The global food system puts enormous pressure on the environment. Managing these pressures requires understanding not only where they occur (i.e., where food is produced), but also who drives them (i.e., where food is consumed). However, the size and complexity of global supply chains make it difficult to trace food production to consumption. Here, we provide the most comprehensive dataset of bilateral trade flows of environmental pressures stemming from food production from producing to consuming nations. The dataset provides environmental pressures for greenhouse gas emissions, water use, nitrogen and phosphorus pollution, and the area of land/water occupancy of food production for crops and animals from land, freshwater, and ocean systems. To produce these data, we improved upon reported food trade and production data to identify producing and consuming nations for each food item, allowing us to match food flows with appropriate environmental pressure data. These data provide a resource for research on sustainable global food consumption and the drivers of environmental impact.

Multicenter clinical trials are essential for evaluating interventions but often face significant challenges in study design, site coordination, participant recruitment, and regulatory compliance. To address these issues, the National Institutes of Health’s National Center for Advancing Translational Sciences established the Trial Innovation Network (TIN). The TIN offers a scientific consultation process, providing access to clinical trial and disease experts who provide input and recommendations throughout the trial’s duration, at no cost to investigators. This approach aims to improve trial design, accelerate implementation, foster interdisciplinary teamwork, and spur innovations that enhance multicenter trial quality and efficiency. The TIN leverages resources of the Clinical and Translational Science Awards (CTSA) program, complementing local capabilities at the investigator’s institution. The Initial Consultation process focuses on the study’s scientific premise, design, site development, recruitment and retention strategies, funding feasibility, and other support areas. As of 6/1/2024, the TIN has provided 431 Initial Consultations to increase efficiency and accelerate trial implementation by delivering customized support and tailored recommendations. Across a range of clinical trials, the TIN has developed standardized, streamlined, and adaptable processes. We describe these processes, provide operational metrics, and include a set of lessons learned for consideration by other trial support and innovation networks.

Vitamin A deficiency (VAD) poses significant health risks and is prevalent in children and adolescents in India. This study aimed to determine the effect of seasonal variation and availability of vitamin A-rich (VA-rich) foods on serum retinol in adolescents. Data on serum retinol levels from adolescents (n 2297, mean age 14 years) from the Comprehensive National Nutrition Survey (2016–2018) in India were analysed, with VAD defined as serum retinol < 0·7 µmol/L. Five states were selected based on a comparable under-five mortality rate and the seasonal spread of the data collection period. Dietary data from adolescents and children ≤ 4 years old were used to assess VA-rich food consumption. A linear mixed model framework was employed to analyse the relationship between serum retinol, month of the year and VA-rich food consumption, with a priori ranking to control for multiple hypothesis testing. Consumption of VA-rich foods, particularly fruits and vegetables/roots and tubers, showed seasonal patterns, with higher consumption during summer and monsoon months. Significant associations were found between serum retinol concentrations and age, month of sampling, consumption of VA-rich foods and fish. VAD prevalence was lowest in August, coinciding with higher consumption of VA-rich fruits and foods. Findings highlight the importance of considering seasonality in assessing VAD prevalence and careful interpretation of survey findings. Intentional design, analysis and reporting of surveys to capture seasonal variation is crucial for accurate assessment and interpretation of VAD prevalence, including during monitoring and evaluation of programmes, and to ensure that public health strategies are appropriately informed.

The First Large Absorption Survey in H i (FLASH) is a large-area radio survey for neutral hydrogen in and around galaxies in the intermediate redshift range $0.4\lt z\lt1.0$, using the 21-cm H i absorption line as a probe of cold neutral gas. The survey uses the ASKAP radio telescope and will cover 24,000 deg$^2$ of sky over the next five years. FLASH breaks new ground in two ways – it is the first large H i absorption survey to be carried out without any optical preselection of targets, and we use an automated Bayesian line-finding tool to search through large datasets and assign a statistical significance to potential line detections. Two Pilot Surveys, covering around 3000 deg$^2$ of sky, were carried out in 2019-22 to test and verify the strategy for the full FLASH survey. The processed data products from these Pilot Surveys (spectral-line cubes, continuum images, and catalogues) are public and available online. In this paper, we describe the FLASH spectral-line and continuum data products and discuss the quality of the H i spectra and the completeness of our automated line search. Finally, we present a set of 30 new H i absorption lines that were robustly detected in the Pilot Surveys, almost doubling the number of known H i absorption systems at $0.4\lt z\lt1$. The detected lines span a wide range in H i optical depth, including three lines with a peak optical depth $\tau\gt1$, and appear to be a mixture of intervening and associated systems. Interestingly, around two-thirds of the lines found in this untargeted sample are detected against sources with a peaked-spectrum radio continuum, which are only a minor (5–20%) fraction of the overall radio-source population. The detection rate for H i absorption lines in the Pilot Surveys (0.3 to 0.5 lines per 40 deg$^2$ ASKAP field) is a factor of two below the expected value. One possible reason for this is the presence of a range of spectral-line artefacts in the Pilot Survey data that have now been mitigated and are not expected to recur in the full FLASH survey. A future paper in this series will discuss the host galaxies of the H i absorption systems identified here.

Aerosol-cloud interactions contribute significant uncertainty to modern climate model predictions. Analysis of complex observed aerosol-cloud parameter relationships is a crucial piece of reducing this uncertainty. Here, we apply two machine learning methods to explore variability in in-situ observations from the NASA ACTIVATE mission. These observations consist of flights over the Western North Atlantic Ocean, providing a large repository of data including aerosol, meteorological, and microphysical conditions in and out of clouds. We investigate this dataset using principal component analysis (PCA), a linear dimensionality reduction technique, and an autoencoder, a deep learning non-linear dimensionality reduction technique. We find that we can reduce the dimensionality of the parameter space by more than a factor of 2 and verify that the deep learning method outperforms a PCA baseline by two orders of magnitude. Analysis in the low dimensional space of both these techniques reveals two consistent physically interpretable regimes—a low pollution regime and an in-cloud regime. Through this work, we show that unsupervised machine learning techniques can learn useful information from in-situ atmospheric observations and provide interpretable results of low-dimensional variability.

Gravity currents are a ubiquitous density-driven flow occurring in both the natural environment and in industry. They include: seafloor turbidity currents, primary vectors of sediment, nutrient and pollutant transport; cold fronts; and hazardous gas spills. However, while the energetics are critical for their evolution and particle suspension, they are included in system-scale models only crudely, so we cannot yet predict and explain the dynamics and run-out of such real-world flows. Herein, a novel depth-averaged framework is developed to capture the evolution of volume, concentration, momentum and turbulent kinetic energy from direct integrals of the full governing equations. For the first time, we show the connection between the vertical profiles, the evolution of the depth-averaged flow and the energetics. The viscous dissipation of mean-flow energy near the bed makes a leading-order contribution, and an energetic approach to entrainment captures detrainment of fluid through particle settling. These observations allow a reconsideration of particle suspension, advancing over 50 years of research. We find that the new formulation can describe the full evolution of a shallow dilute current, with the accuracy depending primarily on closures for the profiles and source terms. Critically, this enables accurate and computationally efficient hazard risk analysis and earth surface modelling.

Surfactant transport is central to a diverse range of natural phenomena with numerous practical applications in physics and engineering. Surprisingly, this process remains relatively poorly understood at the molecular scale. Here, we use non-equilibrium molecular dynamics (NEMD) simulations to study the spreading of sodium dodecyl sulphate on a thin film of liquid water. The molecular form of the control volume is extended to a coordinate system moving with the liquid–vapour interface to track surfactant spreading. We use this to compare the NEMD results to the continuum description of surfactant transport on an interface. By including the molecular details in the continuum model, we establish that the transport equation preserves substantial accuracy in capturing the underlying physics. Moreover, the relative importance of the different mechanisms involved in the transport process is identified. Consequently, we derive a novel exact molecular equation for surfactant transport along a deforming surface. Close agreement between the two conceptually different approaches, i.e. NEMD simulations and the numerical solution of the continuum equation, is found as measured by the surfactant concentration profiles, and the time dependence of the so-called spreading length. The current study focuses on a relatively simple specific solvent–surfactant system, and the observed agreement with the continuum model may not arise for more complicated industrially relevant surfactants and anti-foaming agents. In such cases, the continuum approach may fail to predict accompanying phase transitions, which can still be captured through the NEMD framework.

Inflammation and infections such as malaria affect concentrations of many micronutrient biomarkers and hence estimates of nutritional status. We aimed to assess the relationship between malaria infection and micronutrient biomarker concentrations in pre-school children (PSC), school-age children (SAC) and women of reproductive age (WRA) in Malawi and examine the potential role of malarial immunity on the relationship between malaria and micronutrient biomarkers. Data from the 2015/2016 Malawi micronutrient survey were used. The associations between current or recent malaria infection, detected by rapid diagnostic test and concentration of serum ferritin, soluble transferrin receptor (sTfR), zinc, serum folate, red blood cell folate and vitamin B12 were estimated using multivariable linear regression. Factors related to malarial immunity including age, altitude and presence of hemoglobinopathies were examined as effect modifiers. Serum ferritin, sTfR and zinc were adjusted for inflammation using the BRINDA method. Malaria infection was associated with 68 % (95 % CI 51, 86), 28 % (18, 40) and 34 % (13, 45) greater inflammation-adjusted ferritin in PSC, SAC and WRA, respectively (P < 0·001 for each). In PSC, the positive association was stronger in younger children, high altitude and children who were not carriers of the sickle cell trait. In PSC and SAC, sTfR was elevated (+ 25 % (16, 29) and + 15 % (9, 22) respectively, P < 0·001). Serum folate and erythrocyte folate were elevated in WRA with malaria (+ 18 % (3, 35) and + 11 % (1, 23), P = 0·01 and P = 0·003 respectively). Malaria affects the interpretation of micronutrient biomarker concentrations, and examining factors related to malarial immunity may be informative.

Objectives/Goals: Manual skin assessment in chronic graft-versus-host disease (cGVHD) can be time consuming and inconsistent (>20% affected area) even for experts. Building on previous work we explore methods to use unmarked photos to train artificial intelligence (AI) models, aiming to improve performance by expanding and diversifying the training data without additional burden on experts. Methods/Study Population: Common to many medical imaging projects, we have a small number of expert-marked patient photos (N = 36, n = 360), and many unmarked photos (N = 337, n = 25,842). Dark skin (Fitzpatrick type 4+) is underrepresented in both sets; 11% of patients in the marked set and 9% in the unmarked set. In addition, a set of 20 expert-marked photos from 20 patients were withheld from training to assess model performance, with 20% dark skin type. Our gold standard markings were manual contours around affected skin by a trained expert. Three AI training methods were tested. Our established baseline uses only the small number of marked photos (supervised method). The semi-supervised method uses a mix of marked and unmarked photos with human feedback. The self-supervised method uses only unmarked photos without any human feedback. Results/Anticipated Results: We evaluated performance by comparing predicted skin areas with expert markings. The error was given by the absolute difference between the percentage areas marked by the AI model and expert, where lower is better. Across all test patients, the median error was 19% (interquartile range 6 – 34) for the supervised method and 10% (5 – 23) for the semi-supervised method, which incorporated unmarked photos from 83 patients. On dark skin types, the median error was 36% (18 – 62) for supervised and 28% (14 – 52) for semi-supervised, compared to a median error on light skin of 18% (5 – 26) for supervised and 7% (4 – 17) for semi-supervised. Self-supervised, using all 337 unmarked patients, is expected to further improve performance and consistency due to increased data diversity. Full results will be presented at the meeting. Discussion/Significance of Impact: By automating skin assessment for cGVHD, AI could improve accuracy and consistency compared to manual methods. If translated to clinical use, this would ease clinical burden and scale to large patient cohorts. Future work will focus on ensuring equitable performance across all skin types, providing fair and accurate assessments for every patient.