Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “science data network”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,747 records · Page 97

Optimal Bayesian supervised domain adaptation for RNA sequencing data

Abstract Motivation When learning to subtype complex disease based on next-generation sequencing data, the amount of available data is often limited. Recent works have tried to leverage data from other domains to design better predictors in the target domain of interest with varying degrees of success. But they are either limited to the cases requiring the outcome label correspondence across domains or cannot leverage the label information at all. Moreover, the existing methods cannot usually benefit from other information available a priori such as gene interaction networks. Results In this article, we develop a generative optimal Bayesian supervised domain adaptation (OBSDA) model that can integrate RNA sequencing (RNA-Seq) data from different domains along with their labels for improving prediction accuracy in the target domain. Our model can be applied in cases where different domains share the same labels or have different ones. OBSDA is based on a hierarchical Bayesian negative binomial model with parameter factorization, for which the optimal predictor can be derived by marginalization of likelihood over the posterior of the parameters. We first provide an efficient Gibbs sampler for parameter inference in OBSDA. Then, we leverage the gene-gene network prior information and construct an informed and flexible variational family to infer the posterior distributions of model parameters. Comprehensive experiments on real-world RNA-Seq data demonstrate the superior performance of OBSDA, in terms of accuracy in identifying cancer subtypes by utilizing data from different domains. Moreover, we show that by taking advantage of the prior network information we can further improve the performance. Availability and implementation The source code for implementations of OBSDA and SI-OBSDA are available at the following link. https://github.com/SHBLK/BSDA. Supplementary information Supplementary data are available at Bioinformatics online.

Biochemistry & Molecular Biology↗

Upward propagation of gigantic jets revealed by 3D radio and optical mapping

Occasionally, lightning will exit the top of a thunderstorm and connect to the lower edge of space, forming a gigantic jet. Here, we report on observations of a negative gigantic jet that transferred an extraordinary amount of charge between the troposphere and ionosphere (~300 C). It occurred in unusual circumstances, emerging from an area of weak convection. As the discharge ascended from the cloud top, tens of very high frequency (VHF) radio sources were detected from 22 to 45 km altitude, while simultaneous optical emissions (777.4 nm OI emitted from lightning leaders) remained near cloud top (15 to 20 km altitude). This implies that the high-altitude VHF sources were produced by streamers and the streamer discharge activity can extend all the way from near cloud top to the ionosphere. The simultaneous three-dimensional radio and optical data indicate that VHF lightning networks detect emissions from streamer corona rather than the leader channel, which has broad implications to lightning physics beyond that of gigantic jets.

54 ENVIRONMENTAL SCIENCES↗

Protein Co-Expression Network Analysis (ProCoNA)

Biological networks are important for elucidating disease etiology due to their ability to model complex high dimensional data and biological systems. Proteomics provides a critical data source for such models, but currently lacks robust de novo methods for network construction, which could bring important insights in systems biology. We have evaluated the construction of network models using methods derived from weighted gene co-expression network analysis (WGCNA). We show that approximately scale-free peptide networks, composed of statistically significant modules, are feasible and biologically meaningful using two mouse lung experiments and one human plasma experiment. Within each network, peptides derived from the same protein are shown to have a statistically higher topological overlap and concordance in abundance, which is potentially important for inferring protein abundance. The module representatives, called eigenpeptides, correlate significantly with biological phenotypes. Furthermore, within modules, we find significant enrichment for biological function and known interactions (gene ontology and protein-protein interactions). Biological networks are important tools in the analysis of complex systems. In this paper we evaluate the application of weighted co-expression network analysis to quantitative proteomics data. Protein co-expression networks allow novel approaches for biological interpretation, quality control, inference of protein abundance, a framework for potentially resolving degenerate peptide-protein mappings, and a biomarker signature discovery.

Environmental Molecular Sciences Laboratory↗

My Virtual Cancer

Since each cancer has its own unique characteristics, each one can respond differently to the same treatments. Therefore, the creation of a digital twin (DT) of cancer can assist us in predicting the evolution of an individual's cancer through modeling each tumor's characteristics and response to treatment. Hence, we propose to take advantage of new advances in computational approaches and combine mechanistic, machine learning, and stochastic modeling approaches to create “My Virtual Cancer", a DT platform. To establish a personalized DT, we use patient-specific data for parameter estimations, sensitivity analysis, and uncertainty quantification. For each patient, we will estimate the values of parameters of their QSP model using the patient's data. We perform a multi-dimensional sensitivity analysis and uncertainty quantification on the mechanistic model to find a set of critical interactions and predict the intervals of confidence. Since this QSP model includes the data-driven mechanistic model of cells and molecules' interaction networks, one of the ultimate results of this DT would be the prediction of evolution of tumors.

60 APPLIED LIFE SCIENCES↗

Deep active learning for classifying cancer pathology reports

Abstract Background Automated text classification has many important applications in the clinical setting; however, obtaining labelled data for training machine learning and deep learning models is often difficult and expensive. Active learning techniques may mitigate this challenge by reducing the amount of labelled data required to effectively train a model. In this study, we analyze the effectiveness of 11 active learning algorithms on classifying subsite and histology from cancer pathology reports using a Convolutional Neural Network as the text classification model. Results We compare the performance of each active learning strategy using two differently sized datasets and two different classification tasks. Our results show that on all tasks and dataset sizes, all active learning strategies except diversity-sampling strategies outperformed random sampling, i.e., no active learning. On our large dataset (15K initial labelled samples, adding 15K additional labelled samples each iteration of active learning), there was no clear winner between the different active learning strategies. On our small dataset (1K initial labelled samples, adding 1K additional labelled samples each iteration of active learning), marginal and ratio uncertainty sampling performed better than all other active learning techniques. We found that compared to random sampling, active learning strongly helps performance on rare classes by focusing on underrepresented classes. Conclusions Active learning can save annotation cost by helping human annotators efficiently and intelligently select which samples to label. Our results show that a dataset constructed using effective active learning techniques requires less than half the amount of labelled data to achieve the same performance as a dataset constructed using random sampling.

59 BASIC BIOLOGICAL SCIENCES↗

The bond rupture force for sulfur chains calculated from quantum chemistry simulations and its relevance to the tensile strength of vulcanized rubber

From quantum chemistry simulations using density functional theory, we obtain the total electronic energy of an eight-atom sulfur chain as its end-to-end distance is extended until S–S bond rupture occurs. We find that a sulfur chain can be extended by about 40% beyond its nominally straight conformation, where it experiences rupture at an end-to-end tension of about 1.5 nN. Using this rupture force as the chain failure limit in an explicit polymer network simulation model (EPnet), we predict the tensile failure stress for sulfur crosslinked (vulcanized) natural rubber. Furthermore, quantitative agreement with published experimental data for the failure stress is obtained in these simulations if we assume that only about 30% of the sulfur chains produce viable network crosslinks. Surprisingly, we also find that the failure stress of a rubber network does not scale linearly with the chain failure force limit.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Multiscale Flow for robust and optimal cosmological analysis

We propose Multiscale Flow, a generative Normalizing Flow that creates samples and models the field-level likelihood of two-dimensional cosmological data such as weak lensing. Multiscale Flow uses hierarchical decomposition of cosmological fields via a wavelet basis and then models different wavelet components separately as Normalizing Flows. The log-likelihood of the original cosmological field can be recovered by summing over the log-likelihood of each wavelet term. This decomposition allows us to separate the information from different scales and identify distribution shifts in the data such as unknown scale-dependent systematics. The resulting likelihood analysis can not only identify these types of systematics, but can also be made optimal, in the sense that the Multiscale Flow can learn the full likelihood at the field without any dimensionality reduction. We apply Multiscale Flow to weak lensing mock datasets for cosmological inference and show that it significantly outperforms traditional summary statistics such as power spectrum and peak counts, as well as machine learning–based summary statistics such as scattering transform and convolutional neural networks. We further show that Multiscale Flow is able to identify distribution shifts not in the training data such as baryonic effects. Finally, we demonstrate that Multiscale Flow can be used to generate realistic samples of weak lensing data.

79 ASTRONOMY AND ASTROPHYSICS↗

GNET2: an R package for constructing gene regulatory networks from transcriptomic data

Abstract Motivation The Gene Network Estimation Tool (GNET) is designed to build gene regulatory networks (GRNs) from transcriptomic gene expression data with a probabilistic graphical model. The data preprocessing, model construction and visualization modules of the original GNET software were developed on different programming platforms, which were inconvenient for users to deploy and use. Results Here, we present GNET2, an improved implementation of GNET as an integrated R package. GNET2 provides more flexibility for parameter initialization and regulatory module construction based on the core iterative modeling process of the original algorithm. The data exchange interface of GNET2 is handled within an R session automatically. Given the growing demand for regulatory network reconstruction from transcriptomic data, GNET2 offers a convenient option for GRN inference on large datasets. Availability and implementation The source code of GNET2 is available at https://github.com/jianlin-cheng/GNET2. Supplementary information Supplementary data are available at Bioinformatics online.

59 BASIC BIOLOGICAL SCIENCES↗

Loss of expression of miR-335 is implicated in hepatic stellate cell migration and activation

Activation and migration of resident stellate cells (HSCs) within the hepatic space of Disse play an important role in hepatic fibrosis, which accounts for the increased numbers of activated HSCs in areas of inflammation during hepatic fibrosis. Currently, microRNAs have been found to play essential roles in HSC differentiation, proliferation, apoptosis, fat accumulation and collagen production. However, little is known about microRNA mediated HSC activation and migration. In this study, the miRNA expression profiles of quiescent HSCs, partially activated HSCs and fully activated HSCs were compared in pairs. Gene ontology (GO) and GO-Map network analysis indicated that the activation of HSCs was regulated by microRNAs. Among them miR-335 was confirmed to be significantly reduced during HSC activation by qRT-PCR, and restoring expression of miR-335 inhibited HSC migration and reduced {alpha}-SMA and collagen type I. Previous study revealed that tenascin-C (TNC), an extracellular matrix glycoprotein involved in cell migration, might be a target of miR-335. Therefore, we further studied the TNC expression in miR-335 over-expressed HSCs. Our data showed that exogenous TNC could enhance HSC migration in vitro and miR-335 restoration resulted in a significant inhibition of TNC expression. These results demonstrated that miR-335 restoration inhibited HSC migration, at least in part, via downregulating the TNC expression.

60 APPLIED LIFE SCIENCES↗

Nanostructure Determination by Co-Refining Models to Multiple Datasets

The results of the work are contained in the publications resulting from the grant (which are listed below). Here I summarize the main findings from the last period of the award, 2006-2007: • Published a paper in Science with Igor Levin outlining the “Nanostructure Problem”, our inability to solve structure at the nanoscale. • Published a paper in Nature demonstrating the first ever ab-initio structure determination of a nanoparticle from atomic pair distribution function (PDF) data. • Published one book and 3 overview articles on PDF methods and the nanostructure problem. • Completed a project that sought to find a structural response to the presence of the so-called “intermediate phase” in network glasses which appears close to the rigidity percolation threshold in these systems. The main result was that we did not see convincing evidence for this, which drew into doubt the idea that Ge x Se 1-x glasses were a model system exhibiting rigidity percolation.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

PyDDA: A Pythonic Direct Data Assimilation Framework for Wind Retrievals

This software assimilates data from an arbitrary number of weather radars together with other spatial wind fields (eg numerical weather forecasting model data) in order to retrieve high resolution three dimensional wind fields. PyDDA uses NumPy and SciPy’s optimization techniques combined with the Python Atmospheric Radiation Measurement (ARM) Radar Toolkit (Py-ART) in order to create wind fields using the 3D variational technique (3DVAR). PyDDA is hosted and distributed on GitHub at https://github.com/openradar/PyDDA. PyDDA has the potential to be used by the atmospheric science community to develop high resolution wind retrievals from radar networks. These retrievals can be used for the evaluation of numerical weather forecasting models and plume modelling. This paper shows how wind fields from 2 NEXt generation RADar (NEXRAD) WSR-88D radars and the High Resolution Rapid Refresh can be assimilated together using PyDDA to create a high resolution wind field inside Hurricane Florence.

54 ENVIRONMENTAL SCIENCES↗

An algorithmic approach to crustal deformation analysis

In recent years the analysis of crustal deformation measurements has become important as a result of current improvements in geodetic methods and an increasing amount of theoretical and observational data provided by several earth sciences. A first-generation data analysis algorithm which combines a priori information with current geodetic measurements was proposed. Relevant methods which can be used in the algorithm were discussed. Prior information is the unifying feature of this algorithm. Some of the problems which may arise through the use of a priori information in the analysis were indicated and preventive measures were demonstrated. The first step in the algorithm is the optimal design of deformation networks. The second step in the algorithm identifies the descriptive model of the deformation field. The final step in the algorithm is the improved estimation of deformation parameters. Although deformation parameters are estimated in the process of model discrimination, they can further be improved by the use of a priori information about them. According to the proposed algorithm this information must first be tested against the estimates calculated using the sample data only. Null-hypothesis testing procedures were developed for this purpose. Six different estimators which employ a priori information were examined. Emphasis was put on the case when the prior information is wrong and analytical expressions for possible improvements under incompatible prior information were derived.

Iz, Huseyin Baki↗

A Sundial-Atlas Precursor to the TIMED Mission: A Quick-Response Global Investigation into Coupled Lower Thermospheric, Ionospheric, and Mesospheric Physics

The SUNDIAL-ATLAS effort was a global-scale investigation which responded to the science priorities of the ITM Panel, the Integrated SPD Strategy Implementation Plan as a whole, and the need for potential cost-saving design criteria for the TIMED mission. The investigation focused on coupling processes in the ionospheric-thermospheric system, taking advantage of the timelines of the ATLAS-1 mission (March 1992), and the global-scale ground-based measurement and modeling activities of the SUNDIAL program. The collaborative SUNDIAL-ATLAS activity was the first opportunity for global measurements of the chemistry, kinetics, and electrodynamics which couple the E-, Fl-, and F2-regions into a single interactive system. As such, the program represented an important first step in studying global issues; and accordingly, was an important proof of concept experiment relevant to the strategic mission plans for the ITM community and the upcoming intermediate class satellite program called TIMED. To meet its projected goals, TIMED must perform a number of critical measurements and execute a number of correlations that were to be tried and tested for the first time in the SUNDIAL-ATLAS investigation. This was designed to include global correlations of thermospheric and ionospheric composition during quiet and disturbed conditions and the co-registration of global-scale ground-based measurements with along-track satellite diagnostics. The SUNDIAL component of the current investigation addressed this need by acquiring, reducing, and analyzing a multi-sensor database that complemented and extended that which was generated in the ATLAS mission (Atmospheric Laboratory for Applications and Science). The SUNDIAL data defined the state and condition of the global-scale ionosphere in the altitude range from 100 km to the F2-peak. These data specified the peak heights and densities of the E-, Fl-, and F2-regions, along with the global distributions of intermediate, descending, and sequential layers which play a critical role in the dynamo region of the lower ionospheric-thermospheric domain. The data were collected by the SUNDIAL network of more than 50 ground-based stations utilizing ionosondes, radars, photometers, Fabry-Perot interferometers, and total electron content measurements. The data were acquired during a three-week period centered on the eight-day ATLAS-1 mission, which provided image and photometric sensing of the altitude distributions of the major and minor ions and neutrals in the ITM system. This report focuses on the scientific contributions of the SUNDIAL component of the overall investigation. Specific findings are described in seven papers (attached) published in the Journal of Geophysical Research.

Szuszczewicz, E. P.↗

Orbit Determination and Navigation of the Solar Terrestrial Relations Observatory (STEREO)

This paper provides an overview of the required upgrades necessary for navigation of NASA's twin heliocentric science missions, Solar TErestrial RElations Observatory (STEREO) Ahead and Behind. The orbit determination of the STEREO spacecraft was provided by the NASA Goddard Space Flight Center's (GSFC) Flight Dynamics Facility (FDF) in support of the mission operations activities performed by the Johns Hopkins University Applied Physics Laboratory (APL). The changes to FDF s orbit determination software included modeling upgrades as well as modifications required to process the Deep Space Network X-band tracking data used for STEREO. Orbit results as well as comparisons to independently computed solutions are also included. The successful orbit determination support aided in maneuvering the STEREO spacecraft, launched on October 26, 2006 (00:52 Z), to target the lunar gravity assists required to place the spacecraft into their final heliocentric drift-away orbits where they are providing stereo imaging of the Sun.

Mesarch, Michael↗

Orbit Determination and Navigation of the Solar Terrestrial Relations Observatory (STEREO)

This paper provides an overview of the required upgrades necessary for navigation of NASA's twin heliocentric science missions, Solar TErestrial RElations Observatory (STEREO) Ahead and Behind. The orbit determination of the STEREO spacecraft was provided by the NASA Goddard Space Flight Center's (GSFC) Flight Dynamics Facility (FDF) in support of the mission operations activities performed by the Johns Hopkins University Applied Physics Laboratory (APL). The changes to FDF's orbit determination software included modeling upgrades as well as modifications required to process the Deep Space Network X-band tracking data used for STEREO. Orbit results as well as comparisons to independently computed solutions are also included. The successful orbit determination support aided in maneuvering the STEREO spacecraft, launched on October 26, 2006 (00:52 Z), to target the lunar gravity assists required to place the spacecraft into their final heliocentric drift-away orbits where they are providing stereo imaging of the Sun.

Mesarch, Michael A.↗

Using Open and Interoperable Ways to Publish and Access LANCE AIRS Near-Real Time Data

The Atmospheric Infrared Sounder (AIRS) Near-Real Time (NRT) data from the Land Atmosphere Near real-time Capability for EOS (LANCE) element at the Goddard Earth Sciences Data and Information Services Center (GES DISC) provides information on the global and regional atmospheric state, with very low temporal latency, to support climate research and improve weather forecasting. An open and interoperable platform is useful to facilitate access to, and integration of, LANCE AIRS NRT data. As Web services technology has matured in recent years, a new scalable Service-Oriented Architecture (SOA) is emerging as the basic platform for distributed computing and large networks of interoperable applications. Following the provide-register-discover-consume SOA paradigm, this presentation discusses how to use open-source geospatial software components to build Web services for publishing and accessing AIRS NRT data, explore the metadata relevant to registering and discovering data and services in the catalogue systems, and implement a Web portal to facilitate users' consumption of the data and services.

Zhao, Peisheng↗

Data reduction in deterministic neutron transport calculations using machine learning

Neutron cross section matrices for fission and scattering data are required for each material, temperature, and enrichment level to calculate the neutron transport equation accurately. Here, this information can be a limiting factor when using the multigroup discrete ordinates (S N ) method when the number of energy groups is large. Machine Learning (ML) can be used to replace the need for the cross section matrices by reproducing the function that maps the scalar flux to the scattering and fission sources. Through the use of autoencoders and Deep Jointly-Informed Neural Networks (DJINN), the data storage requirements are reduced by 94% of the original data for a 618 group problem. This is accomplished while preserving the scalar flux, maintaining generality, and decreasing wall clock times.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Network of networks: Time series clustering of AmeriFlux sites

Environmental observation networks, such as AmeriFlux, are foundational for monitoring ecosystem response to climate change, management practices, and natural disturbances; however, their effectiveness depends on their representativeness for the regions or continents. We proposed an empirical, time series approach to quantify the similarity of ecosystem fluxes across AmeriFlux sites. We extracted the diel and seasonal characteristics (i.e., amplitudes, phases) from carbon dioxide, water vapor, energy, and momentum fluxes, which reflect the effects of climate, plant phenology, and ecophysiology on the observations, and explored the potential aggregations of AmeriFlux sites through hierarchical clustering. While net radiation and temperature showed latitudinal clustering as expected, flux variables revealed a more uneven clustering with many small (number of sites < 5), unique groups and a few large (> 100) to intermediate (15–70) groups, highlighting the significant ecological regulations of ecosystem fluxes. Many identified unique groups were from under-sampled ecoregions and biome types of the International Geosphere-Biosphere Programme (IGBP), with distinct flux dynamics compared to the rest of the network. At the finer spatial scale, local topography, disturbance, management, edaphic, and hydrological regimes further enlarge the difference in flux dynamics within the groups. Nonetheless, our clustering approach is a data-driven method to interpret the AmeriFlux network, informing future cross-site syntheses, upscaling, and model-data benchmarking research. Finally, we highlighted the unique and underrepresented sites in the AmeriFlux network, which were found mainly in Hawaii and Latin America, mountains, and at under-sampled IGBP types (e.g., urban, open water), motivating the incorporation of new/unregistered sites from these groups.

54 ENVIRONMENTAL SCIENCES↗