Repository for Publications and Research Data
News from the ETH Library
Recently Added
Urban tree canopy fragmentation and child mortality in low- and middle-income countries
Item type: Journal Article
Chi, Dengkai; Richards, Daniel; Manoli, Gabriele; et al. (2026)
Exposure to urban green spaces is associated with lower risks of several leading causes of under-5 mortality (U5M). While urban development improves infrastructure, healthcare access, and socioeconomic conditions, rapid urbanization in low- and middle-income countries (LMICs) has also accelerated the loss and fragmentation of green spaces. Such fragmentation may weaken ecosystem services critical for child health and increase exposure to vector-borne infections, yet its implications for U5M remain poorly understood and must be weighed against the benefits of urbanization. Here, we quantified associations between urban tree canopy cover and canopy aggregation with all-cause U5M using individual-level longitudinal mortality data between 2000 and 2019 from 57 LMICs, and identified priority areas for expanding canopy cover and improving canopy aggregation across 98 LMICs using counterfactual scenario analyses. The association between canopy cover and U5M was nonlinear, with risk declining steeply at low-to-intermediate levels of canopy cover. At a given level of canopy cover, every 1% increase in tree-to-tree adjacency was associated with a 0.17% reduction in mortality risk (hazard ratio: 0.9983, 95% CI: 0.9972 to 0.9995). The largest contiguous high-priority areas were concentrated in the West Sudanian Savanna, the eastern Punjab region of Pakistan, and the Ganges-Brahmaputra Delta in Bangladesh, where high U5M burdens align with substantial potential to improve tree canopy. This study establishes a robust association between tree canopy characteristics and U5M across LMICs. We identify canopy cover and fragmentation as potentially modifiable risk factors, highlighting important implications of urban greening design for delivering cobenefits for child health and climate mitigation.
Protocol for the REVELIO test-track pilot study: A randomised, controlled, single-centre trial in healthy recreational cannabis users investigating real-time in-vehicle detection of cannabis-impaired driving
Item type: Journal Article
Deuber, Robin; Bechny, Michal; Heck, Christoph Karl; et al. (2026)
Background
Driving under the influence of cannabis is associated with impaired cognitive and psychomotor performance and an increased risk of traffic accidents. Reliable real-time in-vehicle systems for detecting cannabis-related driving impairment are currently lacking; but hold great potential for improving road safety.
Methods
This protocol describes the REVELIO test-track pilot study: a randomised, controlled, open-label, interventional, single-centre trial. The study assesses the feasibility and methodological requirements for developing and evaluating a multimodal in-vehicle detection approach using vehicle and driver state data. A total of 45 healthy recreational cannabis users will be enrolled and randomly allocated to an intervention or a reference control group. During the main study day, all participants will undergo biological sampling for tetrahydrocannabinol (THC) and related metabolites, as well as pre-driving assessments, followed by a sober baseline driving session on a closed test track using a dual-pedal vehicle with a certified driving instructor onboard. Participants in the intervention group will then receive a single controlled inhalative cannabis dose (target 0.67 mg THC per kg body weight), while the reference group will receive no cannabis. All participants will subsequently complete three additional standardized 50-minute driving sessions at predefined time points up to approximately six hours after administration, following identical schedules to enable within- and between-group comparisons. Between driving sessions, structured breaks will include recovery periods, repeated biological sampling, and traffic-medical, traffic-psychological, and pre-driving performance assessments, to characterise the temporal dynamics of cannabis-related impairment.
Discussion
Multimodal data will be collected, including vehicle controller area network (CAN) data, driver monitoring camera (DMC) data, physiological signals using wearables, and biological samples (capillary blood, breath, oral fluid. Machine-learning–based models will be developed and evaluated to distinguish sober from cannabis-influenced driving states under controlled conditions. Secondary analyses will examine changes in driving performance over time and associations between functional measures and biological THC concentrations. As an exploratory pilot study conducted on a secured test track, the protocol aims to generate standardized reference data and quantitative performance metrics to inform both feasibility and system design considerations.
Glow Up: a single-centre observational study protocol to develop digital biomarkers for prediabetes screening and digital phenotyping
Item type: Journal Article
Brügger, Victoria; Fuchs, Magdalena; Jin, Qiuhan; et al. (2026)
Introduction
Prediabetes, a key precursor to type 2 diabetes, is highly prevalent and underdiagnosed, particularly among adults aged ≥45 years with elevated body mass index (BMI). Early detection is critical because lifestyle interventions can delay or prevent progression to type 2 diabetes. The Glucose Observation and Wearable Use for Prevention (Glow Up) study aims to (1) test the feasibility of a digital biomarker for pre-diabetes screening using wearable- and smartphone-derived lifestyle factors (eg, sleep, physical activity and nutrition patterns) in daily life and to (2) characterise individual- and metabolic subgroup-level variability in lifestyle factors and glycaemic control. Specifically, we aim to examine how lifestyle factors relate to glycaemic status and identify personalised predictors of early metabolic dysregulation.
Methods and analysis
Glow Up is a prospective, single-centre, observational case-control study conducted in Switzerland. Adults (n=200) aged ≥45 years with body mass index (BMI) ≥25 kg/m² will be recruited, including n=100 individuals with pre-diabetes and n=100 age- and sex-matched normoglycaemic controls. Participants will undergo 4 weeks of continuous monitoring using a blinded continuous glucose monitor (CGM) and commercial- and medical-grade wearables; for example, capturing physical activity, sleep and physiological markers (heart rate variability, heart rate and skin temperature), in addition to completing daily image-based meal logs using smartphones. Glycated haemoglobin (HbA1c), fasting plasma glucose (FPG) and body anthropometrics will be collected at the enrolment visit (T0) and repeated at follow-up (T3), approximately 5 weeks apart. The primary outcome is glycaemic status at T3, classified from HbA1c and FPG using prespecified thresholds. Secondary outcomes include CGM metrics, lifestyle profiling (sleep, physical activity, stress and nutrition) and adherence to image-based meal logging.
Ethics and dissemination
The study received ethics approval from the Ethics Committee of Eastern Switzerland on 15 May 2025 (BASEC ID: 2025–00972). Written informed consent is obtained from all participants before any study procedures are performed. Findings will be disseminated as articles in international peer-reviewed journals and as posters and oral presentations at national and international conferences. Summaries will be provided to funders and collaborators, and personalised reports to participants.
Trial registration number NCT07373418.
Advancing chemical reaction prediction in data-scarce drug discovery with active and geometric deep learning
Item type: Journal Article
Minot, Mason; Stenzhorn, Yannick; Wolfard, Jens; et al. (2026)
The success of generative molecular design depends on the synthetic accessibility of the molecules it creates. While computational models can help predict chemical reactions, their current effectiveness is limited by a lack of reference data and the difficulty of accurately predicting reaction regioselectivity. Here, to address this, we present an integrated closed-loop workflow that combines efficient reaction data acquisition with high-accuracy deep learning. We developed an active learning strategy, driven by a tree-based ensemble, to guide the acquisition of the most informative laboratory experiments. This experimental campaign yielded a diverse benchmark for C–H borylation. The bespoke dataset allowed us to train and evaluate geometric graph neural networks. We demonstrate that augmenting these symmetry-aware models with self-supervised auxiliary tasks consistently improves performance for both reaction outcome prediction and regioselectivity prediction. Prospective tests on unseen substrates, featuring challenging N-heteroaryl motifs, pinpointed the correct borylation positions in all cases, illustrating the practical utility of our workflow for efficient drug optimization.
Fungal inoculation of human fecal-woodchip substrates produces species-specific effects on indicator bacteria and substrate degradation
Item type: Journal Article
Peter, Jules; Clavijo Daza, Adriana Marcela; Tilley, Elizabeth (2026)
Safely treating human feces in decentralized, waterless toilets remains a sanitation challenge. Filamentous fungi degrade lignocellulose and stay active under the dry conditions that limit bacterial degradation, yet their potential to treat human feces has rarely been quantified. In this exploratory screening study, we inoculated four filamentous fungal species represented by five isolates (two Pleurotus ostreatus isolates and one isolate each of Sordaria fimicola, Trichoderma harzianum, and Ganoderma lucidum) onto fresh human feces mixed with beechwood chips and followed colonization, mold contamination, pH, fecal indicator bacteria (Escherichia coli and Enterococcus spp.), and dry weight over 14 days. Species differed markedly. Sordaria fimicola colonized every sample and fully suppressed competing molds but raised bacterial concentrations and pH. Ganoderma lucidum and both P. ostreatus isolates kept the substrate acidic and suppressed fecal odor; where it established, G. lucidum reduced both indicator bacteria below detection, though its colonization was slow and unreliable. Trichoderma harzianum grew fast but died back after 8 days. Fungal colonization curbed spontaneous molds and for S. fimicola and T. harzianum accelerated substrate degradation, revealing species-specific trade-offs between reliable colonization, suppression of fecal indicator bacteria, and pH control, and pointing to promising candidates and mechanisms for biological fecal treatment.