Search

BackgroundThe number of deaths from COVID-19 continues to surge worldwide. In particular, if a patient’s condition is sufficiently severe to require invasive ventilation, it is more likely to lead to death than to recovery. ObjectiveThe goal of our study was to analyze the factors related to COVID-19 severity in patients and to develop an artificial intelligence (AI) model to predict the severity of COVID-19 at an early stage. MethodsWe developed an AI model that predicts severity based on data from 5601 COVID-19 patients from all national and regional hospitals across South Korea as of April 2020. The clinical severity of COVID-19 was divided into two categories: low and high severity. The condition of patients in the low-severity group corresponded to no limit of activity, oxygen support with nasal prong or facial mask, and noninvasive ventilation. The condition of patients in the high-severity group corresponded to invasive ventilation, multi-organ failure with extracorporeal membrane oxygenation required, and death. For the AI model input, we used 37 variables from the medical records, including basic patient information, a physical index, initial examination findings, clinical findings, comorbid diseases, and general blood test results at an early stage. Feature importance analysis was performed with AdaBoost, random forest, and eXtreme Gradient Boosting (XGBoost); the AI model for predicting COVID-19 severity among patients was developed with a 5-layer deep neural network (DNN) with the 20 most important features, which were selected based on ranked feature importance analysis of 37 features from the comprehensive data set. The selection procedure was performed using sensitivity, specificity, accuracy, balanced accuracy, and area under the curve (AUC). ResultsWe found that age was the most important factor for predicting disease severity, followed by lymphocyte level, platelet count, and shortness of breath or dyspnea. Our proposed 5-layer DNN with the 20 most important features provided high sensitivity (90.2%), specificity (90.4%), accuracy (90.4%), balanced accuracy (90.3%), and AUC (0.96). ConclusionsOur proposed AI model with the selected features was able to predict the severity of COVID-19 accurately. We also made a web application so that anyone can access the model. We believe that sharing the AI model with the public will be helpful in validating and improving its performance.

BackgroundCOVID-19, which is accompanied by acute respiratory distress, multiple organ failure, and death, has spread worldwide much faster than previously thought. However, at present, it has limited treatments. ObjectiveTo overcome this issue, we developed an artificial intelligence (AI) model of COVID-19, named EDRnet (ensemble learning model based on deep neural network and random forest models), to predict in-hospital mortality using a routine blood sample at the time of hospital admission. MethodsWe selected 28 blood biomarkers and used the age and gender information of patients as model inputs. To improve the mortality prediction, we adopted an ensemble approach combining deep neural network and random forest models. We trained our model with a database of blood samples from 361 COVID-19 patients in Wuhan, China, and applied it to 106 COVID-19 patients in three Korean medical institutions. ResultsIn the testing data sets, EDRnet provided high sensitivity (100%), specificity (91%), and accuracy (92%). To extend the number of patient data points, we developed a web application (BeatCOVID19) where anyone can access the model to predict mortality and can register his or her own blood laboratory results. ConclusionsOur new AI model, EDRnet, accurately predicts the mortality rate for COVID-19. It is publicly available and aims to help health care providers fight COVID-19 and improve patients’ outcomes.

Walking is a fundamental aspect of daily life and exercise, with clinical benefits for cardiovascular health and muscle strength. However, accurately measuring energy efficiency during walking poses challenges due to equipment and spatial constraints. In this study, we proposed the cadence-based energy expenditure index (cEEI) and analyzed its correlation with the previously proposed index for measuring energy expenditure under various gait conditions. We enrolled 15 healthy participants and conducted an experimental protocol on a treadmill to measure the following energy expenditure-related indices: oxygen cost index (OCI), energy expenditure index (EEI), and cEEI. The participants underwent stages of walking at different speeds and inclinations that comply with the modified Bruce protocol, whereas their heart rate, oxygen uptake, and cadence were recorded. Participants showed significant increases in heart rate, oxygen uptake, and cadence with higher walking speeds and inclinations. Correlation analysis revealed strong associations between cEEI and OCI, especially during walking conditions. Bland–Altman plots and interclass correlation coefficient analysis demonstrated a favorable agreement between cEEI and OCI, outperforming EEI. In conclusion, this study proposes cEEI as a reliable metric for estimating energy expenditure during walking by proving a strong correlation and agreement with OCI across various gait conditions. This suggests the potential for cEEI to provide real-time, individualized feedback on energy expenditure during walking, facilitating more personalized exercise prescriptions. NEW & NOTEWORTHY: We proposed cEEI as a reliable metric for estimating energy expenditure under various walking conditions. Its demonstrated accuracy and potential for real-time individualized feedback could revolutionize personalized exercise prescriptions for diverse sexes and age groups, as well as individuals with disabilities requiring real-time physical monitoring and home care. [ABSTRACT FROM AUTHOR]

Biological invasion consists of spatially and temporally varying stages, accompanied by ecological and evolutionary changes. Understanding the genomics underlying invasion dynamics provides critical insights into the geographic sources and genetic diversity, contributing to successful invasions across space and time. Here, we used genomic data and model‐based approaches to characterize the invasion dynamics of Hypochaeris radicata L., a noxious weed in Korea. Genetic diversity and assignment patterns were investigated using 3563 SNPs of 283 individuals sampled from 22 populations. We employed a coalescent‐based simulation method to estimate demographic changes for each population and inferred colonization history using both phylogenetic and population genetic model‐based approaches. Our data suggest that H. radicata has been repeatedly been introduced to Korea from multiple genetic sources within the last 50 years, experiencing weak population bottlenecks followed by subsequent population expansions. These findings highlight the potential for further range expansion, particularly in the presence of human‐mediated dispersal. Our study represents the first population‐level genomic research documenting the invasion dynamics of the successful worldwide invader, H. radicata, outside of Europe. [ABSTRACT FROM AUTHOR]

BackgroundIt is increasingly evident that patients with chronic kidney disease (CKD) are more likely to die from heart disease than kidney failure. This study evaluated whether pre- dialysis CKD is an independent risk factor for coronary artery calcium (CAC).MethodsA total of 544 consecutive patients who underwent CAC scoring were analyzed. Eleven patients requiring hemodialysis were excluded. Patients were divided into three groups: normal glomerular filtration rate (GFR) (GFR > 90 mL/min/1.73 m²), mild CKD (90 ≥ GFR > 60 mL/min/1.73 m²), and moderate CKD (60 ≥ GFR > 30 mL/min/1.73 m²). Continuous and categorical variables were compared using analysis of variance and the χ² statistic. A multiple logistic regression model was used for detecting the association between total CAC score and GFR. An unadjusted model was used, followed by a second model adjusted for covariates known to be related to CAC. Another multivariable binary logistic model predicting the presence of CAC (>10) was performed and odds of incidence of CAC (>10) were calculated among the three GFR subgroups.ResultsAfter adjustment for covariates, patients with mild CKD had mean CAC scores 175 points higher than those with the referent normal GFR (P = 0.048), while those with moderate CKD had mean CAC scores 693 points higher than the referent (P < 0.001). After adjustment for covariates, patients with mild CKD were found to be 2.2 times more likely (95% confidence interval 1.3-3.7, P = 0.004) and patients with moderate CKD were 6.4 times more likely (95% confidence interval 2.9-14.3, P < 0.001) to have incident CAC compared with the group with normal GFR.ConclusionMild and moderate pre-dialysis CKD are independent risk factors for increased mean and incident CAC.

The mitigation of fragmentation due to high density road network has been a hot topic among environmentalistsand road construction engineers of South Korea. Over the last ten years 92 wildlife passages, 55 ecoducts and 37 wildlife underpasses, have been constructed on existing roads, and many more will be constructed in the future (Ministry of Environment of the Republic of Korea, 2006). We are at an early stage of data collection on wildlife vehicle collision and the role of traditionally non-wildlife-engineered passages, such as underpasses including bridges, culverts, and human underpasses, for wildlife passages.The objective of this study was to analyze the effectiveness of the number, size, and density of non-wildlife-engineered passages. This study employed three monitoring methods: wildlife vehicle collisions, the passages use ratio (Servheen, 2003) and radio telemetry. The effectiveness of such unintended wildlife passage was evaluated by using the relationship between monthly wildlife vehicle collision data, number of usable passages, use rate of passages, and passage density.The number of usable passages represents all crossing structures after excluding those inundated circular culverts during summer season, since they are impassable for most wildlife species. The use rates of wildlife passages were collected from 14 underpasses. They were seven circular culverts, two box culverts, and four human underpasses, and were selected from 31 structures constructed on a 6.6km segment of a four-lane highway. The landscape of study area mainly consists of rice fields on an alluvial plane and scattered forest, and the road runs along the stream. Every passage has similar surroundings. Wildlife monitoring was carried out for 12 months, from Sept. 2005 to Aug. 2006; using camera traps (an average of 239 camera operating days). The number of recorded mammals was 2,593, consisting of 13 species. We also documented 93 mammal vehicle collisions comprising 12 species by monitoring the same road daily over a period of two years (Sept. 2004-Aug. 2006).The results of our analysis are as follows. First, the use rate of passages and the number of mammal vehicle collisions showed a positive correlation (r=0.890). Second, the fluctuation of the number of usable passages and collisions had no correlation (r=0.402). Third, the density of passages and collisions had a very weak positive correlation (r=0.559, pα=0.05). These results differed from following common expectations: higher numbers and use ratings of passages could cause less frequent collisions, high density areas of passage would cause fewer collisions, and the decreased number of passages would increase the use ratings of remaining passages. Fifth, most monitored mammal species with small-to-medium body sizes used all types of passage structures frequently, but water deer (Hydropotes inermis) rarely used these passage structures of under 0.7 on the openness index. Last, we found by radio telemetry that only one out of 13 radio-collared raccoon dogs was killed by vehicle `collision over a two-year period. However, a total of 12 raccoon dogs that had been killed by cars were found on the same road during the same period.The results of our research can be summarized as follows. First, there were already enough usable passages for wildlife, in spite of seasonal blockage of some passages or the uneven spacing between passages. Second, there were many occurrences of wildlife vehicle collisions, but settlers showed relatively low collision ratio. Third, most collision victims might be wanderers or newcomers unfamiliar to existing passages or occupying settlers. Finally, water deer should be the target species for the construction of wildlife passages, and the size should be O.I of over 0.7. Vehicle collision of other mammal species can be reduced significantly by installing wildlife fences without worsening habitat fragmentation in the case of roads that have many non-wildlife-engineered passages.

Iris is a cosmopolitan genus comprising approximately 280 species distributed throughout the Northern Hemisphere. Although Iris is the most diverse group in the Iridaceae, the number of taxa is debatable owing to various taxonomic issues. Plastid genomes have been widely used for phylogenetic research in plants; however, only limited number of plastid DNA markers are available for phylogenetic study of the Iris. To understand the genomic features of plastids within the genus, including its structural and genetic variation, we newly sequenced and analyzed the complete plastid genome of I. orchioides and compared it with those of 19 other Iris taxa. Potential plastid markers for phylogenetic research were identified by computing the sequence divergence and phylogenetic informativeness. We then tested the utility of the markers with the phylogenies inferred from the markers and whole-plastome data. The average size of the plastid genome was 152,926 bp, and the overall genomic content and organization were nearly identical among the 20 Iris taxa, except for minor variations in the inverted repeats. We identified 10 highly informative regions (matK, ndhF, rpoC2, ycf1, ycf2, rps15-ycf, rpoB-trnC, petA-psbJ, ndhG-ndhI and psbK-trnQ) and inferred a phylogeny from each region individually, as well as from their concatenated data. Remarkably, the phylogeny reconstructed from the concatenated data comprising three selected regions (rpoC2, ycf1 and ycf2) exhibited the highest congruence with the phylogeny derived from the entire plastome dataset. The result suggests that this subset of data could serve as a viable alternative to the complete plastome data, especially for molecular diagnoses among closely related Iris taxa, and at a lower cost. [ABSTRACT FROM AUTHOR]