Flytrap - HTML based gene expression database.

Correlating brain structure and function, across a range of levels of description, is one of the most challenging scientific goal s for the next century. Informatics tools allowing the vast arrays of neurobiological information to be integrated in a seamless and interactive manner will be essential to this exercise. Advances in neural imaging techniques and molecular genetic tools are provid ing exciting new methods for investigating the development and function of the nervous system.

Due to its small size, yet behavioural complexity, the brain of the fruit fly Drosophila is now largely accepted as an imp ortant model for brain function. In Drosophila, development of the P[GAL4] enhancer-trap system has been of particular signif icance as it not only marks cells for visualisation, but also provides a means for manipulating them in the living organism. Several laboratories now have extensive collections genetically transformed fly strains, each with accompanying image data, genetic data an d/or phenotypic data.

We have been developing an interactive database for displaying gene expression patterns, in particular P[GAL4] patterns, via an i ntuitive WWW based interface. This development consists of two components, the first being the html interface to the database and th e second, a tool-kit for constructing and maintaining the database.

The browser component of the project is entirely platform independent; based on javascript and html and therefore only requires a 'standard' browser. This is to facilitate CD-ROM distribution and off-line browsing. Whether on-line or on CD, the basic browser s tructure does not reply on any server based scripts.

Basic searching is now available. The search page uses javascript and will work off-line (i.e. from a CD-ROM copy).

The construction tool-kit is UNIX based and requires an on-line web server. The tool-kit is used to compile the html browser inte rface from a simple database. The tool-kit part comprises a forms based html interface to the datasets allowing new information to b e added and updated very simply. We are also developing a java interface for the tool-kit that will enable us to edit and annotate i mages on-line.

The basic browser interface is complete and a demonstration version can be accessed via the web at http://www.fly-trap.org/. The first working version of the tool-kit is now on-line and is available for use.

J. Douglas Armstrong, Chris Edwards and Kim Kaiser