Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “level set methods”

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 955 records · Page 53

Comparison of scatter rejection and low-contrast performance of scan equalization digital radiography (SEDR), slot-scan digital radiography, and full-field digital radiography systems for chest phantom imaging

Purpose: To investigate and compare the scatter rejection properties and low-contrast performance of the scan equalization digital radiography (SEDR) technique to the slot-scan and conventional full-field digital radiography techniques for chest imaging. Methods: A prototype SEDR system was designed and constructed with an a-Se flat-panel (FP) detector to improve image quality in heavily attenuating regions of an anthropomorphic chest phantom. Slot-scanning geometry was used to reject scattered radiation without attenuating primary x rays. The readout scheme of the FP was modified to erase accumulated scatter signals prior to image readout. A 24-segment beam width modulator was developed to regulate x-ray exposures regionally and compensate for the low x-ray flux in heavily attenuating regions. To measure the scatter-to-primary ratios (SPRs), a 2 mm thick lead plate with a 2-D array of aperture holes was used to measure the primary signals, which were then subtracted from those obtained without the lead plate to determine scatter components. A 2-D array of aluminum beads (3 mm in diameter) was used as the low-contrast objects to measure the contrast ratios (CRs) and contrast-to-noise ratios (CNRs) for evaluating the low-contrast performance in chest phantom images. A set of two images acquired with the same techniques were subtracted from each other to measure the noise levels. SPRs, CRs, and CNRs of the SEDR images were measured in four anatomical regions of chest phantom images and compared to those of slot-scan images and full-field images acquired with and without antiscatter grid. Results: The percentage reduction of SPR (percentage of SPRs reduced with scatter removal/rejection methods relative to that for nongrid full-field imaging) averaged over four anatomical regions was measured to be 80%, 83%, and 71% for SEDR, slot-scan, and full-field with grid, respectively. The average CR over four regions was found to improve over that for nongrid full-field imaging by 259%, 279%, and 145% for SEDR, slot-scan, and full-field with grid, respectively. The average CNR over four regions was found to improve over that for nongrid full-field imaging by 201% for SEDR as compared to 133% for the slot-scan technique and 14% for the antiscatter grid method. Conclusions: Both SEDR and slot-scan techniques outperformed the antiscatter grid method used in standard full-field radiography. For imaging with the same effective exposure, the SEDR technique offers no advantage over the slot-scan method in terms of SPRs and CRs. However, it improves CNRs significantly, especially in heavily attenuating regions. The improvement of low-contrast performance may help improve the detection of the lung nodules or other abnormalities and may offer SEDR the potential for dose reduction in chest radiography.

60 APPLIED LIFE SCIENCES↗

Method for modifying trigger level for adsorber regeneration

A method for modifying a NO.sub.x adsorber regeneration triggering variable. Engine operating conditions are monitored until the regeneration triggering variable is met. The adsorber is regenerated and the adsorbtion efficiency of the adsorber is subsequently determined. The regeneration triggering variable is modified to correspond with the decline in adsorber efficiency. The adsorber efficiency may be determined using an empirically predetermined set of values or by using a pair of oxygen sensors to determine the oxygen response delay across the sensors.

42 ENGINEERING↗

Improving Peptide-Level Mass Spectrometry Analysis via Double Competition

A critical statistical task in the analysis of shotgun proteomics data involves controlling the false discovery rate (FDR) among the reported set of discoveries. This task is most commonly solved at the peptide-spectrum match (PSM) level by using target-decoy competition (TDC), in which a set of observed spectra are searched against a database containing a mixture of real (target) and decoy peptides. The PSM-level procedure can be adapted to the peptide level by selecting the top-scoring PSM per peptide prior to FDR estimation. Here, we investigate both PSM-level and peptide-level FDR control methods and come to two conclusions. First, although the TDC procedure is provably correct under certain assumptions, we observe that one of these assumptions - that incorrect PSMs are independent of one another - is frequently violated. Hence, we empirically demonstrate that TDC-based PSM-level FDR estimates can be liberally biased. Further, we propose that researchers avoid PSM-level results and instead focus on peptide-level analysis. Second, we investigate three ways to carry out peptide level TDC and show that the most common method ("PSM-only") offers the lowest statistical power in practice. The most powerful method, peptide-level FDR with PSM competition ("PSM-and-peptide"), carries out competition first at the PSM level and then again at the peptide level. In our experiments, this approach yields an average increase of 17% more discovered peptides at a1% FDR threshold relative to the PSM-only method.

59 BASIC BIOLOGICAL SCIENCES↗

A Tribute to Klaus Ruedenberg

“I am indebted to my father for living, but to my teacher for living well" said Alexander the Great, the King of Greek Macedonia, about his teacher, the philosopher Aristotle. This statement echoes the widely held belief of the students, research associates, collaborators and admirers of Klaus Ruedenberg regarding his invaluable contribution towards shaping their scientific lives. This special issue of the Journal of Physical Chemistry presents a tribute to Professor Klaus Ruedenberg, on the occasion of his 90th birthday, for his numerous scientific contributions to the field of quantum chemistry. The many outstanding papers that are part of this issue document his seminal contributions to a broad range of quantum chemistry, including the evaluation of electron repulsion integrals, the free-electron network model for conjugated molecules, the origin of covalent bonding, including the central role of the kinetic energy in the covalent bond, and the intrinsic identification of chemical bonding patterns in molecular systems, orbital localization that has enabled a deep understanding of many chemical phenomena, the multi-configurational self-consistent field method and the concept of the Full Optimized Reaction Space (FORS) to study chemical rearrangements and its application to the study of global potential energy surfaces and conical intersections, the first construction of systematic, eventempered sequences of orbital sets that approach the complete basis set limit, and the novel simultaneous extrapolation of basis set and level of theory to achieve nearly exact molecular energies and vibrational spectra. In addition, as past students and research associates of Klaus Ruedenberg, that is people whose scientific and personal lives have been critically affected by a great teacher, we wish to point out a less well-known aspect of his professional career, that of an educator. Professor Klaus Ruedenberg has a passion for communicating scientific ideas and educating students, whether they were at the undergraduate, graduate or post-graduate levels or whether they were visitors to his laboratory. No matter how much of his time it took, there were never any shortcuts or excuses, but also no pressure on the students. Knowledge was just there waiting to be gained - for free - by just mining a vast repository of mathematical, physical and chemical concepts. It is not hyperbole to stress his contribution in molding the scientific lives and in turn the personal success of so many people who have had the privilege to learn from him.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Construction of linearly independent non-orthogonal AGP states

We show how to construct a linearly independent set of antisymmetrized geminal power (AGP) states, which allows us to rewrite our recently introduced geminal replacement models as linear combinations of non-orthogonal AGPs. Furthermore, this greatly simplifies the evaluation of matrix elements and permits us to introduce an AGP-based selective configuration interaction method, which can reach arbitrary excitation levels relative to a reference AGP, balancing accuracy and cost as we see fit.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Using heterogeneous data to identify signatures of dengue outbreaks at fine spatio-temporal scales across Brazil

Dengue virus remains a significant public health challenge in Brazil, and seasonal preparation efforts are hindered by variable intra- and interseasonal dynamics. Here, we present a framework for characterizing weekly dengue activity at the Brazilian mesoregion level from 2010–2016 as time series properties that are relevant to forecasting efforts, focusing on outbreak shape, seasonal timing, and pairwise correlations in magnitude and onset. In addition, we use a combination of 18 satellite remote sensing imagery, weather, clinical, mobility, and census data streams and regression methods to identify a parsimonious set of covariates that explain each time series property. The models explained 54% of the variation in outbreak shape, 38% of seasonal onset, 34% of pairwise correlation in outbreak timing, and 11% of pairwise correlation in outbreak magnitude. Regions that have experienced longer periods of drought sensitivity, as captured by the “normalized burn ratio,” experienced less intense outbreaks, while regions with regular fluctuations in relative humidity had less regular seasonal outbreaks. Both the pairwise correlations in outbreak timing and outbreak trend between mesoresgions were best predicted by distance. Our analysis also revealed the presence of distinct geographic clusters where dengue properties tend to be spatially correlated. Forecasting models aimed at predicting the dynamics of dengue activity need to identify the most salient variables capable of contributing to accurate predictions. Our findings show that successful models may need to leverage distinct variables in different locations and be catered to a specific task, such as predicting outbreak magnitude or timing characteristics, to be useful. This advocates in favor of “adaptive models” rather than “one-size-fits-all” models. The results of this study can be applied to improving spatial hierarchical or target-focused forecasting models of dengue activity across Brazil.

59 BASIC BIOLOGICAL SCIENCES↗

VERA Enhancements for Cross Section Shielding and Geometry Capabilities

Two tasks were undertaken in FY22 that focus on VERA enhancements. The first task involved improvements to the new cross section shielding capability that was added to VERA in FY21. This cell-based capability solves the slowing down problem for each pin cell using Dancoff factors calculated from the whole-core problem. The Dancoff factors are determined for each subgroup level for important sets of materials such as fuel rods, control rods, fuel rods loaded with gad, etc. The cross-section shielding is then performed for each cell using a 1D cylindrical collision probabilities (CP) calculation to obtain the equivalence cross sections required for the core transport calculations. The advantage of this method over the whole-core subgroup calculations is efficiency, since far fewer sweeps of the entire core are required.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Measuring the Density Fields around Bright Quasars at z ~ 6 with XQR-30 Spectra

Measuring the density of the intergalactic medium using quasar sight lines in the epoch of reionization is challenging due to the saturation of Ly α absorption. Near a luminous quasar, however, the enhanced radiation creates a proximity zone observable in the quasar spectra where the Ly α absorption is not saturated. In this study, we use 10 high-resolution ( R ≳ 10,000) z ~ 6 quasar spectra from the extended XQR-30 sample to measure the density field in the quasar proximity zones. We find a variety of environments within 3 pMpc distance from the quasars. We compare the observed density cumulative distribution function (CDF) with models from the Cosmic Reionization on Computers simulation and find a good agreement between 1.5 and 3 pMpc from the quasar. This region is far away from the quasar hosts and hence approaching the mean density of the universe, which allows us to use the CDF to set constraints on the cosmological parameter σ 8 = 0.6 ± 0.3. The uncertainty is mainly due to the limited number of high-quality quasar sight lines currently available. Utilizing the more than 200 known quasars at z ≳ 6, this method will allow us to tighten the constraint on σ 8 to the percent level in the future. In the region closer to the quasar within 1.5 pMpc, we find that the density is higher than predicted in the simulation by 1.23 ± 0.17, suggesting that the typical host dark matter halo mass of a bright quasar ( M 1450 < -26.5) at z ~ 6 is .

79 ASTRONOMY AND ASTROPHYSICS↗

Rejection-free stochastic simulation of BNGL-encoded models

Formal rules encoded using the BioNetGen language (BNGL) can be used to represent the system-level dynamics of molecular interactions. Rules allow one to compactly and implicitly specify the reaction network implied by a set of molecules and their interactions. Typically, the reaction network implied by a set of rules is large, which makes generation of the underlying rule-defined network expensive. Moreover, the cost of conventional simulation methods typically depends on network size. Together these factors have limited application of the rule-based modeling approach. To overcome this limitation, several methods have recently been developed for determining the reaction dynamics implied by rules while avoiding the expensive step of network generation. The cost of these 'network-free' simulation methods is independent of the number of reactions implied by rules. Software implementing such methods is needed for the analysis of rule-based models of biochemical systems. Here, we present a software tool called RuleMonkey that implements a network-free stochastic simulation method for rule-based models. The method is rejection free, unlike other network-free methods that introduce null events (i.e., steps in the simulation procedure that do not change the state of the reaction system being simulated), and the software is capable of simulating models encoded in BNGL, a general-purpose model-specification language. We verify that RuleMonkey produces correct simulation results, and we compare its performance against DYNSTOC, another BNGL-compliant general-purpose simulator for rule-based models, as well as various problem-specific codes that implement network-free simulation methods. RuleMonkey enables the simulation of models defined by rule sets that imply large-scale reaction networks. It is faster than DYNSTOC for stiff problems, although it requires the use of more computer memory. RuleMonkey is freely available for non-commercial use as a stand-alone application (http://public.tgen.org/rulemonkey).

59 BASIC BIOLOGICAL SCIENCES↗

A Self-learning Neural Network Approach for Radio Frequency Interference Detection and Removal in Radio Astronomy

In this work, we present a novel neural network (NN) method for the detection and removal of radio frequency interference (RFI) from the raw digitized signal in the signal processing chain of a typical radio astronomy experiment. The main advantage of our method is that it does not require a training set. Instead, our method relies on the fact that the true signal of interest coming from astronomical sources is thermal and therefore described as a Gaussian random process, which cannot be compressed. We employ a variational encoder/decoder network to find the compressible information in the data stream that can explain the most variance with the fewest degrees of freedom. We demonstrate it on a set of toy problems and stored ring buffers from the Baryon Mapping eXperiment prototype. We find that the RFI subtraction is effective at cleaning simulated time streams: While we find that the power spectra of the RFI-cleaned time streams output by the NN suffer from extra signal consistent with additive noise, we find that it is generally around percent level across the band and sub 10% in contaminated spectral channels even when RFI power is an order of magnitude larger than the signal. We discuss the advantages and limitations of this method and possible implementation in the front end of future radio experiments.

79 ASTRONOMY AND ASTROPHYSICS↗

SU-C-202-01: Incorporating Time-Dependent Hypoxia in IMRT Planning

Purpose: To incorporate uncertainties in oxygenation of tumor cells into radiation therapy planning via robust optimization. The model is demonstrated using a clinical prostate cancer case. Methods: The tumor oxygenation levels are determined based on pre- and mid-treatment PET scans. To account for oxygenation changes, we use a radio-sensitivity factor for the effective dose necessary to treat hypoxic cells. Due to the unpredictable nature of re-oxygenation of hypoxic cells, the radio-sensitivity factor and its change is modeled to reside in a time-dependent uncertainty set. This uncertainty can adapt to both pre- and mid-treatment scans, resulting in a two-stage robust treatment planning model. We develop a robust counterpart reformulation that transforms the original NP-hard problem into a linear program consisting of finitely many linear constraints, which can be solved efficiently. Results: The robust plan is compared to both the corresponding fractionated plan which ignores hypoxia, and to the commonly practiced dose-escalated plan that uniformly increases dose based on initial oxygen concentrations. For PTV, the robust plan improves D{sub 50}, D{sub 95}, and EUD by 4% on average compared to fractionated, and by 11% compared to the optimal escalated plan. Moreover, the robust plan spared organs at risk similar to the fractionated plan, while the escalated plan led to over 170% increase of Bladder D{sub 30}. Conclusion: The proposed robust adaptive model can improve tumor coverage by increasing dose to hypoxic tumor voxels that is overlooked by common fractionated treatments. In addition, the robust model anticipates the re-oxygenation process, thus reducing the excess dose at later fractions in comparison to uniform dose-escalation methods.

60 APPLIED LIFE SCIENCES↗

Sci-Sat AM: Radiation Dosimetry and Practical Therapy Solutions - 09: Stability of a water calorimetry system as a primary standard for absorbed dose to water

Purpose: To investigate the stability of a water calorimetry system as a primary standard for absorbed dose to water using measurements performed in cobalt-60 and high-energy linac photon beams over a span of more than a decade. Methods: Calorimetry measures adsorbed dose directly by recording the amount of heat created when ionizing radiation passes through matter. The radiation-induced temperature rise was measured using two thermistors calibrated against the NRC temperature primary standard, using an AC bridge with lock-in amplifier for precise measurement. The calorimeter system was operated under thermal equilibrium at 4 °C (to eliminate convection) with drifts in water temperature less than 0.1 mK/min. Seven water vessels of various designs were used to make repeated measurements over the course of 17 years. Results: The standard uncertainty achieved for a set of ten calorimeter measurements (4 Gy delivered) was generally well below 0.15 % while the variation between multiple sets for a given vessel was consistent with this value. The long-term stability of the system combined with inter-vessel variations indicated that there was good control of the radiochemistry (chemical heat defect). Conclusions: The measurements performed over a period of several years showed that the combined water calorimeters showed stability at +/− 0.25 % level. Thus, rather than relying on a particular vessel as an artifact one can realize the Gray through the more generalized method of combining a glass vessel, high-purity water and thermistor probes. This provides increased robustness in the dissemination of absorbed dose to Canadian users.

60 APPLIED LIFE SCIENCES↗

Real-space representation of the quasiparticle self-consistent GW self-energy and its application to defect calculations

The quasiparticle self-consistent (QS) GW (G for Green's function, W for screened Coulomb interaction) approach incorporates the corrections of the quasiparticle energies from their Kohn-Sham density functional theory (DFT) eigenvalues by means of an energy-independent and Hermitian self-energy matrix usually given in the basis set of the DFT eigenstates. By expanding these into an atom-centered basis set (specifically here the linearized muffin-tin orbitals) a real space representation of the self-energy corrections becomes possible. In this work, We show that this representation is relatively short-ranged. This offers opportunities to construct the self-energy of a complex system from parts of the system by a cut-and-paste method. Specifically for a point defect, represented in a large supercell, the self-energy can be constructed from those of the host and a smaller defect-containing cell. The self-energy of the periodic host can be constructed simply from a GW calculation for the primitive cell. We show for the case of the As Ga in GaAs that the defect part can already be well represented by a minimal eight-atom cell and allows us to construct the self-energy for a 64-atom cell in good agreement with direct QSGW calculations for the large cell. Using this approach to an even larger 216-atom cell shows the defect band approaches an isolated defect level. The calculations also allow us to identify a second defect band which appears as a resonance near the conduction band minimum. The results on the extracted defect levels agree well with Green's function calculations for an isolated defect and with experimental data.

36 MATERIALS SCIENCE↗

BAYESIAN ANALYSIS OF WHITE NOISE LEVELS IN THE FIVE-YEAR WMAP DATA

We develop a new Bayesian method for estimating white noise levels in CMB sky maps, and apply this algorithm to the five-year Wilkinson Microwave Anisotropy Probe (WMAP) data. We assume that the amplitude of the noise rms is scaled by a constant value, {alpha}, relative to a pre-specified noise level. We then derive the corresponding conditional density, P({alpha} | s, C {sub l}, d), which is subsequently integrated into a general CMB Gibbs sampler. We first verify our code by analyzing simulated data sets, and then apply the framework to the WMAP data. For the foreground-reduced five-year WMAP sky maps and the nominal noise levels initially provided in the five-year data release, we find that the posterior means typically range between {alpha} = 1.005 {+-} 0.001 and {alpha} = 1.010 {+-} 0.001 depending on differencing assembly, indicating that the noise level of these maps are biased low by 0.5%-1.0%. The same problem is not observed for the uncorrected WMAP sky maps. After the preprint version of this letter appeared on astro-ph., the WMAP team has corrected the values presented on their web page, noting that the initially provided values were in fact estimates from the three-year data release, not from the five-year estimates. However, internally in their five-year analysis the correct noise values were used, and no cosmological results are therefore compromised by this error. Thus, our method has already been demonstrated in practice to be both useful and accurate.

79 ASTRONOMY AND ASTROPHYSICS↗

Methods to Evaluate Subcolumn Profiles Based on Two-Point Diagnostics

In atmospheric models, stochastic generation of subgrid-scale profiles or “subcolumns” has been used for a variety of purposes. Such subcolumns can be generated from subgrid probability density functions (PDFs) at different vertical levels, when such PDFs are available. To do so, the generator needs to decide how strongly points should be correlated in the vertical, that is, how much the values should be overlapped. This is sometimes called “PDF overlap.” To assess vertical correlation in a simplified, observable setting, here the vertical correlation of vertical velocity in subcloud layers is examined. Doppler lidar is used to evaluate the vertical profiles of vertical velocity produced by a large-eddy simulation (LES) model and the Subgrid Importance Latin Hypercube Sampler (SILHS) subcolumn generator. In order to diagnose unrealistic features in subcolumn profiles, various statistical diagnostics are examined here, including the bivariate PDF of vertical velocity at two separated points (i.e., altitudes), the two-point velocity correlation, the integral correlation length, the PDF of two-point velocity differences, and the skewness and kurtosis of two-point velocity differences. The profiles produced by LES match lidar well, except that they are too smooth at small scales. The profiles produced by SILHS exhibit sharp jumps from updraft to downdraft that are not observed in the lidar data. To reduce the generation of these unrealistically sharp jumps, the SILHS sampling method is revised. The diagnostics confirm that the revised sampling method reduces the overprediction of sharp jumps.

54 ENVIRONMENTAL SCIENCES↗

Shape-constrained Input Estimation for Efficient Multi-shaker Vibration Testing

Multi-shaker vibration testing is gaining interest from structural dynamics test engineers as it can provide a much more accurate match to complicated field vibration responses than traditional single-axis shaker tests. However, the force capabilities of the small modal shakers typically used in multi-shaker vibration tests has limited the achievable response levels. To date, most multi-shaker vibration tests have been performed using a variety of standard, commercially-available control systems. While these control systems are adequate for a wide range of multiple-input/multiple-output tests, their control algorithms have not been tailored for the specific problem of multi-shaker vibration tests: efficiently coordinating the various shakers to work together to achieve a desired response. Here, a new input estimation algorithm is developed and demonstrated using simulations and actual test data. This algorithm, dubbed shape-constrained input estimation, is shown to effectively coordinate multiple shakers using a set of constraint vectors based on the deflection shapes of the test structure. This is accomplished by using the singular vector shapes of the system frequency response matrix, which allows the constraint vectors to automatically change as a function of frequency. Simulation and test results indicate a significant reduction in the input forces required to achieve a desired response. Finally, the results indicate that shape-constrained input estimation is an effective method to achieve higher response levels from limited shaker forces which will enable higher level multi-shaker vibration tests to be performed.

42 ENGINEERING↗

Metabolic engineering of Caldicellulosiruptor bescii yields increased hydrogen production from lignocellulosic biomass

Background: Members of the anaerobic thermophilic bacterial genus Caldicellulosiruptor are emerging candidates for consolidated bioprocessing (CBP) because they are capable of efficiently growing on biomass without conventional pretreatment. C. bescii produces primarily lactate, acetate and hydrogen as fermentation products, and while some Caldicellulosiruptor strains produce small amounts of ethanol C. bescii does not, making it an attractive background to examine the effects of metabolic engineering. The recent development of methods for genetic manipulation has set the stage for rational engineering of this genus for improved biofuel production. Here, we report the first targeted gene deletion, the gene encoding lactate dehydrogenase (ldh), for metabolic engineering of a member of this genus. Results: A deletion of the C. bescii L-lactate dehydrogenase gene (ldh) was constructed on a non-replicating plasmid and introduced into the C. bescii chromosome by marker replacement. The resulting strain failed to produce detectable levels of lactate from cellobiose and maltose, instead increasing production of acetate and H2 by 21-34% relative to the wild type and pyrFA parent strains. The same phenotype was observed on a real-world substrate switchgrass (Panicum virgatum). Furthermore, the ldh deletion strain grew to a higher maximum optical density than the wild type on maltose and cellobiose, consistent with the prediction that the mutant would gain additional ATP with increased acetate production. Conclusions: Deletion of ldh in C. bescii is the first use of recently developed genetic methods for metabolic engineering of these bacteria. This deletion resulted in a redirection of electron flow from production of lactate to acetate and hydrogen. New capabilities in metabolic engineering combined with intrinsic utilization of lignocellulosic materials position these organisms to provide a new paradigm for consolidated bioprocessing of fuels and other products from biomass.

Biohydrogen↗

Multitask methods for predicting molecular properties from heterogeneous data

Data generation remains a bottleneck in training surrogate models to predict molecular properties. We demonstrate that multitask Gaussian process regression overcomes this limitation by leveraging both expensive and cheap data sources. In particular, we consider training sets constructed from coupled-cluster (CC) and density functional theory (DFT) data. We report that multitask surrogates can predict at CC-level accuracy with a reduction in data generation cost by over an order of magnitude. Of note, our approach allows the training set to include DFT data generated by a heterogeneous mix of exchange–correlation functionals without imposing any artificial hierarchy on functional accuracy. More generally, the multitask framework can accommodate a wider range of training set structures—including the full disparity between the different levels of fidelity—than existing kernel approaches based on Δ-learning although we show that the accuracy of the two approaches can be similar. Consequently, multitask regression can be a tool for reducing data generation costs even further by opportunistically exploiting existing data sources.

Chemistry↗