DATABASE TOOLS FOR INTEGRATING AND SEARCHING MEMBRANE PROPERTY DATA CORRELATED WITH NEURONAL MORPHOLOGY

A critical problem in neuroscience is the lack of database tools for i ntegrating neuronal property data. We report here the development of a combined object oriented-relational database (NeuronDB, http://sensel ab.med.yale.edu/neurondb) that meets these needs by providing tools fo r integrating data within neurons and comparing data across neurons. I t focuses on three types of neuronal properties-voltage-gated channels , neurotransmitter receptors, and neurotransmitters. :The data are org anized in relation to different regions of neurons as represented in c anonical forms; using simple canonical models of complex cells as a ve hicle for indexing information permits the database to be searchable a cross different neurons. Using these multidimensional search tools, us ers can locate specific properties in specific regions of a neuron; ob tain integrated summaries of all properties within a region; and carry out searches to compare properties across equivalent compartments in different neurons. These tools thus permit searches of the multidimens ional neuron property space equivalent to homology searches of sequenc e databases. NeuronDB is accessible over the Internet; it provides imm ediate links to citation indexes and abstracts supporting the deposite d data, and annotations that indicate the state of acceptance of the d ata. Users are encouraged to contribute data. The ability to input the data from NeuronDB directly to NEURON and GENESIS is being developed. As a shared Web resource, NeuronDB should enhance the efforts of neur oscientists and neuronal modellers to analyze and compare the function al operations of different types of neurons. (C) 1998 Elsevier Science B.V. All rights reserved.