Finding and following: a deep learning-based pipeline for tracking platelets during thrombus formation in vivo and ex vivo.

The last decade has seen increasing use of advanced imaging techniques in platelet research. However, there has been a lag in the development of image analysis methods, leaving much of the information trapped in images. Herein, we present a robust analytical pipeline for finding and following individual platelets over time in growing thrombi. Our pipeline covers four steps: detection, tracking, estimation of tracking accuracy, and quantification of platelet metrics. We detect platelets using a deep learning network for image segmentation, which we validated with proofreading by multiple experts. We then track platelets using a standard particle tracking algorithm and validate the tracks with custom image sampling - essential when following platelets within a dense thrombus. We show that our pipeline is more accurate than previously described methods. To demonstrate the utility of our analytical platform, we use it to show that in vivo thrombus formation is much faster than that ex vivo. Furthermore, platelets in vivo exhibit less passive movement in the direction of blood flow. Our tools are free and open source and written in the popular and user-friendly Python programming language. They empower researchers to accurately find and follow platelets in fluorescence microscopy experiments.

In this paper we describe computational tools to find and follow individual platelets in blood clots recorded with fluorescence microscopy. Our tools work in a diverse range of conditions, both in living animals and in artificial flow chamber models of thrombosis. Our work uses deep learning methods to achieve excellent accuracy. We also provide tools for visualizing data and estimating error rates, so you don't have to just trust the output. Our workflow measures platelet density, shape, and speed, which we use to demonstrate differences in the kinetics of clotting in living vessels versus a synthetic environment. The tools we wrote are open source, written in the popular Python programming language, and freely available to all. We hope they will be of use to other platelet researchers.

The knob protein KAHRP assembles into a ring-shaped structure that underpins virulence complex assembly.

Plasmodium falciparum mediates adhesion of infected red blood cells (RBCs) to blood vessel walls by assembling a multi-protein complex at the RBC surface. This virulence-mediating structure, called the knob, acts as a scaffold for the presentation of the major virulence antigen, P. falciparum Erythrocyte Membrane Protein-1 (PfEMP1). In this work we developed correlative STochastic Optical Reconstruction Microscopy-Scanning Electron Microscopy (STORM-SEM) to spatially and temporally map the delivery of the knob-associated histidine-rich protein (KAHRP) and PfEMP1 to the RBC membrane skeleton. We show that KAHRP is delivered as individual modules that assemble in situ, giving a ring-shaped fluorescence profile around a dimpled disk that can be visualized by SEM. Electron tomography of negatively-stained membranes reveals a previously observed spiral scaffold underpinning the assembled knobs. Truncation of the C-terminal region of KAHRP leads to loss of the ring structures, disruption of the raised disks and aberrant formation of the spiral scaffold, pointing to a critical role for KAHRP in assembling the physical knob structure. We show that host cell actin remodeling plays an important role in assembly of the virulence complex, with cytochalasin D blocking knob assembly. Additionally, PfEMP1 appears to be delivered to the RBC membrane, then inserted laterally into knob structures.

Aurora Kinase B, a novel regulator of TERF1 binding and telomeric integrity.

AURKB (Aurora Kinase B) is a serine/threonine kinase better known for its role at the mitotic kinetochore during chromosome segregation. Here, we demonstrate that AURKB localizes to the telomeres in mouse embryonic stem cells, where it interacts with the essential telomere protein TERF1. Loss of AURKB function affects TERF1 telomere binding and results in aberrant telomere structure. In vitro kinase experiments successfully identified Serine 404 on TERF1 as a putative AURKB target site. Importantly, in vivo overexpression of S404-TERF1 mutants results in fragile telomere formation. These findings demonstrate that AURKB is an important regulator of telomere structural integrity.

Synaptic circuits and their variations within different columns in the visual system of Drosophila.

We reconstructed the synaptic circuits of seven columns in the second neuropil or medulla behind the fly's compound eye. These neurons embody some of the most stereotyped circuits in one of the most miniaturized of animal brains. The reconstructions allow us, for the first time to our knowledge, to study variations between circuits in the medulla's neighboring columns. This variation in the number of synapses and the types of their synaptic partners has previously been little addressed because methods that visualize multiple circuits have not resolved detailed connections, and existing connectomic studies, which can see such connections, have not so far examined multiple reconstructions of the same circuit. Here, we address the omission by comparing the circuits common to all seven columns to assess variation in their connection strengths and the resultant rates of several different and distinct types of connection error. Error rates reveal that, overall, <1% of contacts are not part of a consensus circuit, and we classify those contacts that supplement (E+) or are missing from it (E-). Autapses, in which the same cell is both presynaptic and postsynaptic at the same synapse, are occasionally seen; two cells in particular, Dm9 and Mi1, form ≥ 20-fold more autapses than do other neurons. These results delimit the accuracy of developmental events that establish and normally maintain synaptic circuits with such precision, and thereby address the operation of such circuits. They also establish a precedent for error rates that will be required in the new science of connectomics.

Identification of a Stress-Sensitive Anorexigenic Neurocircuit From Medial Prefrontal Cortex to Lateral Hypothalamus.

BACKGROUND: A greater understanding of how the brain controls appetite is fundamental to developing new approaches for treating diseases characterized by dysfunctional feeding behavior, such as obesity and anorexia nervosa. METHODS: By modeling neural network dynamics related to homeostatic state and body mass index, we identified a novel pathway projecting from the medial prefrontal cortex (mPFC) to the lateral hypothalamus (LH) in humans (n = 53). We then assessed the physiological role and dissected the function of this mPFC-LH circuit in mice. RESULTS: In vivo recordings of population calcium activity revealed that this glutamatergic mPFC-LH pathway is activated in response to acute stressors and inhibited during food consumption, suggesting a role in stress-related control over food intake. Consistent with this role, inhibition of this circuit increased feeding and sucrose seeking during mild stressors, but not under nonstressful conditions. Finally, chemogenetic or optogenetic activation of the mPFC-LH pathway is sufficient to suppress food intake and sucrose seeking in mice. CONCLUSIONS: These studies identify a glutamatergic mPFC-LH circuit as a novel stress-sensitive anorexigenic neural pathway involved in the cortical control of food intake.

In Vivo Photometry Reveals Insulin and 2-Deoxyglucose Maintain Prolonged Inhibition of VMH Vglut2 Neurons in Male Mice.

The ventromedial hypothalamic (VMH) nucleus is a well-established hub for energy and glucose homeostasis. In particular, VMH neurons are thought to be important for initiating the counterregulatory response to hypoglycemia, and ex vivo electrophysiology and immunohistochemistry data indicate a clear role for VMH neurons in sensing glucose concentration. However, the temporal response of VMH neurons to physiologically relevant changes in glucose availability in vivo has been hampered by a lack of available tools for measuring neuronal activity over time. Since the majority of neurons within the VMH are glutamatergic and can be targeted using the vesicular glutamate transporter Vglut2, we expressed cre-dependent GCaMP7s in Vglut2 cre mice and examined the response profile of VMH to intraperitoneal injections of glucose, insulin, and 2-deoxyglucose (2DG). We show that reduced available glucose via insulin-induced hypoglycemia and 2DG-induced glucoprivation, but not hyperglycemia induced by glucose injection, inhibits VMH Vglut2 neuronal population activity in vivo. Surprisingly, this inhibition was maintained for at least 45 minutes despite prolonged hypoglycemia and initiation of a counterregulatory response. Thus, although VMH stimulation, via pharmacological, electrical, or optogenetic approaches, is sufficient to drive a counterregulatory response, our data suggest VMH Vglut2 neurons are not the main drivers required to do so, since VMH Vglut2 neuronal population activity remains suppressed during hypoglycemia and glucoprivation.

Multimodal imaging reveals membrane skeleton reorganisation during reticulocyte maturation and differences in dimple and rim regions of mature erythrocytes.

The red blood cell (RBC) is remarkable in its ability to deform as it passages through the vasculature. Its deformability derives from a spectrin-actin protein network that supports the cell membrane and provides strength and flexibility, however questions remain regarding the assembly and maintenance of the skeletal network. Using scanning electron microscopy (SEM) and atomic force microscopy (AFM) we have examined the nanoscale architecture of the cytoplasmic side of membrane discs prepared from reticulocytes and mature RBCs. Immunofluorescence microscopy was used to probe the distribution of spectrin and other membrane skeleton proteins. We found that the cell surface area decreases by up to 30% and the spectrin-actin network increases in density by approximately 20% as the reticulocyte matures. By contrast, the inter-junctional distance and junctional density increase only by 3-4% and 5-9%, respectively. This suggests that the maturation-associated reduction in surface area is accompanied by an increase in spectrin self-association to form higher order oligomers. We also examined the mature RBC membrane in the edge (rim) and face (dimple) regions of mature RBCs and found the rim contains about 1.5% more junctional complexes compared to the dimple region. A 2% increase in band 4.1 density in the rim supports these structural measurements.

Metabolic sensing in AgRP neurons integrates homeostatic state with dopamine signalling in the striatum.

Agouti-related peptide (AgRP) neurons increase motivation for food, however, whether metabolic sensing of homeostatic state in AgRP neurons potentiates motivation by interacting with dopamine reward systems is unexplored. As a model of impaired metabolic-sensing, we used the AgRP-specific deletion of carnitine acetyltransferase (Crat) in mice. We hypothesised that metabolic sensing in AgRP neurons is required to increase motivation for food reward by modulating accumbal or striatal dopamine release. Studies confirmed that Crat deletion in AgRP neurons (KO) impaired ex vivo glucose-sensing, as well as in vivo responses to peripheral glucose injection or repeated palatable food presentation and consumption. Impaired metabolic-sensing in AgPP neurons reduced acute dopamine release (seconds) to palatable food consumption and during operant responding, as assessed by GRAB-DA photometry in the nucleus accumbens, but not the dorsal striatum. Impaired metabolic-sensing in AgRP neurons suppressed radiolabelled 18F-fDOPA accumulation after ~30 min in the dorsal striatum but not the nucleus accumbens. Impaired metabolic sensing in AgRP neurons suppressed motivated operant responding for sucrose rewards during fasting. Thus, metabolic-sensing in AgRP neurons is required for the appropriate temporal integration and transmission of homeostatic hunger-sensing to dopamine signalling in the striatum.

An integrative modular approach to systematically predict gene-phenotype associations.

BACKGROUND: Complex human diseases are often caused by multiple mutations, each of which contributes only a minor effect to the disease phenotype. To study the basis for these complex phenotypes, we developed a network-based approach to identify coexpression modules specifically activated in particular phenotypes. We integrated these modules, protein-protein interaction data, Gene Ontology annotations, and our database of gene-phenotype associations derived from literature to predict novel human gene-phenotype associations. Our systematic predictions provide us with the opportunity to perform a global analysis of human gene pleiotropy and its underlying regulatory mechanisms. RESULTS: We applied this method to 338 microarray datasets, covering 178 phenotype classes, and identified 193,145 phenotype-specific coexpression modules. We trained random forest classifiers for each phenotype and predicted a total of 6,558 gene-phenotype associations. We showed that 40.9% genes are pleiotropic, highlighting that pleiotropy is more prevalent than previously expected. We collected 77 ChIP-chip datasets studying 69 transcription factors binding over 16,000 targets under various phenotypic conditions. Utilizing this unique data source, we confirmed that dynamic transcriptional regulation is an important force driving the formation of phenotype specific gene modules. CONCLUSION: We created a genome-wide gene to phenotype mapping that has many potential implications, including providing potential new drug targets and uncovering the basis for human disease phenotypes. Our analysis of these phenotype-specific coexpression modules reveals a high prevalence of gene pleiotropy, and suggests that phenotype-specific transcription factor binding may contribute to phenotypic diversity. All resources from our study are made freely available on our online Phenotype Prediction Database.

Usefulness and limitations of dK random graph models to predict interactions and functional homogeneity in biological networks under a pseudo-likelihood parameter estimation approach.

BACKGROUND: Many aspects of biological functions can be modeled by biological networks, such as protein interaction networks, metabolic networks, and gene coexpression networks. Studying the statistical properties of these networks in turn allows us to infer biological function. Complex statistical network models can potentially more accurately describe the networks, but it is not clear whether such complex models are better suited to find biologically meaningful subnetworks. RESULTS: Recent studies have shown that the degree distribution of the nodes is not an adequate statistic in many molecular networks. We sought to extend this statistic with 2nd and 3rd order degree correlations and developed a pseudo-likelihood approach to estimate the parameters. The approach was used to analyze the MIPS and BIOGRID yeast protein interaction networks, and two yeast coexpression networks. We showed that 2nd order degree correlation information gave better predictions of gene interactions in both protein interaction and gene coexpression networks. However, in the biologically important task of predicting functionally homogeneous modules, degree correlation information performs marginally better in the case of the MIPS and BIOGRID protein interaction networks, but worse in the case of gene coexpression networks. CONCLUSION: Our use of dK models showed that incorporation of degree correlations could increase predictive power in some contexts, albeit sometimes marginally, but, in all contexts, the use of third-order degree correlations decreased accuracy. However, it is possible that other parameter estimation methods, such as maximum likelihood, will show the usefulness of incorporating 2nd and 3rd degree correlations in predicting functionally homogeneous modules.

Gene Aging Nexus: a web database and data mining platform for microarray data on aging.

The recent development of microarray technology provided unprecedented opportunities to understand the genetic basis of aging. So far, many microarray studies have addressed aging-related expression patterns in multiple organisms and under different conditions. The number of relevant studies continues to increase rapidly. However, efficient exploitation of these vast data is frustrated by the lack of an integrated data mining platform or other unifying bioinformatic resource to enable convenient cross-laboratory searches of array signals. To facilitate the integrative analysis of microarray data on aging, we developed a web database and analysis platform 'Gene Aging Nexus' (GAN) that is freely accessible to the research community to query/analyze/visualize cross-platform and cross-species microarray data on aging. By providing the possibility of integrative microarray analysis, GAN should be useful in building the systems-biology understanding of aging. GAN is accessible at http://gan.usc.edu.

An integrative approach to characterize disease-specific pathways and their coordination: a case study in cancer.

BACKGROUND: The most common application of microarray technology in disease research is to identify genes differentially expressed in disease versus normal tissues. However, it is known that, in complex diseases, phenotypes are determined not only by genes, but also by the underlying structure of genetic networks. Often, it is the interaction of many genes that causes phenotypic variations. RESULTS: In this work, using cancer as an example, we develop graph-based methods to integrate multiple microarray datasets to discover disease-related co-expression network modules. We propose an unsupervised method that take into account both co-expression dynamics and network topological information to simultaneously infer network modules and phenotype conditions in which they are activated or de-activated. Using our method, we have discovered network modules specific to cancer or subtypes of cancers. Many of these modules are consistent with or supported by their functional annotations or their previously known involvement in cancer. In particular, we identified a module that is predominately activated in breast cancer and is involved in tumor suppression. While individual components of this module have been suggested to be associated with tumor suppression, their coordinated function has never been elucidated. Here by adopting a network perspective, we have identified their interrelationships and, particularly, a hub gene PDGFRL that may play an important role in this tumor suppressor network. CONCLUSION: Using a network-based approach, our method provides new insights into the complex cellular mechanisms that characterize cancer and cancer subtypes. By incorporating co-expression dynamics information, our approach can not only extract more functionally homogeneous modules than those based solely on network topology, but also reveal pathway coordination beyond co-expression.

Functional annotation and network reconstruction through cross-platform integration of microarray data.

The rapid accumulation of microarray data translates into a need for methods to effectively integrate data generated with different platforms. Here we introduce an approach, 2(nd)-order expression analysis, that addresses this challenge by first extracting expression patterns as meta-information from each data set (1(st)-order expression analysis) and then analyzing them across multiple data sets. Using yeast as a model system, we demonstrate two distinct advantages of our approach: we can identify genes of the same function yet without coexpression patterns and we can elucidate the cooperativities between transcription factors for regulatory network reconstruction by overcoming a key obstacle, namely the quantification of activities of transcription factors. Experiments reported in the literature and performed in our lab support a significant number of our predictions.

A new Python library to analyse skeleton images confirms malaria parasite remodelling of the red blood cell membrane skeleton.

We present Skan (Skeleton analysis), a Python library for the analysis of the skeleton structures of objects. It was inspired by the "analyse skeletons" plugin for the Fiji image analysis software, but its extensive Application Programming Interface (API) allows users to examine and manipulate any intermediate data structures produced during the analysis. Further, its use of common Python data structures such as SciPy sparse matrices and pandas data frames opens the results to analysis within the extensive ecosystem of scientific libraries available in Python. We demonstrate the validity of Skan's measurements by comparing its output to the established Analyze Skeletons Fiji plugin, and, with a new scanning electron microscopy (SEM)-based method, we confirm that the malaria parasite Plasmodium falciparum remodels the host red blood cell cytoskeleton, increasing the average distance between spectrin-actin junctions.

Analyzing microtomography data with Python and the scikit-image library.

The exploration and processing of images is a vital aspect of the scientific workflows of many X-ray imaging modalities. Users require tools that combine interactivity, versatility, and performance. scikit-image is an open-source image processing toolkit for the Python language that supports a large variety of file formats and is compatible with 2D and 3D images. The toolkit exposes a simple programming interface, with thematic modules grouping functions according to their purpose, such as image restoration, segmentation, and measurements. scikit-image users benefit from a rich scientific Python ecosystem that contains many powerful libraries for tasks such as visualization or machine learning. scikit-image combines a gentle learning curve, versatile image processing capabilities, and the scalable performance required for the high-throughput analysis of X-ray imaging data.

Graph-based active learning of agglomeration (GALA): a Python library to segment 2D and 3D neuroimages.

The aim in high-resolution connectomics is to reconstruct complete neuronal connectivity in a tissue. Currently, the only technology capable of resolving the smallest neuronal processes is electron microscopy (EM). Thus, a common approach to network reconstruction is to perform (error-prone) automatic segmentation of EM images, followed by manual proofreading by experts to fix errors. We have developed an algorithm and software library to not only improve the accuracy of the initial automatic segmentation, but also point out the image coordinates where it is likely to have made errors. Our software, called gala (graph-based active learning of agglomeration), improves the state of the art in agglomerative image segmentation. It is implemented in Python and makes extensive use of the scientific Python stack (numpy, scipy, networkx, scikit-learn, scikit-image, and others). We present here the software architecture of the gala library, and discuss several designs that we consider would be generally useful for other segmentation packages. We also discuss the current limitations of the gala library and how we intend to address them.

scikit-image: image processing in Python.

scikit-image is an image processing library that implements algorithms and utilities for use in research, education and industry applications. It is released under the liberal Modified BSD open source license, provides a well-documented API in the Python programming language, and is developed by an active, international team of collaborators. In this paper we highlight the advantages of open source to achieve the goals of the scikit-image library, and we showcase several real-world image processing applications that use scikit-image. More information can be found on the project homepage, http://scikit-image.org.

Machine learning of hierarchical clustering to segment 2D and 3D images.

We aim to improve segmentation through the use of machine learning tools during region agglomeration. We propose an active learning approach for performing hierarchical agglomerative segmentation from superpixels. Our method combines multiple features at all scales of the agglomerative process, works for data with an arbitrary number of dimensions, and scales to very large datasets. We advocate the use of variation of information to measure segmentation accuracy, particularly in 3D electron microscopy (EM) images of neural tissue, and using this metric demonstrate an improvement over competing algorithms in EM and natural images.

Electron Microscopy Reconstruction of Brain Structure Using Sparse Representations Over Learned Dictionaries.

A central problem in neuroscience is reconstructing neuronal circuits on the synapse level. Due to a wide range of scales in brain architecture such reconstruction requires imaging that is both high-resolution and high-throughput. Existing electron microscopy (EM) techniques possess required resolution in the lateral plane and either high-throughput or high depth resolution but not both. Here, we exploit recent advances in unsupervised learning and signal processing to obtain high depth-resolution EM images computationally without sacrificing throughput. First, we show that the brain tissue can be represented as a sparse linear combination of localized basis functions that are learned using high-resolution datasets. We then develop compressive sensing-inspired techniques that can reconstruct the brain tissue from very few (typically five) tomographic views of each section. This enables tracing of neuronal processes and, hence, high throughput reconstruction of neural circuits on the level of individual synapses.

Betamax: towards optimal sampling strategies for high-throughput screens.

Sample size is a critical component in the design of any high-throughput genetic screening approach. Sample size determination from assumptions or limited data at the planning stages, though standard practice, may at times be unreliable because of the difficulty of a priori modeling of effect sizes and variance. Methods to update the sample size estimate during the course of the study could improve statistical power. In this article, we introduce an approach to estimate the power and update it continuously during the screen. We use this estimate to decide where to sample next to achieve maximum overall statistical power. Finally, in simulations, we demonstrate significant gains in study recall over the naive strategy of equal sample sizes while maintaining the same total number of samples.