University of Iceland

Background Glial fibrillary acidic protein (GFAP) is an astrocytic cytoskeletal protein and a promising blood biomarker for Alzheimer's disease (AD) and other neurodegenerative diseases. To date, the genetic architecture of plasma GFAP has not been characterized. We conducted a multi‐ancestry meta‐analyses of genome‐wide association studies (GWAS) in diverse population‐based cohorts to identify genetic variants associated with plasma levels of GFAP and to investigate their implication for neurological diseases. Methods Circulating GFAP levels were assayed by ultrasensitive single molecule array (Simoa) immunoassay. We performed ancestry‐specific and ancestry‐combined meta‐analyses of GWAS in 17,028 individuals (mean age: 64.5 years; 60% women) of diverse ancestry (African (AA), N=1163; Hispanic (HA), N=7405; European (EA), N=8460) from 12 community‐based cohorts. Results In the combined sample, we identified 7 loci associated with plasma GFAP. These were located in or near genes encoding APOE, CAPN1, CAPN2, GFAP, TMEM106B, PPP1R3C, and LOC105375672. The strongest association was with the APOE4 variant rs429358 (P=2.6x10‐40). However, this association was heterogeneous by ancestry, with stronger effects in EA, weaker in HA, and no association in AA. Ancestry‐specific meta‐analyses identified additional loci, including ZFAND3 in EA (Figure 1) and CAPN2 in HA (Figure 2), which were not observed in the other ancestries. No locus reached genome‐wide significance in AA. Using fine‐mapping methods incorporating functional annotation, gene expression, and linkage disequilibrium, we further prioritized potential causal variants and genes at the identified loci and investigated their implication for major neurological diseases, including AD. Conclusion Our GWAS of plasma GFAP has identified multiple loci with ancestry‐specific or cross‐ancestry effects. Our multi‐ancestry, integrative approach extends knowledge about the biology of GFAP and its shared genetic architecture with neurological disease.

Background Epigenetic age acceleration (EAA) is a valuable tool for predicting all‐cause mortality and assessing disease risk. Differences in EAA reflects biologic aging (BA) and suggests underlying differences in morbidities. We examined whether EAA moderated the association of dementia to mortality risk. Method Analyses are based on data from the Age, Gene/Environment Susceptibility‐ Reykjavik Study (AGES‐RS) study (n = 2602, 57.5% females, mean age = 75.8 years, 30% demented). Data were collected from 2002 to 2006, with mortality data available until 2015. EAA was computed with the DunedinPACE algorithm and grouped as follows: EAA scores > 1 SD of the mean were labeled as ‘fast’, EAA < 1 SD as ‘slow’, and EAA scores within ±1 SD as ‘average’. Dementia was ascertained in a study exam or through follow‐up of health records. Cox proportional hazard analysis was used to evaluate the hazard ratio (HR, 95%CI) of mortality and the interaction between EAA and dementia. Models were adjusted for sex, chronological age, education, smoking, diabetes and stroke. Result Fast agers had higher mortality HRs than average agers (Figure 1A). Fast agers with dementia had a significantly increased risk of death (interaction p‐value = 4.08×10⁻²) compared to non‐demented average agers (Figure 1C); this interaction was reduced after adjusting for total brain volume (Figure 1B). Conversely, slow aging was associated with reduced mortality risk (HR = 0.71, 95% CI: 0.52‐0.98, p‐value = 3.62×10⁻²); however slow agers with dementia had risk estimates similar to non‐demented average agers (interaction p‐value = 1.14×10⁻¹) but lower than demented fast agers. Controlling for total brain volume reduced the differences between slow and average agers with and without dementia. Conclusion Individuals aging at a faster rate than their chronological age were at a higher risk for death. The rate of BA had an effect on the association between dementia and mortality. The association was partially explained by brain atrophy, suggesting other factors related to dementia contributing to mortality risk in people with dementia.

Background Blood‐based biomarkers, glial fibrillary acidic protein (GFAP) and neurofilament light chain (NfL), show potential for dementia risk stratification. Yet, their predictive performance for incident dementia in heterogeneous older population with multi‐morbidities, is not well tested. This study evaluates their predictive ability for incident dementia accounting for differences in factors affecting health. Method Data from 910 baseline non‐demented participants of the population‐based Age, Gene/Environment Susceptibility‐Reykjavik Study (AGES‐RS) from 2002 to 2015 (average age 76.6 years, 54.2% female, 6.7% incident dementia) were analyzed. Factors (n = 360) (sociodemographic, clinical, sensory, cardiometabolic, musculoskeletal, medical history, and lifestyle) clustered into 34 groups. Incident dementia cases were identified following the baseline exam until December 2015. Logistic regression models, with 5000 bootstraps, evaluated the predictive power of GFAP and NfL for incident dementia. Predictor sets included: biomarker only; plus age and sex; plus basic factors (education, APOE e4, eGFR); and a full model with associated peripheral clusters. An additional model with age and sex served as the reference for assessing performance without biomarker. Model effectiveness was assessed using area under the precision‐recall curve (AUPRC), precision, and sensitivity (recall), suitable for low dementia prevalence contexts. Result Biomarker levels between cases and non‐cases showed significant overlap (Figure 1a). In predictive performance for incident dementia, NfL outperformed GFAP in the biomarker‐only model (AUPRCs, GFAP: 0.17, NfL: 0.20). Integrating peripheral clusters associated with these biomarkers improved the models’ performance, surpassing both biomarker‐only and basic models, especially for NfL (Figure 1b). Precision rates for the final GFAP and NfL models were 0.304 and 0.61, respectively, reflecting approximately 30% and 61% prediction accuracy for incident dementia. The GFAP’s AUPRC improved to 0.339, accurately identifying 5 out of 100 incident dementia cases, while the NfL’s AUPRC was 0.367, predicted 17% of cases. Conclusion Our findings indicate that due to the low prevalence of dementia in the general population, using plasma GFAP and NfL as dementia screening tools is not highly effective. However, the addition of peripheral factor clusters does enhance their predictive performance, with NfL having a slightly better performance over GFAP.

Migratory birds depend on a suite of sites across their annual cycles, making them vulnerable to a wide variety of anthropogenic pressures. Current area‐based conservation measures have been found inadequate to safeguard migratory birds, in part due to a lack of consideration for the connectivity between sites mediated by the movements of individuals. To address this issue, we develop a network analysis integrating different types of individual movement data for a migratory shorebird, the Black‐tailed Godwit (Limosa limosa), across the East Atlantic Flyway. Leveraging metal‐ring recoveries, colour‐ring re‐sightings and satellite tracking from over 10,000 individual godwits, we quantify variation in connectivity between sites across the migratory range, using two weighted metrics to address sampling biases. Colour‐ring re‐sightings provided the largest number of sites (70%) and links (60% and 43% per season) overall, followed by tracking data (50% of sites, 49% and 63% of links per season) and ring recoveries (25% of sites, <1% of links per season), with clear regional variation in datatype contributions. Sampling completeness of the network structure varied with longitude, with information particularly lacking in central and eastern countries of both Europe and Africa. We identified 49 sites playing a disproportionate role in the site network, each with direct connections to 48 (interquartile range 32–84) other sites, on average. Just 23 (47%) top sites are formally recognized for their international importance for Black‐tailed Godwits, and 33 (67%) were robust to sampling incompleteness. Across all 1058 sites, 20% lacked protected area coverage, and per site, 44% (44% ± SD) of bird relocations fell within protected areas. Integrating multiple sources of data improved geographical coverage and completeness of the site network, allowing us to quantify the importance of sites in terms of connectivity across the flyway. Our results highlight shortcomings of existing area‐based conservation measures and add value to ongoing efforts to identify important sites for migratory birds. Policy implications. The increasing availability of individual movement data provides valuable opportunities to reveal the inter‐dependence of sites used by migratory species, which can help identify priority areas and facilitate flyway‐scale management.

Despite the amount of data sampled by current and upcoming satellite missions, the temporal evolution of oceanic turbulence is for now poorly understood. Here we use state-of-the-art satellite measurements of Sea Surface Height timeseries, combined with global oceanic re-analysis and idealized numerical simulations to study the properties of the frequency spectrum for oceanic turbulence. We show that the meridional gradient of Potential Vorticity, that supports planetary Rossby Waves, and the barotropic current are the main parameters that shape the frequency spectrum of Sea Surface Height as observed by satellite altimetry.

Major curricular reforms have appeared via the revised Icelandic National Curriculum for Secondary Education released in 2015. Few studies in Iceland have explored teachers’ views and beliefs of motivational strategies for English teaching or considered the impact of the national curriculum on the learning experience in Iceland (Ingvarsdóttir, 2018, Sigurgeirsson et al., 2018). Consequently, the primary goal of this study is to capture English teachers’ views at the secondary level regarding (a) the role of motivational strategies in their own classroom practice and (b) its influence on the language learning experience of their students. Focus group interviews with beginning and experienced teachers, i.e., teachers in their first 3 years of teaching and those with 10 or more years of teaching experience, comprise the primary data. Interviews were transcribed and analyzed for significant themes in relation to the research focus (Braun & Clarke, 2021). Results are explored in relation to Dörnyei’s L2 Experience dimension, particularly regarding teacher practices (Dörnyei, 2009) relating to how English teachers in Iceland understand their role in learners’ L2 experience, how they think about their students’ motivation, how they develop their own motivational practices, and the interplay between these aspects.

Objectives Randomized clinical trials informing clinical practice (e.g., like large, pragmatic, and late-phase trials) should ideally mostly use harmonized outcomes that are important to patients, family members, clinicians, and researchers. Core outcome sets for specific subsets of ICU patients exist, for example, respiratory failure, delirium, and COVID-19, but not for ICU patients in general. Accordingly, we aimed to develop a core outcome set for adult general ICU patients. Design We developed a core outcome set in Denmark following the Core Outcome Measures in Effectiveness Trials Handbook. We used a modified Delphi consensus process with multiple methods design, including literature review, survey, semi-structured interviews, and discussions with initially five Danish research panels. The core outcome set was internationally validated and revised based on feedback from research panels in all countries. Setting There were five Danish research panels and 17 panels in 13 other countries. Interviews and the three-round Delphi survey was conducted in Denmark, followed by validation of the core outcome set across 14 countries in Europe, Australasia, and India. Subjects Adult ICU survivors, family members, clinicians, and researchers. Interventions None. Measurements and Main Results We identified 329 published outcomes, of which 50 were included in the 264 participant Delphi survey. In semi-structured interviews of 82, no additional outcomes were added. The first Delphi survey round was completed by 249 (94%) participants, and 202 (82%) contributed to the third and final round. The initial core outcome set comprised six outcomes. International validation involved 217 research panel members and resulted in the final core outcome set comprising survival, free of life support, free of delirium, out of hospital, health-related quality of life, and cognitive function. Conclusions We developed and internationally validated a core outcome set with six core outcomes to be used in research, specifically clinical trials involving adult general ICU patients.

Purpose Lamb meat is a crucial protein source in Icelanders’ diets. Extensive grazing lands, locally grown hay feed, and traditional farming methods are often used as arguments for Icelandic lamb meat’s environmental friendliness. However, no life cycle assessment (LCA) study exists to corroborate these arguments. This study conducts a national-level LCA to evaluate the environmental performance of lamb meat based on two functional units to identify key hotspot processes in its production. Method This study conducts a cradle-to-farm gate LCA at the national level for two functional units: 1 kg of edible lamb meat (ELM) and 100 g of ELM protein produced in Iceland in 2019. The multifunctionality between wool and meat is handled using mass allocation. The environmental impacts were estimated using the ReCiPe 2016 v1.1 mid-point (H) impact assessment method, emphasizing selected environmental categories: global warming, fossil resource scarcity, land use, and terrestrial ecotoxicity. In addition, the study conducts a scenario-based variability analysis by taking minimum and maximum values of inventory data to estimate the possible range of environmental impacts. Lastly, an overall uncertainty analysis and a global sensitivity analysis of the key hotspot process shed light on the variability and sensitivity of the LCA results. Result and discussion For the 18 ReCiPe impact categories, animal and feed (hay) production are the hotspot processes, followed by feed (grazing) as a hotspot for land use. The global warming impact for 1 kg of ELM ranges 41–53 kg CO2 equivalent, and for 100 g of protein, 19–29 kg CO2 equivalent. Fossil resource scarcity impact for 1 kg of ELM impact ranges 2.5–3.6 kg oil equivalent, and for 100 g protein, 1–2 kg oil equivalent. Terrestrial ecotoxicity impacts for 1 kg of ELM range 46–69 kg 1,4-DCB, and for 100 g protein, 21–37 kg 1,4-DCB. Lastly, land use impacts for 1 kg of ELM range 562–2448 m²a crop equivalent, and for 100 g protein, 261–1324 m²a crop equivalent. Conclusion With its traditional farming practices, Icelandic lamb meat production is close to an extensive farming system, which is in line with its higher global warming impact per kg ELM. Additionally, due to low hay yield and high fertilization rate, the impact on other impact categories is still higher compared to an extensive system. This perspective shifts when analyzed per 100 g of protein, where it performs close to the global average.

Achieving Land cover change detection (LCCD) through remote sensed images (RSIs) is important in the observation of the changes on the Earth’s surface. In such detection, spectral-reflectance noise and the uncertainty of the imaging external conditions for the bitemporal RSIs usually caused some salt and pepper noisy pixels in the results and reduces the change detection accuracy. In this paper, a hierarchical feature-fusion triple network (HFTN) is proposed to improve the performance of LCCD with RSIs. Overall, the proposed HFTN aims at learning representative features to improve the change detection performance via two feature learning enhancement strategies and a hierarchical feature fusion mechanism. First, an image feature difference model is proposed for generating the input feature for the middle branch and guiding the learning performance. Second, a progressive denoising module is proposed and applied to each temporal image to reduce the noise before feeding the features into the backbone of the proposed HFTN. Finally, a hierarchical feature fusion module is proposed to fuse the learned deep feature for generating a change-magnitude image. Additionally, multi-scale convolution, cross-scale fusion, and a shared weight are adopted in the backbone of the proposed HFTN to further enhance the feature learning performance. Compared with eight state-of-the-art methods, experimental results verified the feasibility and superiority of the proposed HFTN for LCCD with RSIs. For example, the proposed HFTN achieved improvement rates of approximately 0.43%-11.83% for overall accuracy (OA) and 0.11%-4.81% for false alarms (FA) across six pairs of real RSIs. The code can be available at: https://github.com/ImgSciGroup/HFTN-NET.git.

Background Melanoma cells frequently dedifferentiate in response to inflammation which can increase responses to certain cytokines. Interferon-γ (IFNγ) is an integral part of the anti-tumor immune response and can directly induce both differentiational changes and expression of immunosuppressive proteins in melanoma cells. How the differentiation status of melanoma cells affects IFNγ responses remains unclear. Methods Dedifferentiation of melanoma cells was induced via either siRNA or shRNA mediated MITF knockdown and the cells were subsequently treated with IFNγ. Effects of MITF knockdown and IFNγ treatment on gene expression were evaluated via qPCR and RNA sequencing. A Luminex assay was used to analyze the effects of dedifferentiation and IFNγ treatment on cytokine secretion. Effects on PD-L1 protein expression were analyzed via flow cytometry and western blotting. Inhibition of the JAK kinases, NF-κB and STAT3 with small molecule inhibitors, and siRNA mediated knockdown of STAT1 and IRF1 was applied to investigate the molecular mechanism behind IFNγ induced PD-L1 expression in dedifferentiated melanoma cells. The effects of inhibitor treatments and siRNA mediated knockdowns were evaluated via qPCR and western blotting. Bioinformatic analysis of publicly available RNA sequencing data, consisting of 45 patient derived melanoma cell lines, with or without IFNγ treatment, was conducted to assess the generalizability of the in vitro results. Results Dedifferentiation renders 624Mel melanoma cells hypersensitive to IFNγ stimulation in a context-dependent manner, resulting in non-additive upregulation of IFNγ-induced genes, increased PD-L1 protein expression and amplified secretion of CCL2, CXCL10 and IL-10. Furthermore, the intensified PD-L1 protein expression occurs through the JAK-STAT1-IRF1 axis. Lastly, dedifferentiated patient derived melanoma cell lines showed enhanced inflammatory signaling in response to IFNγ compared to differentiated cells, and tended to have higher PD-L1 expression, associated with increased IRF1 expression and activity. Conclusions Together, these findings indicate the existence of a molecular context linking dedifferentiation and IFNγ signaling in melanoma which may lead to immune evasion. Additionally, the variability in PD-L1 expression among MITF low and MITF high cells suggests that high IFNγ-induced PD-L1 expression associates with enhanced inflammatory gene expression. These results imply that modulating melanoma differentiation may help shape IFNγ responsiveness. Graphical Abstract

Empirical studies worldwide show that warming has variable effects on plant litter decomposition, leaving the overall impact of climate change on decomposition uncertain. We conducted a meta‐analysis of 109 experimental warming studies across seven continents, using natural and standardised plant material, to assess the overarching effect of warming on litter decomposition and identify potential moderating factors. We determined that at least 5.2° of warming is required for a significant increase in decomposition. Overall, warming did not have a significant effect on decomposition at a global scale. However, we found that warming reduced decomposition in warmer, low‐moisture areas, while it slightly increased decomposition in colder regions, although this increase was not significant. This is particularly relevant given the past decade's global warming trend at higher latitudes where a large proportion of terrestrial carbon is stored. Future changes in vegetation towards plants with lower litter quality, which we show were likely to be more sensitive to warming, could increase carbon release and reduce the amount of organic matter building up in the soil. Our findings highlight how the interplay between warming, environmental conditions, and litter characteristics improves predictions of warming's impact on ecosystem processes, emphasising the importance of considering context‐specific factors.

Background The Convention on the Rights of the Child states that children need to be protected from ‘any work that is likely to be hazardous or to interfere with the child’s education, or to be harmful to the child’s health or physical, mental, spiritual, moral or social development’. We aimed to determine the prevalence and correlates of child labour in five low-income African countries using the sixth wave of UNICEF Multiple Indicator Cluster Surveys (MICS6). Methods Data on child labour, reported by the household respondent for a randomly selected child (5–17 years), were extracted from MICS6 reports from Chad, Guinea Bissau, Malawi, Sierra Leone and Togo. Prevalence rates were extracted from three categories of child labour (household chores, economic activity and hazardous work) stratified by gender, age, wealth and residence. Results 140 598 children aged 5–17 years (61.2% 5–11; 22.0% 12–14; 16.8% 15–17) were included in the reports; 59 090 (42%) were engaged in child labour. The highest proportion of child labourers by age was 12–14 years old (61.1%) followed by 15–17 years old (51.1%) and 5–11 years old (36.1%). There were differences between countries, with Chad having the highest proportion of working children. Gender differences in working were limited (43.9% boys vs 40.3% girls); rural children were almost twice as likely to be working compared with urban children (47.5% rural vs 25.6% urban) as were children in the poorest quintile compared with those in the wealthiest quintile (46.9% vs 23.7%). Over a third (35.3%) of working children were exposed to hazardous conditions. Older, male, rural or poor children were over-represented among those in hazardous work. Conclusions Almost half of all children in these five sub-Saharan African countries are engaged in labour, of which one-third are in hazardous work. MICS6 surveys do not report on working children’s health; however, working puts their health and development at risk.