Multicellular organisms are inherently three-dimensional. This leads to complex intercellular interactions that cannot be reproduced in two-dimensional cell culture. Instead, three-dimensional spheroids, ball-shaped cell aggregates, arise as model systems. Spheroids provide an accurate in vitro representation of the three-dimensional organization of cells in tissues, and compared to a real tissue, they excel with well-defined experimental conditions, easy handling, and suitability for high-quality imaging. Therefore, spheroids are an experimental system that can be readily combined with mathematical modeling. This chapter shows how image-based systems biology is implemented for multicellular spheroids to study three-dimensional cell–cell interactions. The chapter is intended for experimentalists and theoreticians who plan to extend their research by linkage with other disciplines. The relevant concepts for experimental approaches and quantitative imaging are introduced and linked to mathematical models of spheroids. This results in a list of potential systems biology workflows for typical spheroid research areas in cell biology, cancer biology, and bioprinting. In all three areas, there is a large gap between the details of the mathematical models and the available imaging data. The aim of this chapter is to encourage more interactions of experimentalists and theoreticians to fill this gap in spheroid research.

The rise of omics techniques has resulted in an explosion of molecular data in modern biomedical research. Together with information from medical images and clinical data, the field of omics has driven the implementation of personalized medicine. Biomedical and omics datasets are complex and heterogeneous, and extracting meaningful knowledge from this vast amount of information is by far the most important challenge for bioinformatics and machine learning researchers. In this context, there is an increasing interest in the potential of deep learning (DL) methods to create predictive models and to identify complex patterns from these large datasets. This chapter provides an overview of the main applications of DL methods in biomedical research, with focus on omics data analysis and precision medicine applications. DL algorithms and the most popular architectures are introduced first. This is followed by a review of some of the main applications and problems approached by DL in omics data and medical image analysis. Finally, implementations for improving the diagnosis, treatment, and classification of complex diseases are discussed.

Our understanding of biology has undergone a revolution in the past 20 years, driven by our ability to capture, store, interrogate and analyze the ever-increasing volumes of ‘omics’ data. Computational Biology, an integrated approach employing high performance computers, state-of-the art software and algorithms, mathematical modeling and statistical analyses have enabled us to unravel the seemingly impenetrable complexity of biological systems. This book draws together many of the latest cutting-edge developments in the field of Computational Biology. Each chapter draws on the expertise of leading researchers in the field to highlight the utility of specific technologies. The breadth of the text is impressive - from integrative biology in human diseases through the various branches of epigenomics, metabolomics and proteomics to biological sequencing and deep learning. Computational biology approaches for image-based analysis of multicellular spheroids, feature selection using entropy and cellular cryo-electron tomography structural pattern mining are covered. In addition, the key role of statistics in the analysis of high-dimensional multiset omics data and RNA sequencing are discussed in dedicated chapters. This book would have broad appeal to anyone with an interest in cutting edge computational biology.

The quantitative evaluation of computed tomography (CT) images is widely investigated and applied in clinical diagnosis. However, the CT number of tissue can vary with scanners or applied tube voltages because of the x-ray spectrum dependence of measured linear attenuation coefficients that degrades evaluation accuracy and limits multicenter or multimodality research. This study proposed a novel CT image synthesis method to correct the spectrum dependence of CT numbers by normalizing them to the same spectrum condition. Stoichiometric calibration was performed to derive the spectrum characteristic parameters (SCPs) of six spectra from two CT scanners with different applied tube voltages. Subsequently, conversion relationships between CT numbers and tissue parameters (TPs) were determined using the SCPs and standard tissue data. The CT number of a tissue measured from a spectrum condition was converted to TPs using these relationships, and the results were used to estimate the CT number of the tissue in another spectrum condition using the corresponding SCPs. Phantom, cadaver, and patient studies were performed to evaluate the proposed method. In the phantom study, image synthesis reduced the mean difference between the CT numbers of tissue-equivalent phantoms measured using different spectra from 57.96 to 33.94 HU. In the cadaveric study, the mean difference between the CT numbers of a temporal bone flap measured using different spectra was lowered by over 57%. In the patient image study, a significant difference of 81.5 HU was observed between the mean CT numbers of femoral shafts obtained from the two scanners; this difference was reduced to less than 17 HU, which was nonsignificant, when the proposed method was used. The proposed image synthesis method could reduce the spectrum dependence of CT numbers measured with different spectra and could be applied clinically to improve the accuracy of multicenter and multimodality evaluation and research.

Cone-beam computed tomography (CBCT) scans are commonly used in diagnosing and planning surgical or orthodontic treatment to correct craniomaxillofacial (CMF) deformities. Based on CBCT images, it is clinically essential to generate an accurate 3D model of CMF structures (e.g., midface, and mandible) and digitize anatomical landmarks. This process often involves two tasks, i.e., bone segmentation and anatomical landmark digitization. Because landmarks usually lie on the boundaries of segmented bone regions, the tasks of bone segmentation and landmark digitization could be highly associated. Also, the spatial context information (e.g., displacements from voxels to landmarks) in CBCT images is intuitively important for accurately indicating the spatial association between voxels and landmarks. However, most of the existing studies simply treat bone segmentation and landmark digitization as two standalone tasks without considering their inherent relationship, and rarely take advantage of the spatial context information contained in CBCT images. To address these issues, we propose a Joint bone Segmentation and landmark Digitization (JSD) framework via context-guided fully convolutional networks (FCNs). Specifically, we first utilize displacement maps to model the spatial context information in CBCT images, where each element in the displacement map denotes the displacement from a voxel to a particular landmark. An FCN is learned to construct the mapping from the input image to its corresponding displacement maps. Using the learned displacement maps as guidance, we further develop a multi-task FCN model to perform bone segmentation and landmark digitization jointly. We validate the proposed JSD method on 107 subjects, and the experimental results demonstrate that our method is superior to the state-of-the-art approaches in both tasks of bone segmentation and landmark digitization.

Tissue segmentation in whole-slide images is an important task in digital pathology, required for efficient and accurate computer-aided diagnostics. Precise tissue segmentation is particularly significant for a correct diagnosis in cases, when tissue structure of a specimen is very porous, such as skin specimens. In this paper, we addressed the problem of fore- and background segmentation in histopatological images of skin specimens stained with hematoxylin and eosin (H&E), which has not been solved yet, by a novel method based on a combination of statistical analysis, color thresholding, and binary morphology. We validated our algorithm on large extracts from 60 high-resolution whole slide images, with differing staining quality and captured under varying imaging conditions, from three laboratories. The size of extracts varies from 2000×1500 to 20000×30000 pixels and the number of images used in our study matches the number of H&E images used by other research teams. We compared our method to the published ones (GrabCut and FESI) and showed that our approach outperforms its counterparts (Jaccard index of 0.929 vs. 0.776 and 0.695).

Atomic force microscopy has a tremendous number of applications in a wide variety of fields, particularly in the semiconductor area for the 3D-stacked device. Imaging three-dimensional (3D) structures with blind features has progressively become a critical technique. Recently, a 3D-atomic force microscopy (AFM) technique has been proposed to image 3D features, especially those having sharp apices, like silicon pillars. However, the scanning strategy has drawbacks, such as long scanning time, and unstable operation, based on the premature algorithm. Herein, an improved 3D-AFM algorithm is reported that overcomes the aforementioned problems by an intelligent 3D scanning algorithm that incorporates sidewall history tracking, troubleshooting for sharp sidewall and sticking, and reactive direction adjustment. The proposed algorithm enables the 3D imagery of ZnO nano-rods and silicon nano-pillars to be achieved by using a high aspect-ratio multiwall carbon nanotube-based AFM probe, without time-consuming disorientation. This study establishes a method to construct a 3D image of arbitrary shape in reduced scanning time.

High-resolution three-dimensional imaging is key to our understanding of biological tissue formation and function. Recent developments in synchrotron-based X-Ray tomography techniques provide unprecedented morphological information on relatively large sample volumes with a spatial resolution better than 50 nm. However, the analysis of the generated data, in particular image segmentation - separation into structure and background - still present a significant challenge, especially when considering complex biomineralized structures that exhibit hierarchical arrangement of their constituents across many length scales - from millimeters down to nanometers. In the present work, synchrotron-based holographic nano-tomography data are combined with state-of-the-art machine learning methods to image and analyze the nacreous architecture in the bivalve Unio pictorum in 3D. Using kinetic and thermodynamic considerations known from physics of materials, the obtained spatial information is then used to provide a quantitative description of the structural and topological evolution of nacre during shell formation. Ultimately, this study establishes a workflow for high-resolution three-dimensional analysis of fine highly-mineralized biological tissues while providing a detailed analytical view on nacre morphogenesis.

When circularly polarized light interacts with a nanostructure, the optical response depends on the geometry of the structure. If the nanostructure is chiral (i.e. it cannot be superimposed on its mirror image) then its optical response, both in near-field and far-field, depends on the handedness of the incident light. In contrast, achiral structures exhibit identical far-field responses for left- and right-circular polarization. Here, we show that a perfectly achiral nanostructure, a plasmonic metamolecule with trigonal D3h symmetry, exhibits a near-field response that is sensitive to the handedness of light. This effect stems from the near-field interference between the different plasmonic modes sustained by the plasmonic metamolecule under circularly polarized light excitation. Moreover, the intensity of the local optical chirality can be tuned by the strength of the near-field coupling, as controlled by the gap distance between the individual meta-atoms. The local chirality in a plasmonic trimer is then experimentally evidenced with nanoscale resolution using a molecular probe. Our experiments demonstrate that the optical near-field chirality can be imprinted into the photosensitive polymer, turning the optical chirality into a geometrical chirality that can be imaged using atomic force microscopy. These results are of interest for the field of polarization-sensitive photochemistry.