scholarly journals Learning cellular morphology with neural networks

2018 ◽  
Author(s):  
Philipp J Schubert ◽  
Sven Dorkenwald ◽  
Michał Januszewski ◽  
Viren Jain ◽  
Joergen Kornfeld
Keyword(s):  
Neural Networks ◽  
Cell Types ◽  
Cellular Morphology ◽  
Data Sets ◽  
Neuron Reconstruction ◽  
Diverse Application ◽  
Unsupervised Training ◽  
Microscopy Data ◽  
Volume Electron Microscopy ◽  
Made In

AbstractReconstruction and annotation of volume electron microscopy data sets of brain tissue is challenging, but can reveal invaluable information about neuronal circuits. Significant progress has recently been made in automated neuron reconstruction, as well as automated detection of synapses. However, methods for automating the morphological analysis of nanometer-resolution reconstructions are less established, despite their diverse application possibilities. Here, we introduce cellular morphology neural networks (CMNs), based on multi-view projections sampled from automatically reconstructed cellular fragments of arbitrary size and shape. Using unsupervised training we inferred morphology embeddings (“Neuron2vec”) of neuron reconstructions and trained CMNs to identify glia cells in a supervised classification paradigm which was then used to resolve neuron reconstruction errors. Finally, we demonstrate that CMNs can be used to identify subcellular compartments and the cell types of neuron reconstructions.

scholarly journals Image Colorization Progress: A Review of Deep Learning Techniques for Automation of Colorization

2021 ◽  
Vol 10 (4) ◽  
pp. 2908-2915
Keyword(s):  
Neural Networks ◽  
Generative Models ◽  
Generative Adversarial Networks ◽  
Data Sets ◽  
Image Colorization ◽  
Black And White ◽  
The Past ◽  
Adversarial Networks ◽  
Learning Techniques ◽  
Made In

Image colorization is the process of taking an input gray- scale (black and white) image and then producing an output colorized image that represents the semantic color tones of the input. Since the past few years, the process of automatic image colorization has been of significant interest and a lot of progress has been made in the field by various researchers. Image colorization finds its application in many domains including medical imaging, restoration of historical documents, etc. There have been different approaches to solve this problem using Convolutional Neural Networks as well as Generative Adversarial Networks. These colorization networks are not only based on different architectures but also are tested on varied data sets. This paper aims to cover some of these proposed approaches through different techniques. The results between the generative models and traditional deep neural networks are compared along with presenting the current limitations in those. The paper proposes a summarized view of past and current advances in the field of image colorization contributed by different authors and researchers.

A Semi-Online Training Algorithm for the Radial Basis Function Neural Networks

2008 ◽  
pp. 243-260 ◽  
Author(s):  
Vadlamani Ravi ◽  
P. Ravi Kumar ◽  
Eruku Ravi Srinivas ◽  
Nikola K. Kasabov
Keyword(s):  
Neural Networks ◽  
Radial Basis Function ◽  
Basis Function ◽  
Clustering Algorithm ◽  
Ordinary Least Squares ◽  
Bankruptcy Prediction ◽  
Data Sets ◽  
Performance Metric ◽  
Radial Basis ◽  
Unsupervised Training

This chapter presents an algorithm to train radial basis function neural networks (RBFN) in a semi-online manner. It employs the online, evolving clustering algorithm of Kasabov and Song (2002) in the unsupervised training part of the RBFN and the ordinary least squares estimation technique for the supervised training part. Its effectiveness is demonstrated on two problems related to bankruptcy prediction in financial engineering. In all the cases, 10-fold cross validation was performed. The present algorithm, implemented in two variants, yielded more sensitivity compared to the multi layer perceptron trained by backpropagation (MLP) algorithm over all the problems studied. Based on the results, it can be inferred that the semi-online RBFN without linear terms is better than other neural network techniques. By taking the Area Under the ROC curve (AUC) as the performance metric, the proposed algorithms viz., semi-online RBFN with and without linear terms are compared with classifiers such as ANFIS, TreeNet, SVM, MLP, Linear RBF, RSES and Orthogonal RBF. Out of them TreeNet outperformed both the variants of the semi-online RBFN in both data sets considered here.

Data Management in an Epidemiological Study : Experiences from the Iowa 500 Field Follow-up and Family Study

1980 ◽  
Vol 19 (01) ◽  
pp. 37-41
Author(s):  
R. F. Woolson ◽  
M. T. Tsuang ◽  
L. R. Urban
Keyword(s):  
Data Management ◽  
Epidemiological Study ◽  
Family Study ◽  
Data Sets ◽  
Data Handling ◽  
Management Decisions ◽  
Numerous Data ◽  
Made In

We are now conducting a forty-year follow-up and family study of 200 schizophrenics, 325 manic-depressives and 160 surgical controls. This study began in 1973 and has continued to the present date. Numerous data handling and data management decisions were made in the course of collecting the data for the project. In this report some of the practical difficulties in the data handling and computer management of such large and bulky data sets are enumerated.

scholarly journals Generation of geometric interpolations of building types with deep variational autoencoders

2020 ◽  
Vol 6 ◽  
Author(s):  
Jaime de Miguel Rodríguez ◽  
Maria Eugenia Villafañe ◽  
Luka Piškorec ◽  
Fernando Sancho Caparrini
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Large Data ◽  
Learning Model ◽  
Large Data Sets ◽  
Data Sets ◽  
Connectivity Map ◽  
Data Set ◽  
3D Objects ◽  
Machine Learning Model

Abstract This work presents a methodology for the generation of novel 3D objects resembling wireframes of building types. These result from the reconstruction of interpolated locations within the learnt distribution of variational autoencoders (VAEs), a deep generative machine learning model based on neural networks. The data set used features a scheme for geometry representation based on a ‘connectivity map’ that is especially suited to express the wireframe objects that compose it. Additionally, the input samples are generated through ‘parametric augmentation’, a strategy proposed in this study that creates coherent variations among data by enabling a set of parameters to alter representative features on a given building type. In the experiments that are described in this paper, more than 150 k input samples belonging to two building types have been processed during the training of a VAE model. The main contribution of this paper has been to explore parametric augmentation for the generation of large data sets of 3D geometries, showcasing its problems and limitations in the context of neural networks and VAEs. Results show that the generation of interpolated hybrid geometries is a challenging task. Despite the difficulty of the endeavour, promising advances are presented.

Development of convolutional neural networks for recognition of tenogenic differentiation based on cellular morphology

2021 ◽  
pp. 106279
Author(s):  
Gözde Dursun ◽  
Saurabh Balkrishna Tandale ◽  
Rutwik Gulakala ◽  
Jörg Eschweiler ◽  
Mersedeh Tohidnezhad ◽  
...  
Keyword(s):  
Neural Networks ◽  
Convolutional Neural Networks ◽  
Cellular Morphology ◽  
Tenogenic Differentiation

scholarly journals Profiling Bioactivity of the ToxCast Chemical Library Using BioMAP Primary Human Cell Systems

2009 ◽  
Vol 14 (9) ◽  
pp. 1054-1066 ◽  
Author(s):  
Keith A. Houck ◽  
David J. Dix ◽  
Richard S. Judson ◽  
Robert J. Kavlock ◽  
Jian Yang ◽  
...  
Keyword(s):  
Human Cell ◽  
Regulatory Networks ◽  
Cell Types ◽  
Environmental Chemicals ◽  
Data Sets ◽  
Chemical Library ◽  
Data Set ◽  
Cell Systems ◽  
Disease Biology ◽  
Nfκb Pathway

The complexity of human biology has made prediction of health effects as a consequence of exposure to environmental chemicals especially challenging. Complex cell systems, such as the Biologically Multiplexed Activity Profiling (BioMAP) primary, human, cell-based disease models, leverage cellular regulatory networks to detect and distinguish chemicals with a broad range of target mechanisms and biological processes relevant to human toxicity. Here the authors use the BioMAP human cell systems to characterize effects relevant to human tissue and inflammatory disease biology following exposure to the 320 environmental chemicals in the Environmental Protection Agency’s (EPA’s) ToxCast phase I library. The ToxCast chemicals were assayed at 4 concentrations in 8 BioMAP cell systems, with a total of 87 assay endpoints resulting in more than 100,000 data points. Within the context of the BioMAP database, ToxCast compounds could be classified based on their ability to cause overt cytotoxicity in primary human cell types or according to toxicity mechanism class derived from comparisons to activity profiles of BioMAP reference compounds. ToxCast chemicals with similarity to inducers of mitochondrial dysfunction, cAMP elevators, inhibitors of tubulin function, inducers of endoplasmic reticulum stress, or NFκB pathway inhibitors were identified based on this BioMAP analysis. This data set is being combined with additional ToxCast data sets for development of predictive toxicity models at the EPA. ( Journal of Biomolecular Screening 2009:1054-1066)

scholarly journals Chromatin interactions reveal novel gene targets for drug repositioning in rheumatic diseases

2019 ◽  
Vol 78 (8) ◽  
pp. 1127-1134 ◽  
Author(s):  
Paul Martin ◽  
James Ding ◽  
Kate Duffus ◽  
Vasanthi Priyadarshini Gaddi ◽  
Amanda McGovern ◽  
...  
Keyword(s):  
Rheumatic Diseases ◽  
Pathway Analysis ◽  
Drug Targets ◽  
Drug Repositioning ◽  
Healthcare Providers ◽  
Cell Types ◽  
Data Sets ◽  
High Confidence ◽  
Causal Genes ◽  
Additional Cell

ObjectivesThere is a need to identify effective treatments for rheumatic diseases, and while genetic studies have been successful it is unclear which genes contribute to the disease. Using our existing Capture Hi-C data on three rheumatic diseases, we can identify potential causal genes which are targets for existing drugs and could be repositioned for use in rheumatic diseases.MethodsHigh confidence candidate causal genes were identified using Capture Hi-C data from B cells and T cells. These genes were used to interrogate drug target information from DrugBank to identify existing treatments, which could be repositioned to treat these diseases. The approach was refined using Ingenuity Pathway Analysis to identify enriched pathways and therefore further treatments relevant to the disease.ResultsOverall, 454 high confidence genes were identified. Of these, 48 were drug targets (108 drugs) and 11 were existing therapies used in the treatment of rheumatic diseases. After pathway analysis refinement, 50 genes remained, 13 of which were drug targets (33 drugs). However considering targets across all enriched pathways, a further 367 drugs were identified for potential repositioning.ConclusionCapture Hi-C has the potential to identify therapies which could be repositioned to treat rheumatic diseases. This was particularly successful for rheumatoid arthritis, where six effective, biologic treatments were identified. This approach may therefore yield new ways to treat patients, enhancing their quality of life and reducing the economic impact on healthcare providers. As additional cell types and other epigenomic data sets are generated, this prospect will improve further.

scholarly journals Moving Forward With Accelerometer-Assessed Physical Activity: Two Strategies to Ensure Meaningful, Interpretable, and Comparable Measures

2018 ◽  
Vol 30 (4) ◽  
pp. 450-456 ◽  
Author(s):  
Alex V. Rowlands
Keyword(s):  
Physical Activity ◽  
Black Box ◽  
Data Sets ◽  
Accelerometer Data ◽  
Future Health ◽  
The Past ◽  
Acceleration Data ◽  
Night And Day ◽  
Very High ◽  
Made In

Significant advances have been made in the measurement of physical activity in youth over the past decade. Monitors and protocols promote very high compliance, both night and day, and raw measures are available rather than “black box” counts. Consequently, many surveys and studies worldwide now assess children’s physical behaviors (physical activity, sedentary behavior, and sleep) objectively 24 hours a day, 7 days a week using accelerometers. The availability of raw acceleration data in many of these studies is both an opportunity and a challenge. The richness of the data lends itself to the continued development of innovative metrics, whereas the removal of proprietary outcomes offers considerable potential for comparability between data sets and harmonizing data. Using comparable physical activity outcomes could lead to improved precision and generalizability of recommendations for children’s present and future health. The author will discuss 2 strategies that he believes may help ensure comparability between studies and maximize the potential for data harmonization, thereby helping to capitalize on the growing body of accelerometer data describing children’s physical behaviors.

scholarly journals Pharmacologically targeting tribbles gene expression in colorectal cancer

2021 ◽  
Vol 31 (Supplement_2) ◽  
Author(s):  
Victor Yassuda ◽  
Ana Luísa De Sousa-Coelho
Keyword(s):  
Gene Expression ◽  
Colorectal Cancer ◽  
Cell Types ◽  
Disease Recurrence ◽  
Rosemary Extract ◽  
Acid Oxidation ◽  
Data Sets ◽  
Compensatory Mechanism ◽  
Dependent Manner ◽  
Pharmacological Potential

Abstract Background TRIB1, TRIB2 and TRIB3 belong to the mammalian Tribbles family of pseudokinases proteins. Several studies reported Tribbles oncogenic role in different types of cancer, including colorectal cancer (CRC). Though current CRC treatment can be curative, patients are in risk of disease recurrence, meaning novel pharmacological targets and strategies are required. Our goal was to analyze Tribbles gene expression in CRC in response to different drugs. Methods Tribbles transcript levels were obtained from GEO profiles database (NCBI). Gene data sets (GDS) were selected based on experimental drug treatment description. Statistical analysis was performed at GraphPadPrism. Results Compared to non-treated control, TRIB2 expression was ∼2-fold increased in colorectal adenocarcinoma samples from patients treated with cyclooxygenase-2 inhibitor celecoxib (GDS3384), though not statistically significant (P < 0.1). TRIB1 was unaltered and data for TRIB3 was not available. By contrast, all Tribbles showed differential expression after treatment of SW620 colon cancer cells with supercritical rosemary extract in progressive increasing doses (0, 30, 60, 100 μg/mL) (P < 0.01;GDS5416). While both TRIB1 and TRIB3 were moderately increased in a dose-dependent manner (∼18% and 13%, respectively), TRIB2 was maximally down-regulated by ∼15% after 60 μg/mL. Conclusions Although celecoxib exhibits antiproliferative effects in different cancer cell types, TRIB2 gene expression showed a trend to be induced after treatment, in contrast to several genes involved in fatty acid oxidation that were down-regulated, which could result from a compensatory mechanism based on a metabolic shift. Since TRIB1/TRIB3 and TRIB2 were oppositely modulated in response to rosemary extract, additional studies are needed to validate its specific pharmacological potential interest for CRC treatment.

Sign in / Sign up

Export Citation Format

Share Document