Chapter 10. GenBank

GenBank (Genetic Sequence Data Bank) is a rapidly growing international repository of known genetic sequences from a variety of organisms. Its use is central to modern biology and to bioinformatics.

This chapter shows you how to write Perl programs to extract information from GenBank files and libraries. Exercises include looking for patterns; creating special libraries; and parsing the flat-file format to extract the DNA, annotation, and features. You will learn how to make a DBM database to create your own rapid-access lookups on selected data in a GenBank library.

Perl is a great tool for dealing with GenBank files. It enables you to extract and use any of the detailed data in the sequence and in the annotation, such as in the FEATURES table and elsewhere. When I first started using Perl, I wrote a program that searched GenBank for all sequence records annotated as being located on human chromosome 22. I found many genes where that information was so deeply buried within the annotation, that the major gene mapping database, Genome Database (GDB), hadn't included them in their chromosome map. I think you'll discover the same feeling of power over the information when you start applying Perl to GenBank files.

Most biologists are familiar with GenBank. Researchers can perform a search, e.g., a BLAST search on some query sequence, and collect a set of GenBank files of related sequences as a result. Because the GenBank records are maintained by the individual scientists who discovered the sequences, if you find some new sequence of interest, you can publish it in GenBank.

GenBank files have a great deal of information in them in addition to sequence data, including identifiers such as accession numbers and gene names, phylogenetic classification, and references to published literature. A GenBank file may also include a detailed FEATURES table that summarizes facts about the sequence, such as the location of the regulatory regions, the protein translation, and exons and introns.

GenBank is sometimes referred to as a databank or data store, which is different from a database. Databases typically have a relational structure imposed upon the data, including associated indices and links and a query language. GenBank in comparison is a flat file, that is, an ASCII text file that is easily readable by humans.^[1]

^[1] GenBank is also distributed in ASN.1 format, for which you need specialized tools, provided by NCBI.

From its humble beginnings GenBank has rapidly grown, and the flat-file format has seen signs of strain during the growth. With a quickly advancing body of knowledge, especially one that's growing as quickly as genetic data, it's difficult for the design of a databank to keep up. Several reworkings of GenBank have been done, but the flat-file format—in all its frustrating glory—still remains.

Due to a certain flexibility in the content of some sections of a GenBank record, extracting the information you're looking for can be tricky. This flexibility is good, in that it allows you to put what you think is most important into the data's annotation. It's bad, because that same flexibility makes it harder to write programs that to find and extract the desired annotations. As a result, the trend has been towards more structure in the annotations.

Since Perl's data structures and its use of regular expressions make it a good tool for manipulating flat files, Perl is especially well-suited to deal with GenBank data. Using these features in Perl and building on the skills you've developed from previous chapters, you can write programs to access the accumulated genetic knowledge of the scientific community in GenBank.

Since this is a beginning book that requires no programming experience, you should not expect to find the most finished, multipurpose software here. Instead you'll find a solid introduction to parsing and building fast lookup tables for GenBank files. If you've never done so, I strongly recommend you explore the National Center for Biotechnology Information (NCBI) at the National Institutes of Health (NIH) (http://www.ncbi.nlm.nih.gov). While you're at it, stop by the European Bioinformatics Institute (EBI) at http://www.ebi.ac.uk and the bioinformatics arm of the European Molecular Biology Laboratory (EMBL) at http://www.embl-heidelberg.de/. These are large, heavily funded governmental bioinformatics powerhouses, and they have (and distribute) a great deal of state-of-the-art bioinformatics software.