by Guanyang Zhang
Letter published in 'Science'
In his Perspective “GenBank—natural history in the 21st century?” (24 October 2008, p. 537), B. J. Strasser claims... more
In his Perspective “GenBank—natural history in the 21st century?” (24 October 2008, p. 537), B. J. Strasser claims that GenBank follows the tradition of natural history studies. I argue that GenBank is inconsistent with some important aspects of the tradition of natural history and it alone does not and will not constitute natural history.
First, GenBank has been designed to store molecular information about model organisms or humans, whereas natural history serves to explore and document biodiversity. GenBank's format is incapable of handling unique aspects of biodiversity studies such as diverse and large collections of specimens, and taxonomic uncertainties and revisions. Second, GenBank does not require vouchering of specimens, DNA extracts, or other molecular products, whereas the study of natural history always anchors information with specimens. Without such anchoring, revisionary work (which is part of the tradition of natural history) cannot be conducted. As a result, GenBank contains a considerable amount of unidentified or misidentified sequences (1–3). Third, GenBank is unlikely to reach amateurs in the way that natural history has. Study of natural history generally does not require expensive equipment and is exciting and rewarding. In comparison, DNA sequencing is costly, offers little aesthetic reward or recreational value, and requires specialist knowledge. It is hard to imagine amateurs comparing or exchanging their collections of DNA sequence instead of actual specimens.
Natural history documentation has many important facets, such as specimen locality data, images of morphology, and ecological notes. The Global Biodiversity Information Facility (http://www.gbif.org/) and Morphbank (http://www.morphbank.net/) are good examples. Specimen collections will continue serving as the primary record of natural history and biodiversity, while molecular data are supplementary.
Download (.pdf) 
