Maximizing value of genetic sequence data requires an enabling environment and urgency.

Severe price spikes of the major grain commodities and rapid expansion of cultivated area in the past two decades are symptoms of a severely stressed global food supply. Scientific discovery and improved agricultural productivity are needed and are enabled by unencumbered access to, and use of, genetic sequence data. In the same way the world witnessed rapid development of vaccines for COVID-19, genetic sequence data afford enormous opportunities to improve crop production. In addition to an enabling regulatory environment that allowed for the sharing of genetic sequence data, robust funding fostered the rapid development of coronavirus diagnostics and COVID-19 vaccines. A similar level of commitment, collaboration, and cooperation is needed for agriculture.

The data-driven future of high-energy-density physics.

High-energy-density physics is the field of physics concerned with studying matter at extremely high temperatures and densities. Such conditions produce highly nonlinear plasmas, in which several phenomena that can normally be treated independently of one another become strongly coupled. The study of these plasmas is important for our understanding of astrophysics, nuclear fusion and fundamental physics-however, the nonlinearities and strong couplings present in these extreme physical systems makes them very difficult to understand theoretically or to optimize experimentally. Here we argue that machine learning models and data-driven methods are in the process of reshaping our exploration of these extreme systems that have hitherto proved far too nonlinear for human researchers. From a fundamental perspective, our understanding can be improved by the way in which machine learning models can rapidly discover complex interactions in large datasets. From a practical point of view, the newest generation of extreme physics facilities can perform experiments multiple times a second (as opposed to approximately daily), thus moving away from human-based control towards automatic control based on real-time interpretation of diagnostic data and updates of the physics model. To make the most of these emerging opportunities, we suggest proposals for the community in terms of research design, training, best practice and support for synthetic diagnostics and data analysis.

Designing accurate emulators for scientific processes using calibration-driven deep models.

Predictive models that accurately emulate complex scientific processes can achieve speed-ups over numerical simulators or experiments and at the same time provide surrogates for improving the subsequent analysis. Consequently, there is a recent surge in utilizing modern machine learning methods to build data-driven emulators. In this work, we study an often overlooked, yet important, problem of choosing loss functions while designing such emulators. Popular choices such as the mean squared error or the mean absolute error are based on a symmetric noise assumption and can be unsuitable for heterogeneous data or asymmetric noise distributions. We propose Learn-by-Calibrating, a novel deep learning approach based on interval calibration for designing emulators that can effectively recover the inherent noise structure without any explicit priors. Using a large suite of use-cases, we demonstrate the efficacy of our approach in providing high-quality emulators, when compared to widely-adopted loss function choices, even in small-data regimes.

A measurement of the equation of state of carbon envelopes of white dwarfs.

White dwarfs represent the final state of evolution for most stars1-3. Certain classes of white dwarfs pulsate4,5, leading to observable brightness variations, and analysis of these variations with theoretical stellar models probes their internal structure. Modelling of these pulsating stars provides stringent tests of white dwarf models and a detailed picture of the outcome of the late stages of stellar evolution6. However, the high-energy-density states that exist in white dwarfs are extremely difficult to reach and to measure in the laboratory, so theoretical predictions are largely untested at these conditions. Here we report measurements of the relationship between pressure and density along the principal shock Hugoniot (equations describing the state of the sample material before and after the passage of the shock derived from conservation laws) of hydrocarbon to within five per cent. The observed maximum compressibility is consistent with theoretical models that include detailed electronic structure. This is relevant for the equation of state of matter at pressures ranging from 100 million to 450 million atmospheres, where the understanding of white dwarf physics is sensitive to the equation of state and where models differ considerably. The measurements test these equation-of-state relations that are used in the modelling of white dwarfs and inertial confinement fusion experiments7,8, and we predict an increase in compressibility due to ionization of the inner-core orbitals of carbon. We also find that a detailed treatment of the electronic structure and the electron degeneracy pressure is required to capture the measured shape of the pressure-density evolution for hydrocarbon before peak compression. Our results illuminate the equation of state of the white dwarf envelope (the region surrounding the stellar core that contains partially ionized and partially degenerate non-ideal plasmas), which is a weak link in the constitutive physics informing the structure and evolution of white dwarf stars9.

Scalable Topological Data Analysis and Visualization for Evaluating Data-Driven Models in Scientific Applications.

With the rapid adoption of machine learning techniques for large-scale applications in science and engineering comes the convergence of two grand challenges in visualization. First, the utilization of black box models (e.g., deep neural networks) calls for advanced techniques in exploring and interpreting model behaviors. Second, the rapid growth in computing has produced enormous datasets that require techniques that can handle millions or more samples. Although some solutions to these interpretability challenges have been proposed, they typically do not scale beyond thousands of samples, nor do they provide the high-level intuition scientists are looking for. Here, we present the first scalable solution to explore and analyze high-dimensional functions often encountered in the scientific data analysis pipeline. By combining a new streaming neighborhood graph construction, the corresponding topology computation, and a novel data aggregation scheme, namely topology aware datacubes, we enable interactive exploration of both the topological and the geometric aspect of high-dimensional data. Following two use cases from high-energy-density (HED) physics and computational biology, we demonstrate how these capabilities have led to crucial new insights in both applications.

Theoretical and experimental investigation of the equation of state of boron plasmas.

We report a theoretical equation of state (EOS) table for boron across a wide range of temperatures (5.1×10^{4}-5.2×10^{8} K) and densities (0.25-49 g/cm^{3}) and experimental shock Hugoniot data at unprecedented high pressures (5608±118 GPa). The calculations are performed with first-principles methods combining path-integral Monte Carlo (PIMC) at high temperatures and density-functional-theory molecular-dynamics (DFT-MD) methods at lower temperatures. PIMC and DFT-MD cross-validate each other by providing coherent EOS (difference <1.5 Hartree/boron in energy and <5% in pressure) at 5.1×10^{5} K. The Hugoniot measurement is conducted at the National Ignition Facility using a planar shock platform. The pressure-density relation found in our shock experiment is on top of the shock Hugoniot profile predicted with our first-principles EOS and a semiempirical EOS table (LEOS 50). We investigate the self-diffusivity and the effect of thermal and pressure-driven ionization on the EOS and shock compression behavior in high-pressure and -temperature conditions. We also study the sensitivity of a polar direct-drive exploding pusher platform to pressure variations based on applying pressure multipliers to LEOS 50 and by utilizing a new EOS model based on our ab initio simulations via one-dimensional radiation-hydrodynamic calculations. The results are valuable for future theoretical and experimental studies and engineering design in high-energy density research.

Evaluation of fitness in F2 generations of Africa Biofortified Sorghum event 188 and weedy Sorghum bicolor ssp. drummondii

Background: Introgression of transgenes from crops to their wild species may enhance the adaptive advantage and therefore the invasiveness of and weedy forms. The study evaluated the effect of Africa Biofortified Sorghum (ABS) genes from ABS event 188 on the vegetative and reproductive features of the F2 populations derived from crosses with Sorghum bicolor subsp. drummondii. Results: F1 populations were obtained from reciprocal crosses involving ABS event 188 and its null segregant with inbred weedy parents from S. bicolor subsp. drummondii. Four F2 populations and four parental populations were raised in RCBD with 4 replications in a confined field plot for two seasons. Vegetative and reproductive traits were evaluated. The vigour shown in the F2 populations from the reciprocal crosses involving ABS event 188 and S. bicolor subsp. drummondii was similar to that in the crosses involving the null segregant and S. bicolor subsp. drummondii. Differences in vegetative and reproductive parameters were observed between the parental controls and the F2 populations. Examination of the above and below ground vegetative biomass showed lack of novel weedy related features like rhizomes. Conclusions: Therefore, release of crops with ABS 188 transgenes into cropping systems is not likely to pose a risk of conferring additional adaptive advantage in the introgressing populations. The interaction of ABS genes in weedy backgrounds will also not have an effect towards enhancing the weedy features in these populations.

Is there a place for nutrition-sensitive agriculture?

The focus of the review paper is to discuss how biotechnological innovations are opening new frontiers to mitigate nutrition in key agricultural crops with potential for large-scale health impact to people in Africa. The general objective of the Africa Biofortified Sorghum (ABS) project is to develop and deploy sorghum with enhanced pro-vitamin A to farmers and end-users in Africa to alleviate vitamin A-related micronutrient deficiency diseases. To achieve this objective the project technology development team has developed several promising high pro-vitamin A sorghum events. ABS 203 events are so far the most advanced and well-characterised lead events with about 12 µg ß-carotene/g tissue which would supply about 40-50 % of the daily recommended vitamin A at harvest. Through gene expression optimisation other events with higher amounts of pro-vitamin A, including ABS 214, ABS 235, ABS 239 with 25, 30-40, 40-50 µg ß-carotene/g tissue, respectively, have been developed. ABS 239 would provide twice recommended pro-vitamin A at harvest, 50-90 % after 3 months storage and 13-45 % after 6 months storage for children. Preliminary results of introgression of ABS pro-vitamin A traits into local sorghum varieties in target countries Nigeria and Kenya show stable introgression of ABS vitamin A into local farmer-preferred sorghums varieties. ABS gene Intellectual Property Rights and Freedom to Operate have been donated for use royalty free for Africa. Prior to the focus on the current target countries, the project was implemented by fourteen institutions in Africa and the USA. For the next 5 years, the project will complete ABS product development, complete regulatory science data package and apply for product deregulation in target African countries.

Ultraviolet surprise: Efficient soft x-ray high-harmonic generation in multiply ionized plasmas.

High-harmonic generation is a universal response of matter to strong femtosecond laser fields, coherently upconverting light to much shorter wavelengths. Optimizing the conversion of laser light into soft x-rays typically demands a trade-off between two competing factors. Because of reduced quantum diffusion of the radiating electron wave function, the emission from each species is highest when a short-wavelength ultraviolet driving laser is used. However, phase matching--the constructive addition of x-ray waves from a large number of atoms--favors longer-wavelength mid-infrared lasers. We identified a regime of high-harmonic generation driven by 40-cycle ultraviolet lasers in waveguides that can generate bright beams in the soft x-ray region of the spectrum, up to photon energies of 280 electron volts. Surprisingly, the high ultraviolet refractive indices of both neutral atoms and ions enabled effective phase matching, even in a multiply ionized plasma. We observed harmonics with very narrow linewidths, while calculations show that the x-rays emerge as nearly time-bandwidth-limited pulse trains of ~100 attoseconds.

Mapping nanoscale absorption of femtosecond laser pulses using plasma explosion imaging.

We make direct observations of localized light absorption in a single nanostructure irradiated by a strong femtosecond laser field, by developing and applying a technique that we refer to as plasma explosion imaging. By imaging the photoion momentum distribution resulting from plasma formation in a laser-irradiated nanostructure, we map the spatial location of the highly localized plasma and thereby image the nanoscale light absorption. Our method probes individual, isolated nanoparticles in vacuum, which allows us to observe how small variations in the composition, shape, and orientation of the nanostructures lead to vastly different light absorption. Here, we study four different nanoparticle samples with overall dimensions of ∼100 nm and find that each sample exhibits distinct light absorption mechanisms despite their similar size. Specifically, we observe subwavelength focusing in single NaCl crystals, symmetric absorption in TiO2 aggregates, surface enhancement in dielectric particles containing a single gold nanoparticle, and interparticle hot spots in dielectric particles containing multiple smaller gold nanoparticles. These observations demonstrate how plasma explosion imaging directly reveals the diverse ways in which nanoparticles respond to strong laser fields, a process that is notoriously challenging to model because of the rapid evolution of materials properties that takes place on the femtosecond time scale as a solid nanostructure is transformed into a dense plasma.

Observation and control of shock waves in individual nanoplasmas.

Using an apparatus that images the momentum distribution of individual, isolated 100-nm-scale plasmas, we make the first experimental observation of shock waves in nanoplasmas. We demonstrate that the introduction of a heating pulse prior to the main laser pulse increases the intensity of the shock wave, producing a strong burst of quasimonoenergetic ions with an energy spread of less than 15%. Numerical hydrodynamic calculations confirm the appearance of accelerating shock waves and provide a mechanism for the generation and control of these shock waves. This observation of distinct shock waves in dense plasmas enables the control, study, and exploitation of nanoscale shock phenomena with tabletop-scale lasers.