Deep Fish.

Zebrafish ( Danio rerio) is an important vertebrate model organism in biomedical research, especially suitable for morphological screening due to its transparent body during early development. Deep learning has emerged as a dominant paradigm for data analysis and found a number of applications in computer vision and image analysis. Here we demonstrate the potential of a deep learning approach for accurate high-throughput classification of whole-body zebrafish deformations in multifish microwell plates. Deep learning uses the raw image data as an input, without the need of expert knowledge for feature design or optimization of the segmentation parameters. We trained the deep learning classifier on as few as 84 images (before data augmentation) and achieved a classification accuracy of 92.8% on an unseen test data set that is comparable to the previous state of the art (95%) based on user-specified segmentation and deformation metrics. Ablation studies by digitally removing whole fish or parts of the fish from the images revealed that the classifier learned discriminative features from the image foreground, and we observed that the deformations of the head region, rather than the visually apparent bent tail, were more important for good classification performance.

Compaction of rolling circle amplification products increases signal integrity and signal-to-noise ratio.

Rolling circle amplification (RCA) for generation of distinct fluorescent signals in situ relies upon the self-collapsing properties of single-stranded DNA in commonly used RCA-based methods. By introducing a cross-hybridizing DNA oligonucleotide during rolling circle amplification, we demonstrate that the fluorophore-labeled RCA products (RCPs) become smaller. The reduced size of RCPs increases the local concentration of fluorophores and as a result, the signal intensity increases together with the signal-to-noise ratio. Furthermore, we have found that RCPs sometimes tend to disintegrate and may be recorded as several RCPs, a trait that is prevented with our cross-hybridizing DNA oligonucleotide. These effects generated by compaction of RCPs improve accuracy of visual as well as automated in situ analysis for RCA based methods, such as proximity ligation assays (PLA) and padlock probes.

AUTOMATED QUANTIFICATION OF ZEBRAFISH TAIL DEFORMATION FOR HIGH-THROUGHPUT DRUG SCREENING.

Zebrafish (Danio rerio) is an important vertebrate model organism in biomedical research thanks to its ease of handling and translucent body, enabling in vivo imaging. Zebrafish embryos undergo spinal deformation upon exposure to chemical agents that inhibit DNA repair. Automated image-based quantification of spine deformation is therefore attractive for whole-organism based assays for use in early-phase drug discovery. We propose an automated method for accurate high-throughput measurement of tail deformations in multi-fish micro-plate wells. The method generates refined medial representations of partial tail-segments. Subsequently, these disjoint segments are analyzed and fused to generate complete tails. Based on estimated tail curvatures we reach a classification accuracy of 91% on individual animals as compared to known control treatment. This accuracy is increased to 95% when combining scores for fish in the same well.

Recovering the dynamics of root growth and development using novel image acquisition and analysis methods.

Roots are highly responsive to environmental signals encountered in the rhizosphere, such as nutrients, mechanical resistance and gravity. As a result, root growth and development is very plastic. If this complex and vital process is to be understood, methods and tools are required to capture the dynamics of root responses. Tools are needed which are high-throughput, supporting large-scale experimental work, and provide accurate, high-resolution, quantitative data. We describe and demonstrate the efficacy of the high-throughput and high-resolution root imaging systems recently developed within the Centre for Plant Integrative Biology (CPIB). This toolset includes (i) robotic imaging hardware to generate time-lapse datasets from standard cameras under infrared illumination and (ii) automated image analysis methods and software to extract quantitative information about root growth and development both from these images and via high-resolution light microscopy. These methods are demonstrated using data gathered during an experimental study of the gravitropic response of Arabidopsis thaliana.

Extraction of the plane of minimal cross-sectional area of the corpus callosum using template-driven segmentation.

Changes in corpus callosum (CC) size are typically quantified in clinical studies by measuring the CC cross-sectional area on a midsagittal plane. We propose an alternative measurement plane based on the role of the CC as a bottleneck structure in determining the rate of interhemispheric neural transmission. We designate this plane as the Minimum Corpus Callosum Area Plane (MCCAP), which captures the cross section of the CC that best represents an upper bound on interhemispheric transmission. Our MCCAP extraction method uses a nested optimization framework, segmenting the CC as it appears on each candidate plane, using registration-based segmentation. We demonstrate the robust convergence and high accuracy of our method for magnetic resonance images and present preliminary clinical results showing higher sensitivity to disease-induced atrophy.

Longitudinal, regional and deformation-specific corpus callosum shape analysis for multiple sclerosis.

The corpus callosum (CC) is an anatomical structure which connects the two brain hemispheres. Neurological diseases can cause atrophy of the CC resulting in a change in its size and shape. The measurement and analysis of this change is one of the goals of clinical research. We perform statistical analysis of the shape of the CC extracted from MR brain scans of a group of multiple sclerosis patients undergoing a longitudinal (serial) study. In contrast to the classical boundary-based, global shape variability measures, e.g. principal component analysis (PCA) of CC boundary vertices, we perform a deformation-specific PCA for analyzing the global and regional shape of the CC. This deformation-specific PCA is based on a medial-based shape representation. The adopted shape representation describes shape variability in terms of intuitive deformations (e.g. bending, stretching and thickness). We present qualitative and quantitative results for 412 MR images of the CC. We show that our method is successful in identifying and quantifying the effect of each type of deformation on the shape variability of the CC. In addition to analyzing the spatial shape variability in the CC, we explore shape changes as the disease progresses. Our method allows the exploration of the shape variability quantitatively (e.g. the amount of variance explained by a particular principal mode of shape variation) as well as in a qualitative visual manner (e.g. by visualizing, say, the 2nd principal mode of shape variation due to bending at the 4th sub-region of the CC) which is useful for developing an intuitive understanding of the effects of MS on the CC shape.