Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data set”

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 1,189 records · Page 66

A linear programming approach for estimating the structure of a sparse linear genetic network from transcript profiling data

Background: A genetic network can be represented as a directed graph in which a node corresponds to a gene and a directed edge specifies the direction of influence of one gene on another. The reconstruction of such networks from transcript profiling data remains an important yet challenging endeavor. A transcript profile specifies the abundances of many genes in a biological sample of interest. Prevailing strategies for learning the structure of a genetic network from highdimensional transcript profiling data assume sparsity and linearity. Many methods consider relatively small directed graphs, inferring graphs with up to a few hundred nodes. This work examines large undirected graphs representations of genetic networks, graphs with many thousands of nodes where an undirected edge between two nodes does not indicate the direction of influence, and the problem of estimating the structure of such a sparse linear genetic network (SLGN) from transcript profiling data. Results: The structure learning task is cast as a sparse linear regression problem which is then posed as a LASSO (l 1-constrained fitting) problem and solved finally by formulating a Linear Program (LP). A bound on the Generalization Error of this approach is given in terms of the LeaveOne-Out Error. The accuracy and utility of LP-SLGNs is assessed quantitatively and qualitatively using simulated and real data. The Dialogue for Reverse Engineering Assessments and Methods (DREAM) initiative provides gold standard data sets and evaluation metrics that enable and facilitate the comparison of algorithms for deducing the structure of networks. The structures of LP-SLGNs estimated from the INSILICO1, INSILICO2 and INSILICO3 simulated DREAM2 data sets are comparable to those proposed by the first and/or second ranked teams in the DREAM2 competition. The structures of LP-SLGNs estimated from two published Saccharomyces cerevisae cell cycle transcript profiling data sets capture known regulatory associations. In each S. cerevisiae LP-SLGN, the number of nodes with a particular degree follows an approximate power law suggesting that its degree distributions is similar to that observed in real-world networks. Inspection of these LP-SLGNs suggests biological hypotheses amenable to experimental verification.

59 BASIC BIOLOGICAL SCIENCES↗

Prospects for de novo phasing with de novo protein models

In a first systematic exploration of phasing with Rosetta de novo models, it is shown that all-atom refinement of coarse-grained models significantly improves both the model quality and performance in molecular replacement with the Phaser software. The prospect of phasing diffraction data sets ‘de novo’ for proteins with previously unseen folds is appealing but largely untested. In a first systematic exploration of phasing with Rosetta de novo models, it is shown that all-atom refinement of coarse-grained models significantly improves both the model quality and performance in molecular replacement with the Phaser software. 15 new cases of diffraction data sets that are unambiguously phased with de novo models are presented. These diffraction data sets represent nine space groups and span a large range of solvent contents (33–79%) and asymmetric unit copy numbers (1–4). No correlation is observed between the ease of phasing and the solvent content or asymmetric unit copy number. Instead, a weak correlation is found with the length of the modeled protein: larger proteins required somewhat less accurate models to give successful molecular replacement. Overall, the results of this survey suggest that de novo models can phase diffraction data for approximately one sixth of proteins with sizes of 100 residues or less. However, for many of these cases, ‘de novo phasing with de novo models’ requires significant investment of computational power, much greater than 10{sup 3} CPU days per target. Improvements in conformational search methods will be necessary if molecular replacement with de novo models is to become a practical tool for targets without homology to previously solved protein structures.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

A toolkit for detecting technical surprise.

The detection of a scientific or technological surprise within a secretive country or institute is very difficult. The ability to detect such surprises would allow analysts to identify the capabilities that could be a military or economic threat to national security. Sandia's current approach utilizing ThreatView has been successful in revealing potential technological surprises. However, as data sets become larger, it becomes critical to use algorithms as filters along with the visualization environments. Our two-year LDRD had two primary goals. First, we developed a tool, a Self-Organizing Map (SOM), to extend ThreatView and improve our understanding of the issues involved in working with textual data sets. Second, we developed a toolkit for detecting indicators of technical surprise in textual data sets. Our toolkit has been successfully used to perform technology assessments for the Science & Technology Intelligence (S&TI) program.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Multifrequency radar observations of marine clouds during the EPCAPE campaign

The Eastern Pacific Cloud Aerosol Precipitation Experiment (EPCAPE) was a year-round campaign conducted by the US Department of Energy at the Scripps Institution of Oceanography in La Jolla, CA, USA, with a focus on characterizing atmospheric processes at a coastal location. The ground-based prototype of a new Ka-, W-, and G-band (35.75, 94.88, and 238.8 GHz) profiling atmospheric radar, named CloudCube, which was developed at the Jet Propulsion Laboratory, took part in the experiment during 6 weeks in March and April 2023. This article describes the unique data sets that were obtained during the field campaign from a variety of marine clouds and light precipitation. These are, to the best of the authors' knowledge, the first observations of atmospheric clouds using simultaneous multifrequency measurements including 238.8 GHz. These data sets therefore provide an exceptional opportunity to study and analyze hydrometeors with diameters in the millimeter- and submillimeter size range that can be used to better understand cloud and precipitation structure, formation, and evolution. The data sets referenced in this article are intended to provide a complete, extensive, and high-quality collection of G-band data in the form of Doppler spectra and Doppler moments. In addition, Ka-band and W-band reflectivity and Ka-, W-, and G-band reflectivity ratio profiles are included for several cases of interest on 6 different days.

54 ENVIRONMENTAL SCIENCES↗

BatAnalysis - A Comprehensive Python Pipeline for Swift BAT Survey Analysis

The Swift Burst Alert Telescope (BAT) is a coded-aperture gamma-ray instrument with a large field of view that primarily operates in survey mode when it is not triggering on transient events. The survey data consist of 80- channel detector plane histograms that accumulate photon counts over periods of at least 5 minutes. These histograms are processed on the ground and are used to produce the survey data set between 14 and 195 keV. Survey data comprise >90% of all BAT data by volume and allow for the tracking of long-term light curves and spectral properties of cataloged and uncataloged hard X-ray sources. Until now, the survey data set has not been used to its full potential due to the complexity associated with its analysis and the lack of easily usable pipelines. Here, we introduce the BatAnalysis Python package, a wrapper for HEASoftpy, which provides a modern, opensource pipeline to process and analyze BAT survey data. BatAnalysis allows members of the community to use BAT survey data in more advanced analyses of astrophysical sources, including pulsars, pulsar wind nebula, active galactic nuclei, and other known/unknown transient events that may be detected in the hard X-ray band. We outline the steps taken by the Python code and exemplify its usefulness and accuracy by analyzing survey data of the Crab Nebula, NGC 2992, and a previously uncataloged MAXI transient. The BatAnalysis package allows for ~18 yr of BAT survey data to be used in a systematic way to study a large variety of astrophysical sources.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

BatAnalysis - A Comprehensive Python Pipeline for Swift BAT Survey Analysis

The Swift Burst Alert Telescope (BAT) is a coded-aperture gamma-ray instrument with a large field of view that primarily operates in survey mode when it is not triggering on transient events. The survey data consist of 80-channel detector plane histograms that accumulate photon counts over periods of at least 5 minutes. These histograms are processed on the ground and are used to produce the survey data set between 14 and 195 keV. Survey data comprise >90% of all BAT data by volume and allow for the tracking of long-term light curves and spectral properties of cataloged and uncataloged hard X-ray sources. Until now, the survey data set has not been used to its full potential due to the complexity associated with its analysis and the lack of easily usable pipelines. Here, we introduce the BatAnalysis Python package, a wrapper for HEASoftpy, which provides a modern, open-source pipeline to process and analyze BAT survey data. BatAnalysis allows members of the community to use BAT survey data in more advanced analyses of astrophysical sources, including pulsars, pulsar wind nebula, active galactic nuclei, and other known/unknown transient events that may be detected in the hard X-ray band. We outline the steps taken by the Python code and exemplify its usefulness and accuracy by analyzing survey data of the Crab Nebula, NGC 2992, and a previously uncataloged MAXI transient. The BatAnalysis package allows for ~18 yr of BAT survey data to be used in a systematic way to study a large variety of astrophysical sources.

79 ASTRONOMY AND ASTROPHYSICS↗

A Validated Prediction Model for Overall Survival From Stage III Non-Small Cell Lung Cancer: Toward Survival Prediction for Individual Patients

Purpose: Although patients with stage III non-small cell lung cancer (NSCLC) are homogeneous according to the TNM staging system, they form a heterogeneous group, which is reflected in the survival outcome. The increasing amount of information for an individual patient and the growing number of treatment options facilitate personalized treatment, but they also complicate treatment decision making. Decision support systems (DSS), which provide individualized prognostic information, can overcome this but are currently lacking. A DSS for stage III NSCLC requires the development and integration of multiple models. The current study takes the first step in this process by developing and validating a model that can provide physicians with a survival probability for an individual NSCLC patient. Methods and Materials: Data from 548 patients with stage III NSCLC were available to enable the development of a prediction model, using stratified Cox regression. Variables were selected by using a bootstrap procedure. Performance of the model was expressed as the c statistic, assessed internally and on 2 external data sets (n=174 and n=130). Results: The final multivariate model, stratified for treatment, consisted of age, gender, World Health Organization performance status, overall treatment time, equivalent radiation dose, number of positive lymph node stations, and gross tumor volume. The bootstrapped c statistic was 0.62. The model could identify risk groups in external data sets. Nomograms were constructed to predict an individual patient's survival probability ( (www.predictcancer.org)). The data set can be downloaded at (https://www.cancerdata.org/10.1016/j.ijrobp.2015.02.048). Conclusions: The prediction model for overall survival of patients with stage III NSCLC highlights the importance of combining patient, clinical, and treatment variables. Nomograms were developed and validated. This tool could be used as a first building block for a decision support system.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

CHARACTERIZING THE FORMATION HISTORY OF MILKY WAY LIKE STELLAR HALOS WITH MODEL EMULATORS

We use the semi-analytic model ChemTreeN, coupled to cosmological N-body simulations, to explore how different galaxy formation histories can affect observational properties of Milky Way like galaxies' stellar halos and their satellite populations. Gaussian processes are used to generate model emulators that allow one to statistically estimate a desired set of model outputs at any location of a p-dimensional input parameter space. This enables one to explore the full input parameter space orders of magnitude faster than could be done otherwise. Using mock observational data sets generated by ChemTreeN itself, we show that it is possible to successfully recover the input parameter vectors used to generate the mock observables if the merger history of the host halo is known. However, our results indicate that for a given observational data set, the determination of 'best-fit' parameters is highly susceptible to the particular merger history of the host. Very different halo merger histories can reproduce the same observational data set, if the 'best-fit' parameters are allowed to vary from history to history. Thus, attempts to characterize the formation history of the Milky Way using these kind of techniques must be performed statistically, analyzing large samples of high-resolution N-body simulations.

79 ASTRONOMY AND ASTROPHYSICS↗

The Chemical and Products Database, a resource for exposure-relevant data on chemicals in consumer products

Quantitative data on product chemical composition is a necessary parameter for characterizing near-field exposure. This data set comprises reported and predicted information on more than 75,000 chemicals and more than 15,000 consumer products. The data’s primary intended use is for exposure, risk, and safety assessments. The data set includes specific products with quantitative or qualitative ingredient information, which has been publicly disclosed through material safety data sheets (MSDS) and ingredient lists. A single product category from a refined and harmonized set of categories has been assigned to each product. The data set also contains information on the functional role of chemicals in products, which can inform predictions of the concentrations in which they occur. These data will be useful to exposure and risk assessors evaluating chemical and product safety.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Fast Response Temperature by airborne measurements over BNF

The original data were collected during the AAF Engineering Flights (AEF2025) in the vicinity of the ARM Bankhead National Forest (BNF) Atmospheric Observatory (https://www.arm.gov/capabilities/observatories/bnf ) in northwestern Alabama in March 2025. The ARM Aerial Facility ArcticShark uncrewed aerial system (UAS, https://www.arm.gov/capabilities/observatories/aaf/uas) was based at the public-use airport of Posey Field, Alabama (FAA LID: 1M4, 34.28027778° N, 87.60055556° W, 283m MSL) from March 9 through March 24, 2025. The ArcticShark UAS performed nine flights, including eight research flights over the AMF3 (BNF Main Site) and Supplemental Facilities to measure atmospheric state, turbulence, surface IR temperature and imagery, and aerosol number concentration and size distribution. The current data set presents fast response temperature in the atmospheric boundary layer and lower free troposphere measured on the airborne platform throughout the field campaign. The primary instruments used to create the current data set were the fine wire thermocouple probe, the Aircraft Integrated Meteorological Measurement System (AIMMS-30), the Pitot-static system (part of UAS flight control), and the infrared gas analyzer sensor for H2O and CO2 (LI-840). All parameters used in temperature calculations (static pressure, True Air Speed, and absolute humidity in form of dew point temperature) were included in the data set. For user convenience, one additional parameter was also included: the type of flight flag (level, up, down, turn, and combination of thereof).

Air temperature, fast response↗

CHELAX-BNF: Fast Response Temperature by airborne measurements

The original data were collected on board the ARM Aerial Facility ArcticShark uncrewed aerial system (UAS; https://www.arm.gov/capabilities/observatories/aaf/uas ) during the “Characterizing HEterogeneous Land-Atmosphere eXchanges at BNF” field campaign (CHEAX-BNF; https://arm.gov/research/campaigns/aaf2025CHELAX-BNF ). The ARM Aerial Facility ArcticShark UAS was based at the public-use airport of Posey Field, AL (FAA LID: 1M4, 34.28027778° N, 87.60055556° W, 283m MSL) from May 28 to June 23, 2025. The ArcticShark UAS performed 5 flights, including 4 research flights over the BNF Main Site (ARM Mobile Facility 3, https://arm.gov/capabilities/observatories/amf ) and Supplemental Facilities to measure atmospheric state, turbulence, surface IR temperature and imagery, aerosol number concentration, and aerosol size distribution. The current data set presents fast response temperature in the atmospheric boundary layer and lower free troposphere measured on the airborne platform throughout the field campaign. The primary instruments used to create the current data set were the fine wire thermocouple probe, the Aircraft Integrated Meteorological Measurement System (AIMMS-30), the Pitot-static system (part of UAS flight control), and the infrared gas analyzer sensor for H2O and CO2 (LI-840). All parameters used in temperature calculations (static pressure, True Air Speed, and absolute humidity in the form of dew point temperature) were included in the data set. For user convenience, one additional parameter was also included: the type of flight flag (level, up, down, turn, and combination of thereof).

Air temperature, fast response↗

Cosmic constraint to DGP brane model: Geometrical and dynamical perspectives

In this paper, the Dvali-Gabadadze-Porrati (DGP) brane model is confronted by current cosmic observational data sets from geometrical and dynamical perspectives. On the geometrical side, the recently released Union2 557 of type Ia supernovae (SN Ia), the baryon acoustic oscillation from the Sloan Digital Sky Survey and the Two Degree Galaxy Redshift Survey (transverse and radial to line-of-sight data points), the cosmic microwave background measurement given by the seven-year Wilkinson Microwave Anisotropy Probe observations [shift parameters R, l{sub a}(z{sub *}) and redshift at the last scatter surface z{sub *}], ages of high redshifts galaxies, i.e., the lookback time and the high redshift gamma ray bursts are used. On the dynamical side, data points about the growth function of matter linear perturbations are used. Using the same data set combination, we also constrain the flat {Lambda}CDM model as a comparison. The results show that current geometrical and dynamical observational data sets much favor the flat {Lambda}CDM model and the departure from it is above 4{sigma}(6{sigma}) for the spatially flat DGP model with (without) SN systematic errors. The consistence of growth function data points is checked in terms of a relative departure of redshift-distance relation.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Evaluating Variable‐Resolution CESM Over China and Western United States for Use in Water‐Energy Nexus and Impacts Modeling

Abstract Climate models are critical tools to study earth system processes and are often further downscaled to provide refined‐resolution data sets for regional water‐energy impact analyses. In this study, the Variable‐Resolution Community Earth System Model (VR‐CESM), a new dynamical downscaling method, is employed to generate 1/8° simulations for China and the western United States over 1970–2006. Precipitation, temperature, snowpack, radiation, and wind speed are compared with two 1° global climate models, regional climate model ensemble, and reference data sets to evaluate their fidelity. Simulated precipitation skill is dependent upon the reference data sets used, a slight (<+10%) and spatially consistent wet bias in China during summer and the western United States during winter. Precipitation bias in VR‐CESM is the smallest among models and demonstrates the added value provided by resolution refinement. However, temperature biases are generally greater than precipitation due to more complicated interactions among processes. VR‐CESM simulates +2°C warm bias in the low‐elevation regions in China but −2°C cold bias in mountains. However, DJF cold bias (−2.23°C to−3.37°C) and JJA warm bias (+1.13°C to +1.87°C) exist in the western United States. The modeling skills of precipitation and temperature at various topographies are similar in the two evaluated regions. The cold bias was likely affected by downward shortwave radiation deficit (−15%) and slow bias of surface wind speeds (−16% to −29%). Despite the wet and cold bias, snow water equivalent and snow cover are underestimated in mountains. We conclude that VR‐CESM outperforms other models evaluated and provides the best estimation of downscaled climate model data for water‐energy studies.

54 ENVIRONMENTAL SCIENCES↗

SU-F-J-161: Prostate Contouring in Patients with Bilateral Hip Prostheses: Impact of Using Artifact-Reduced CT Images and MRI

Purpose: To investigate the usefulness of metal artifact reduction in CT images of patients with bilateral hip prostheses (BHP) for contouring the prostate and determine if the inclusion of MR images provides additional benefits. Methods: Five patients with BHP were CT scanned using our clinical protocol (140kV, 300mAs, 3mm slices, 1.5mm increment, Philips Medical Systems, OH). Images were reconstructed with the orthopaedic metal artifact reduction (O-MAR) algorithm. MRI scanning was then performed (1.5T, GE Healthcare, WI) with a flat table-top (T{sub 2}-weighted, inherent body coil, FRFSE, 3mm slices with 0mm gap). All images were transferred to Pinnacle (Version 9.2, Philips Medical Systems). For each patient, two data sets were produced: one containing the O-MAR-corrected CT images and another containing fused MRI and O-MAR-corrected CT images. Four genito-urinary radiation oncologists contoured the prostate of each patient on the O-MAR-corrected CT data. Two weeks later, they contoured the prostate on the fused data set, blinded to all other contours. During each contouring session, the oncologists reported their confidence in the contours (1=very confident, 3=not confident) and the contouring difficulty that they experienced (1=really easy, 4=very challenging). Prostate volumes were computed from the contours and the conformity index was used to evaluate inter-observer variability. Results: Larger prostate volumes were found on the O-MAR-corrected CT set than on the fused set (p< 0.05, median=36.9cm{sup 3} vs. 26.63 cm{sup 3}). No significant differences were noted in the inter-observer variability between the two data sets (p=0.3). Contouring difficulty decreased with the addition of MRI (p<0.05) while the radiation oncologists reported more confidence in their contours when MRI was fused with the O-MAR-corrected CT data (p<0.05). Conclusion: This preliminary work demonstrated that, while O-MAR correction to CT images improves visualization of anatomy, the addition of MRI enhanced the oncologists’ confidence in contouring the prostate in patients with BHP.

60 APPLIED LIFE SCIENCES↗

Dis3- and exosome subunit-responsive 3 Prime mRNA instability elements

Highlights: Black-Right-Pointing-Pointer Successful use of a novel RNA-specific bioinformatic tool, RNA SCOPE. Black-Right-Pointing-Pointer Identified novel 3 Prime UTR cis-acting element that destabilizes a reporter mRNA. Black-Right-Pointing-Pointer Show exosome subunits are required for cis-acting element-mediated mRNA instability. Black-Right-Pointing-Pointer Define precise sequence requirements of novel cis-acting element. Black-Right-Pointing-Pointer Show that microarray-defined exosome subunit-regulated mRNAs have novel element. -- Abstract: Eukaryotic RNA turnover is regulated in part by the exosome, a nuclear and cytoplasmic complex of ribonucleases (RNases) and RNA-binding proteins. The major RNase of the complex is thought to be Dis3, a multi-functional 3 Prime -5 Prime exoribonuclease and endoribonuclease. Although it is known that Dis3 and core exosome subunits are recruited to transcriptionally active genes and to messenger RNA (mRNA) substrates, this recruitment is thought to occur indirectly. We sought to discover cis-acting elements that recruit Dis3 or other exosome subunits. Using a bioinformatic tool called RNA SCOPE to screen the 3 Prime untranslated regions of up-regulated transcripts from our published Dis3 depletion-derived transcriptomic data set, we identified several motifs as candidate instability elements. Secondary screening using a luciferase reporter system revealed that one cassette-harboring four elements-destabilized the reporter transcript. RNAi-based depletion of Dis3, Rrp6, Rrp4, Rrp40, or Rrp46 diminished the efficacy of cassette-mediated destabilization. Truncation analysis of the cassette showed that two exosome subunit-sensitive elements (ESSEs) destabilized the reporter. Point-directed mutagenesis of ESSE abrogated the destabilization effect. An examination of the transcriptomic data from exosome subunit depletion-based microarrays revealed that mRNAs with ESSEs are found in every up-regulated mRNA data set but are underrepresented or missing from the down-regulated data sets. Taken together, our findings imply a potentially novel mechanism of mRNA turnover that involves direct Dis3 and other exosome subunit recruitment to and/or regulation on mRNA substrates.

60 APPLIED LIFE SCIENCES↗

Linear systematics mitigation in galaxy clustering in the Dark Energy Survey Year 1 Data

ABSTRACT We implement a linear model for mitigating the effect of observing conditions and other sources of contamination in galaxy clustering analyses. Our treatment improves upon the fiducial systematics treatment of the Dark Energy Survey (DES) Year 1 (Y1) cosmology analysis in four crucial ways. Specifically, our treatment (1) does not require decisions as to which observable systematics are significant and which are not, allowing for the possibility of multiple maps adding coherently to give rise to significant bias even if no single map leads to a significant bias by itself, (2) characterizes both the statistical and systematic uncertainty in our mitigation procedure, allowing us to propagate said uncertainties into the reported cosmological constraints, (3) explicitly exploits the full spatial structure of the galaxy density field to differentiate between cosmology-sourced and systematics-sourced fluctuations within the galaxy density field, and (4) is fully automated, and can therefore be trivially applied to any data set. The updated correlation function for the DES Y1 redMaGiC catalogue minimally impacts the cosmological posteriors from that analysis. Encouragingly, our analysis does improve the goodness-of-fit statistic of the DES Y1 3 × 2pt data set (Δχ2 = −6.5 with no additional parameters). This improvement is due in nearly equal parts to both the change in the correlation function and the added statistical and systematic uncertainties associated with our method. We expect the difference in mitigation techniques to become more important in future work as the size of cosmological data sets grows.

79 ASTRONOMY AND ASTROPHYSICS↗

A Method to Automate the Segmentation of the GTV and ITV for Lung Tumors

Four-dimensional computed tomography (4D-CT) is a useful tool in the treatment of tumors that undergo significant motion. To fully utilize 4D-CT motion information in the treatment of mobile tumors such as lung cancer, autosegmentation methods will need to be developed. Using autosegmentation tools in the Pinnacle{sup 3} v8.1t treatment planning system, 6 anonymized 4D-CT data sets were contoured. Two test indices were developed that can be used to evaluate which autosegmentation tools to apply to a given gross tumor volume (GTV) region of interest (ROI). The 4D-CT data sets had various phase binning error levels ranging from 3% to 29%. The appropriate autosegmentation method (rigid translational image registration and deformable surface mesh) was determined to properly delineate the GTV in all of the 4D-CT phases for the 4D-CT data sets with binning errors of up to 15%. The ITV was defined by 2 methods: a mask of the GTV in all 4D-CT phases and the maximum intensity projection. The differences in centroid position and volume were compared with manual segmentation studies in literature. The indices developed in this study, along with the autosegmentation tools in the treatment planning system, were able to automatically segment the GTV in the four 4D-CTs with phase binning errors of up to 15%.

61 RADIATION PROTECTION AND DOSIMETRY↗

BraggNet: integrating Bragg peaks using neural networks

Neutron crystallography offers enormous potential to complement structures from X-ray crystallography by clarifying the positions of low-Z elements, namely hydrogen. Macromolecular neutron crystallography, however, remains limited, in part owing to the challenge of integrating peak shapes from pulsed-source experiments. To advance existing software, this article demonstrates the use of machine learning to refine peak locations, predict peak shapes and yield more accurate integrated intensities when applied to whole data sets from a protein crystal. The artificial neural network, based on the U-Net architecture commonly used for image segmentation, is trained using about 100 000 simulated training peaks derived from strong peaks. After 100 training epochs (a round of training over the whole data set broken into smaller batches), training converges and achieves a Dice coefficient of around 65%, in contrast to just 15% for negative control data sets. Integrating whole peak sets using the neural network yields improved intensity statistics compared with other integration methods, including k-nearest neighbours. These results demonstrate, for the first time, that neural networks can learn peak shapes and be used to integrate Bragg peaks. In conclusion, it is expected that integration using neural networks can be further developed to increase the quality of neutron, electron and X-ray crystallography data.

97 MATHEMATICS AND COMPUTING↗