ATGGATTTATCTGCTCTTCGCGTT...' on the next — just a name and the letters, the lingua franca that lets any program read any sequence.
GenBank stores the world's sequences; simple formats like FASTA let any tool read them.
Public databases are lightly curated. A sequence in GenBank means someone submitted it, not that it is correct — errors, contamination, and stale annotations are common and get copied onward.