R. von Collenberg, C. | Search bioRxiv

bioRxiv2019

Essential role of the mouse synapse associated protein Syap1 in circuits for spontaneous motor activity and rotarod balance

Synapse-associated protein 1 (Syap1) is the mammalian homologue of synapse-associated protein of 47 kDa (Sap47) in Drosophila. Genetic deletion of Sap47 leads to deficiencies in short-term plasticity and associative memory processing in flies. In mice, Syap1 is prominently expressed in the nervous system, but its function is still unclear. We have generated Syap1 knockout mice and tested motor behaviour and memory. These mice are viable and fertile but display distinct deficiencies in motor behaviour. Locomotor activity specifically appears to be reduced in early phases when voluntary movement is initiated. On the rotarod, a more demanding motor test involving control by sensory feedback, Syap1-deficient mice dramatically fail to adapt to accelerated speed or to a change in rotation direction. Syap1 is highly expressed in cerebellar Purkinje cells and cerebellar nuclei. Thus, this distinct motor phenotype could be due to a so far unknown function of Syap1 in cerebellar sensorimotor control. The observed motor defects are highly specific since other tests in the modified SHIRPA test, as well as cognitive tasks like novel object recognition, Pavlovian fear conditioning, and anxiety-like behaviour in open field, dark-light transition, and elevated plus maze, do not appear to be strongly affected in Syap1 knockout mice.

animal behavior and cognition↗

bioRxiv2018

DeepFLaSh, a deep learning pipeline for segmentation of fluorescent labels in microscopy images

Fluorescent labeling of biomolecules is widely used for bioimage analyses throughout the life sciences. Recent advances in deep learning (DL) have opened new possibilities to scale the image analysis processes through automation. However, the annotation of fluorescent features with a low signal-to-noise ratio is frequently based on subjective criteria. Training on subjective annotations may ultimately lead to biased DL models yielding irreproducible results. An end-to-end analysis process that integrates data annotation, ground truth estimation, and model training can mitigate this risk. To highlight the importance of this integrated process, we compare different DL-based analysis approaches. Based on data from different laboratories, we show that ground truth estimation from multiple human annotators is indispensable to establish objectivity in fluorescent feature annotations. We demonstrate that ensembles of multiple models trained on the estimated ground truth establish reliability and validity. Our research provides guidelines for reproducible and transparent bioimage analyses using DL methods.

neuroscience↗