Virtual Brain, an anatomical standard for the Drosophila-brain
Project management at the University of Würzburg:
The project is part of a larger effort to develope software for the visualization, handling and processing of 3D image data in neuroanatomy (Virtual Neuroanatomy Laboratory, VNL). In particular, the project is designed to support confocal (CLS-) microscopy in behavioral neurogenetics of Drosophila melanogaster. The VNL should make it possible (1) to work on 3D-data sets of whole-mount preparations of fly brains generated by confocal microscopy; (2) to visualize at the same time two complete data sets on the screen and to superimpose them manually in real time; (3) to determine average values and variances from several brains ("Standard brain"); (4) to quantitatively evaluate differences between brains and brain regions of different animals; (5) to match gene expression patterns and single cells to the Standard brain. The data of the VNL should be made available in the internet.
On the basis of the visualization software AMIRA, different tools were developed in close cooperation with M. Zoeckler, Berlin to support confocal microscopy of the Drosophila brain. After collection of the raw data at the confocal microscope these are converted automatically into the AMIRA data format. As a first ?preprocessing? step the scaling needs to be corrected for the specific refractive index of the preparation. If necessary, it is possible to change the number of data points (?voxels?) (?resampling?) and to reduce the data sets to the regions of interest (?cropping?). The data sets are then brought into the same orientation (?alignment?). This can either be performed manually or fully automatically using different procedures (e.g. grey value correlation). All neuropile areas of interest are now marked in the grey value data sets and receive their appropriate names (?segmentation?, ?labelling?). This procedure is a labour-intensive but necessary step to compare brains, to decribe their variability using quantitative anatomy (s.u.), to calculate brain average values (Standard brain) and to integrate gene expression-patterns.
Volume visualizations which are changable in real time, now make it possible to interactively explore the brain working with the 3D-data set (?data mining?) and thus create a completely new possibility of data analysis - a ?virtual anatomy?: Virtual sections can be moved in real time under visual control through the data set. An ?electronic scalpel? can remove parts of the data (?clipping?) and ?virtual staining? can emphasize certain areas of the data set by changing their color and opacity. The combination of data sets from different individuals on the same screen provides completely new possibilities for comparison and analysis.
Fly brains differ among individuals, just like human brains. For many problems in neuroanatomy, one has to quantify this variability. The analysis of volume, shape and position differences of neuropile regions is possible using ?average intensity maps?, ?probabilistic maps? and special graphs.
The goals of the research proposal have been largely met. The Virtual Neuroanatomy Laboratory (VNL) is a reality and has been successfully applied to Drosophila neurogenetics. The commercial product, an extension of the visualization software AMIRA, is available form ?Indeed - Visual Concepts GmbH?. It contains contributions from several laboratories. The Standard brain for females and males of the wildtype strain "CantonS", and for the comparison of the local wildtype strain "Lindelbach" are available in the Internet. Applications demonstrating the use of the VNL and the Standard brain (comparison of the two wildtype strains, gender differences, comparison with mutants, representation of gene expression patterns), are published.
The current goal is to spread the VNL in order to establish a general data formate for Drosophila neurogenetics and a data base for standardized gene expression patterns. At present the main emphasis is on the characterisation of gene expression patterns. In the long run, a genetic brain atlas will develop connecting all molecular information of gene function with the anatomy of the brain. The modification of gene expression patterns with time and under specific environmental influences (learning, experience) is a further dimension of this new kind of atlas.
Projekt period: from 09.2000 to 08.2003
Landeshaushalt Wissenschaftsministerium ( BMBF ) ,Granting date: 28.07.2000