Image analysis platforms for exploring genetic and neuronal mechanisms regulating animal behavior

Abstract

An important aim of neuroscience is to understand how gene interactions and neuronal networks regulate animal behavior. The larvae of the marine annelid Platynereis dumerilii provide a convenient system for such integrative studies. These larvae exhibit a wide range of behaviors, including phototaxis, chemotaxis and gravitaxis and at the same time exhibit relatively simple nervous system organization. Due to its small size and transparent body, the Platynereis larva is compatible with whole-body light microscopic imaging following tissue staining protocols. It is also suitable for serial electron microscopic imaging and subsequent neuronal connectome reconstruction. Despite advances in imaging techniques, automated computational tools for large data analysis are not well-established in Platynereis. In the current work, I developed image analysis software for exploring genetic and nervous system mechanisms modulating Platynereis behavior. Exploring gene expression patterns Current labeling and imaging techniques restrict the number of gene expression patterns that can be labelled and visualized in a single specimen, which hinders the study of behaviors driven by multi-molecular interactions. To address this problem, I employed image registration to generate a gene expression atlas that integrates gene expression information from multiple specimens in a common reference space. The gene expression atlas was used to investigate mechanisms regulating larval locomotion, settlement and phototaxis in Platynereis. The atlas can assist in the identification of inter-individual and inter-species variations in gene expression. To provide a representation convenient for exploring gene expression patterns, I created a model of the atlas using 3D graphics software, which enabled convenient data visualization and efficient data storage and sharing. Exploring neuronal networks regulating behavior Neuronal circuitry can be reconstructed from the images obtained from electron microscopy, which resolves very fine structures such as neuron morphology or synapses. The amount of data resulting from electron microscopy and the complexity of neuronal networks represent a significant challenge for manual analysis. To solve this problem, I developed the NeuroDetective software, which models a neuronal circuitry and analyzes the information flow within it. The software combines the advantages of 3D visualization and graph analysis software by integrating neuron morphology and spatial distribution together with synaptic connectivity. NeuroDetective allowed studying the neuronal circuitry responsible for phototaxis in Platynereis larvae, revealing the connections and the neurons important for the network functionality. NeuroDetective facilitated the establishment of a relationship between the function and the structure of the neuronal circuitry in Platynereis phototaxis. Integrating gene expression patterns with neuronal connectivity Neuronal circuitry and its associated modulating biomolecules, such as neurotransmitters and neuropeptides, are thought to be the main factors regulating animal behavior. Therefore it was important to integrate both genetic and neuronal information in order to fully understand how biomolecules in conjunction with neuronal anatomy elicit certain animal behavior. To resolve the difference in specimen preparation for gene expression versus electron microscopy preparations, I developed an image registration procedure to match the signals from these two different datasets. This method enabled the integration the spatial distribution of specific modulators into the analysis of neuronal networks, leading to an improved understanding of the genetic and neuronal mechanisms modulating behavior in Platynereis.