Bookshelf (http://www.ncbi.nlm.nih.gov/books/) is a full-text electronic literature
resource of books and documents in life sciences and health care at the National Center
for Biotechnology Information (NCBI). Created in 1999 with a single book as an
encyclopedic reference for resources such as PubMed and GenBank, it has grown to its
current size of >1300 titles. Unlike other NCBI databases, such as GenBank and Gene,
which have a strict data structure, books come in all forms; they are diverse in
publication types, formats, sizes and authoring models. The Bookshelf data format is XML
tagged in the NCBI Book DTD (Document Type Definition), modeled after the National Library
of Medicine journal article DTDs. The book DTD has been used for systematically tagging
the diverse data formats of books, a move that has set the foundation for the growth of
this resource. Books at NCBI followed the route of journal articles in the PubMed Central
project, using the PubMed Central architectural framework, workflows and processes.
Through integration with other NCBI molecular databases, books at NCBI can be used to
provide reference information for biological data and facilitate its discovery. This
article describes Bookshelf at NCBI: its growth, data handling and retrieval and
integration with molecular databases.