An AUC-based permutation variable importance measure for random forests

Application Value of the Automated Machine Learning Model Based on Modified Computed Tomography Severity Index Combined With Serological Indicators in the Early Prediction of Severe Acute Pancreatitis

BACKGROUND AND AIMS

Machine learning (ML) algorithms are widely applied in building models of medicine due to their powerful studying and generalizing ability. To assess the value of the Modified Computed Tomography Severity Index (MCTSI) combined with serological indicators for early prediction of severe acute pancreatitis (SAP) by automated ML (AutoML).

PATIENTS AND METHODS

The clinical data, of the patients with acute pancreatitis (AP) hospitalized in Hospital 1 and hospital 2 from January 2017 to December 2021, were retrospectively analyzed. Serological indicators within 24 hours of admission were collected. MCTSI score was completed by noncontrast computed tomography within 24 hours of admission. Data from the hospital 1 were adopted for training, and data from the hospital 2 were adopted for external validation. The diagnosis of AP and SAP was based on the 2012 revised Atlanta classification of AP. Models were built using traditional logistic regression and AutoML analysis with 4 types of algorithms. The performance of models was evaluated by the receiver operating characteristic curve, the calibration curve, and the decision curve analysis based on logistic regression and decision curve analysis, feature importance, SHapley Additive exPlanation Plot, and Local Interpretable Model Agnostic Explanation based on AutoML.

RESULTS

A total of 499 patients were used to develop the models in the training data set. An independent data set of 201 patients was used to test the models. The model developed by the Deep Neural Net (DL) outperformed other models with an area under the receiver operating characteristic curve (areas under the curve) of 0.907 in the test set. Furthermore, among these AutoML models, the DL and gradient boosting machine models achieved the highest sensitivity values, both exceeding 0.800.

CONCLUSION

The AutoML model based on the MCTSI score combined with serological indicators has good predictive value for SAP in the early stage.

The impact of parental acceptance and childhood maltreatment on mental health and physical pain in Burundian survivors of childhood sexual abuse

BACKGROUND

Parental support has been suggested to mitigate mental and physical consequences following childhood sexual abuse (CSA). However, many CSA survivors experience parental rejection post-CSA.

OBJECTIVE

We aimed to understand the impact of abuse-specific parental acceptance on post-traumatic stress disorder (PTSD) and physical pain in Burundian CSA-survivors. We further assessed the significance of parental acceptance among known risk factors for predicting PTSD.

METHODS, PARTICIPANTS, AND SETTINGS

Participants (N = 131, 80.9 % female, mean age 16.21 years) were recruited via primary health care centers for survivors of sexual violence which survivors approached post-CSA. Survivors reported on PTSD symptoms, daytime/nighttime pain, and adverse childhood experiences in semi-structured interviews. Parental acceptance levels were categorized (acceptance, no acceptance, no contact) for mothers and fathers separately. Kruskal-Wallis tests assessed group differences. Conditional random forests (CRF) evaluated the significance of parental acceptance in predicting PTSD symptom severity.

RESULTS

No significant differences regarding PTSD symptoms and physical pain between levels of maternal acceptance were obtained. Pairwise comparisons revealed significant differences in PTSD symptom severity between paternal acceptance and no acceptance (d = 1.04) and paternal acceptance and no contact (d = 0.81). The CRF identified paternal acceptance as important variable for the prediction of PTSD symptom severity. Even though results were less conclusive, medium effect sizes hint at less pain perception within the paternal acceptance group.

CONCLUSIONS

The results highlight paternal acceptance as a potential risk or protective factor regarding psychological and possibly physical well-being in the aftermath of CSA, even in the context of other known risk factors.

Predicting carob tree physiological parameters under different irrigation systems using Random Forest and Planet satellite images

Introduction

In the context of climate change, monitoring the spatial and temporal variability of plant physiological parameters has become increasingly important. Remote spectral imaging and GIS software have shown effectiveness in mapping field variability. Additionally, the application of machine learning techniques, essential for processing large data volumes, has seen a significant rise in agricultural applications. This research was focused on carob tree, a drought-resistant tree crop spread through the Mediterranean basin. The study aimed to develop robust models to predict the net assimilation and stomatal conductance of carob trees and to use these models to analyze seasonal variability and the impact of different irrigation systems.

Methods

Planet satellite images were acquired on the day of field data measurement. The reflectance values of Planet spectral bands were used as predictors to develop the models. The study employed the Random Forest modeling approach, and its performances were compared with that of traditional multiple linear regression.

Results and discussion

The findings reveal that Random Forest, utilizing Planet spectral bands as predictors, achieved high accuracy in predicting net assimilation (R² = 0.81) and stomatal conductance (R² = 0.70), with the yellow and red spectral regions being particularly influential. Furthermore, the research indicates no significant difference in intrinsic water use efficiency between the various irrigation systems and rainfed conditions. This work highlighted the potential of combining satellite remote sensing and machine learning in precision agriculture, with the goal of the efficient monitoring of physiological parameters.

Looming-eyes buoys fail to reduce seabird bycatch in the Icelandic lumpfish fishery: depth-based fishing restrictions are an alternative

Bycatch in gillnets from lumpfish (Cyclopterus lumpus) fisheries is an important conservation issue in the north Atlantic, with up to 30 000 seabirds potentially killed each year. To date, no technical solutions exist to reduce the bycatch of seabirds in gillnet fisheries, but research on above-water bird deterrents as a form of bycatch mitigation has shown promising results. Here, we tested whether a floating device called 'looming-eyes buoy' (LEB) would consistently reduce the bycatch of seabirds in the Icelandic lumpfish fishery. We conducted 61 controlled trials with sets of normal gillnets and experimental nets equipped with LEBs. We compared both fish catch and bycatch between net types while accounting for exposure time, water depth and season, and found no effect of LEBs on both target lumpfish catch and bycatch. Our analysis indicated however a strong correlation between bycatch rates and fishing depths, suggesting that depth-based fishing restrictions could virtually eliminate the bycatch of seabirds in this fishery. We estimated that limiting fishing to waters more than 50 m deep could save between 5000 and 9300 seabirds every year, arrest the population decline of endangered black guillemots in Iceland, while having only a marginal effect on target fish catch.

Detection of diabetic patients in people with normal fasting glucose using machine learning

BACKGROUND

Diabetes mellitus (DM) is a chronic metabolic disease that could produce severe complications threatening life. Its early detection is thus quite important for the timely prevention and treatment. Normally, fasting blood glucose (FBG) by physical examination is used for large-scale screening of DM; however, some people with normal fasting glucose (NFG) actually have suffered from diabetes but are missed by the examination. This study aimed to investigate whether common physical examination indexes for diabetes can be used to identify the diabetes individuals from the populations with NFG.

METHODS

The physical examination data from over 60,000 individuals with NFG in three Chinese cohorts were used. The diabetes patients were defined by HbA1c ≥ 48 mmol/mol (6.5%). We constructed the models using multiple machine learning methods, including logistic regression, random forest, deep neural network, and support vector machine, and selected the optimal one on the validation set. A framework using permutation feature importance algorithm was devised to discover the personalized risk factors.

RESULTS

The prediction model constructed by logistic regression achieved the best performance with an AUC, sensitivity, and specificity of 0.899, 85.0%, and 81.1% on the validation set and 0.872, 77.9%, and 81.0% on the test set, respectively. Following feature selection, the final classifier only requiring 13 features, named as DRING (diabetes risk of individuals with normal fasting glucose), exhibited reliable performance on two newly recruited independent datasets, with the AUC of 0.964 and 0.899, the balanced accuracy of 84.2% and 81.1%, the sensitivity of 100% and 76.2%, and the specificity of 68.3% and 86.0%, respectively. The feature importance ranking analysis revealed that BMI, age, sex, absolute lymphocyte count, and mean corpuscular volume are important factors for the risk stratification of diabetes. With a case, the framework for identifying personalized risk factors revealed FBG, age, and BMI as significant hazard factors that contribute to an increased incidence of diabetes. DRING webserver is available for ease of application ( http://www.cuilab.cn/dring ).

CONCLUSIONS

DRING was demonstrated to perform well on identifying the diabetes individuals among populations with NFG, which could aid in early diagnosis and interventions for those individuals who are most likely missed.

Transcriptome-guided metabolic network analysis reveals rearrangements of carbon flux distribution in Neisseria gonorrhoeae during neutrophil co-culture

The ability of bacterial pathogens to metabolically adapt to the environmental conditions of their hosts is critical to both colonization and invasive disease. Infection with Neisseria gonorrhoeae (the gonococcus, Gc) is characterized by the influx of neutrophils [polymorphonuclear leukocytes (PMNs)], which fail to clear the bacteria and make antimicrobial products that can exacerbate tissue damage. The inability of the human host to clear Gc infection is particularly concerning in light of the emergence of strains that are resistant to all clinically recommended antibiotics. Bacterial metabolism represents a promising target for the development of new therapeutics against Gc. Here, we generated a curated genome-scale metabolic network reconstruction (GENRE) of Gc strain FA1090. This GENRE links genetic information to metabolic phenotypes and predicts Gc biomass synthesis and energy consumption. We validated this model with published data and in new results reported here. Contextualization of this model using the transcriptional profile of Gc exposed to PMNs revealed substantial rearrangements of Gc central metabolism and induction of Gc nutrient acquisition strategies for alternate carbon source use. These features enhanced the growth of Gc in the presence of neutrophils. From these results, we conclude that the metabolic interplay between Gc and PMNs helps define infection outcomes. The use of transcriptional profiling and metabolic modeling to reveal new mechanisms by which Gc persists in the presence of PMNs uncovers unique aspects of metabolism in this fastidious bacterium, which could be targeted to block infection and thereby reduce the burden of gonorrhea in the human population. IMPORTANCE The World Health Organization designated Gc as a high-priority pathogen for research and development of new antimicrobials. Bacterial metabolism is a promising target for new antimicrobials, as metabolic enzymes are widely conserved among bacterial strains and are critical for nutrient acquisition and survival within the human host. Here we used genome-scale metabolic modeling to characterize the core metabolic pathways of this fastidious bacterium and to uncover the pathways used by Gc during culture with primary human immune cells. These analyses revealed that Gc relies on different metabolic pathways during co-culture with human neutrophils than in rich media. Conditionally essential genes emerging from these analyses were validated experimentally. These results show that metabolic adaptation in the context of innate immunity is important to Gc pathogenesis. Identifying the metabolic pathways used by Gc during infection can highlight new therapeutic targets for drug-resistant gonorrhea.

The Canadian Optimized Statistical Smoke Exposure Model (CanOSSEM): A machine learning approach to estimate national daily fine particulate matter (PM2.5) exposure

Exposure to biomass smoke has been associated with a wide range of acute and chronic health outcomes. Over the past decades, the frequency and intensity of wildfires has increased in many areas, resulting in longer smoke episodes with higher concentrations of fine particulate matter (PM_2.5). There are also many communities where seasonal open burning and residential wood heating have short- and long-term impacts on ambient air quality. Understanding the acute and chronic health effects of biomass smoke exposure requires reliable estimates of PM_2.5 concentrations during the wildfire season and throughout the year, particularly in areas without regulatory air quality monitoring stations. We have developed a machine learning approach to estimate PM_2.5 across all populated regions of Canada from 2010 to 2019. The random forest machine learning model uses potential predictor variables integrated from multiple data sources and estimates daily mean (24-hour) PM_2.5 concentrations at a 5 km × 5 km spatial resolution. The training and prediction datasets were generated using observations from National Air Pollution Surveillance (NAPS) network. The Root Mean Squared Error (RMSE) between predicted and observed PM_2.5 concentrations was 2.96 μg/m³ for the entire prediction set, and more than 96 % of the predictions were within 5 μg/m³ of the NAPS PM_2.5 measurements. The model was evaluated using 10-fold, leave one-region-out, and leave-one-year-out cross-validations. Overall, CanOSSEM performed well but performance was sensitive to removal of large wildfire events such as the Fort McMurray interface fire in May 2016 or the extreme 2017 and 2018 wildfire seasons in British Columbia. Exposure estimates from CanOSSEM will be useful for epidemiologic studies on the acute and chronic health effects associated with PM_2.5 exposure, especially for populations affected by biomass smoke where routine air quality measurements are not available.

Identifying Evacuation Needs and Resources Based on Volunteered Geographic Information: A Case of the Rainstorm in July 2021, Zhengzhou, China

Recently, global climate change has led to a high incidence of extreme weather and natural disasters. How to reduce its impact has become an important topic. However, the studies that both consider the disaster's real-time geographic information and environmental factors in severe rainstorms are still not enough. Volunteered geographic information (VGI) data that was generated during disasters offered possibilities for improving the emergency management abilities of decision-makers and the disaster self-rescue abilities of citizens. Through the case study of the extreme rainstorm disaster in Zhengzhou, China, in July 2021, this paper used machine learning to study VGI issued by residents. The vulnerable people and their demands were identified based on the SOS messages. The importance of various indicators was analyzed by combining open data from socio-economic and built-up environment elements. Potential safe areas with shelter resources in five administrative districts in the disaster-prone central area of Zhengzhou were identified based on these data. This study found that VGI can be a reliable data source for future disaster research. The characteristics of rainstorm hazards were concluded from the perspective of affected people and environmental indicators. The policy recommendations for disaster prevention in the context of public participation were also proposed.

Quantitative Structure-Property Relationship Analysis for the Prediction of Propylene Adsorption Capacity in Pure Silicon Zeolites at Various Pressure Levels

This work is devoted to the development of quantitative structure-property relationship (QSPR) models using various regression analyses to predict propylene (C₃H₆) adsorption capacity at various pressures in zeolites from a topologically diverse International Zeolite Association database. Based on univariate and multilinear regression analysis, the accessible volume and largest cavity diameter are the most crucial factors determining C₃H₆ uptake at high and low pressures, respectively. An artificial neural network (ANN) model with five structural descriptors is sufficient to predict C₃H₆ uptake at high pressures. For combined pressures, the prediction of an ANN model with pore size distribution is pleasing. The isosteric heat of adsorption (Q _st) has a significant impact on the improvement of the prediction of low-pressure gas adsorption, which finely classifies zeolites into high or low C₃H₆ adsorbers. The conjunction of high-throughput screening and QSPR models contributes to being able to prescreen the database rapidly and accurately for top performers and perform further detailed and time-consuming computational-intensive molecular simulations on these candidates for other gas adsorption applications.

Using machine learning to predict individual patient toxicities from cancer treatments

PURPOSE

Machine learning (ML) is a powerful tool for interrogating datasets and learning relationships between multiple variables. We utilized a ML model to identify those early breast cancer (EBC) patients at highest risk of developing severe vasomotor symptoms (VMS).

METHODS

A gradient boosted decision model utilizing cross-sectional survey data from 360 EBC patients was created. Seventeen patient- and treatment-specific variables were considered in the model. The outcome variable was based on the Hot Flush Night Sweats (HFNS) Problem Rating Score, and individual scores were dichotomized around the median to indicate individuals with high and low problem scores. Model accuracy was assessed using the area under the receiver operating curve, and conditional partial dependence plots were constructed to illustrate relationships between variables and the outcome of interest.

RESULTS

The model area under the ROC curve was 0.731 (SD 0.074). The most important variables in the model were as follows: the number of hot flashes per week, age, the prescription, or use of drug interventions to manage VMS, whether patients were asked about VMS in routine follow-up visits, and the presence or absence of changes to breast cancer treatments due to VMS. A threshold of 17 hot flashes per week was identified as being more predictive of severe VMS. Patients between the ages of 49 and 63 were more likely to report severe symptoms.

CONCLUSION

Machine learning is a unique tool for predicting severe VMS. The use of ML to assess other treatment-related toxicities and their management requires further study.

Simulation of the potential distribution of rare and endangered Satyrium species in China under climate change

Satyrium is an endangered and rare genus of plant that has various pharmacodynamic functions. In this study, optimized MaxEnt models were used in analyzing potential geographical distributions under current and future climatic conditions (the 2050s and 2070s) and dominant environmental variables influencing their geographic distribution. The results provided reference for implementation of long‐term conservation and management approaches for the species. The results showed that the area of the total suitable habitat for Satyrium ciliatum (S. ciliatum) in China is 32.51 × 10⁴ km², the total suitable habitat area for Satyrium nepalense (S. nepalense) in China is 61.76 × 10⁴ km², and the area of the total suitable habitat for Satyrium yunnanense (S. yunnanense) in China is 89.73 × 10⁴ km² under current climatic conditions. The potential suitable habitat of Satyrium is mainly distributed in Southwest China. The major environmental variables influencing the geographical distribution of S. ciliatum were isothermality (bio3), temperature seasonality (bio4), and mean temperature of coldest quarter (bio11). Environmental variables such as isothermality (bio3), temperature seasonality (bio4), and precipitation of coldest quarter (bio19) affected the geographical distribution of S. nepalense; and environmental variables such as isothermality (bio3), temperature seasonality (bio4), and lower temperature of coldest month (bio6) affected the geographical distribution of S. yunnanense. The distribution range of Satyrium was extended as global warming increased, showing emissions of greenhouse gases with lower concentration (SSP1‐2.6) and higher concentration (SSP5‐8.5). According to the study, the distribution of suitable habitat will shift with a change to higher elevation areas and higher latitude areas in the future.

Automated Machine Learning for the Early Prediction of the Severity of Acute Pancreatitis in Hospitals

Background

Machine learning (ML) algorithms are widely applied in building models of medicine due to their powerful studying and generalizing ability. This study aims to explore different ML models for early identification of severe acute pancreatitis (SAP) among patients hospitalized for acute pancreatitis.

Methods

This retrospective study enrolled patients with acute pancreatitis (AP) from multiple centers. Data from the First Affiliated Hospital and Changshu No. 1 Hospital of Soochow University were adopted for training and internal validation, and data from the Second Affiliated Hospital of Soochow University were adopted for external validation from January 2017 to December 2021. The diagnosis of AP and SAP was based on the 2012 revised Atlanta classification of acute pancreatitis. Models were built using traditional logistic regression (LR) and automated machine learning (AutoML) analysis with five types of algorithms. The performance of models was evaluated by the receiver operating characteristic (ROC) curve, the calibration curve, and the decision curve analysis (DCA) based on LR and feature importance, SHapley Additive exPlanation (SHAP) Plot, and Local Interpretable Model Agnostic Explanation (LIME) based on AutoML.

Results

A total of 1,012 patients were included in this study to develop the AutoML models in the training/validation dataset. An independent dataset of 212 patients was used to test the models. The model developed by the gradient boost machine (GBM) outperformed other models with an area under the ROC curve (AUC) of 0.937 in the validation set and an AUC of 0.945 in the test set. Furthermore, the GBM model achieved the highest sensitivity value of 0.583 among these AutoML models. The model developed by eXtreme Gradient Boosting (XGBoost) achieved the highest specificity value of 0.980 and the highest accuracy of 0.958 in the test set.

Conclusions

The AutoML model based on the GBM algorithm for early prediction of SAP showed evident clinical practicability.

Machine Learning-Based Analysis and Prediction of Unplanned 30-Day Readmissions After Pituitary Adenoma Resection: A Multi-Institutional Retrospective Study With External Validation

BACKGROUND

Unplanned readmission after transsphenoidal resection of pituitary adenoma can occur in up to 10% of patients but is unpredictable.

OBJECTIVE

To develop a reliable system for predicting unplanned readmission and create a validated method for stratifying patients by risk.

METHODS

Data sets were retrospectively collected from the National Surgical Quality Improvement Program and 2 tertiary academic medical centers. Eight machine learning classifiers were fit to the National Surgical Quality Improvement Program data, optimized using Bayesian parameter optimization and evaluated on the external data. Permutation analysis identified the relative importance of predictive variables, and a risk stratification system was built using the trained machine learning models.

RESULTS

Readmissions were accurately predicted by several classification models with an area under the receiving operator characteristic curve of 0.76 (95% CI 0.68-0.83) on the external data set. Permutation analysis identified the most important variables for predicting readmission as preoperative sodium level, returning to the operating room, and total operation time. High-risk and medium-risk patients, as identified by the proposed risk stratification system, were more likely to be readmitted than low-risk patients, with relative risks of 12.2 (95% CI 5.9-26.5) and 4.2 (95% CI 2.3-8.7), respectively. Overall risk stratification showed high discriminative capability with a C-statistic of 0.73.

CONCLUSION

In this multi-institutional study with outside validation, unplanned readmissions after pituitary adenoma resection were accurately predicted using machine learning techniques. The features identified in this study and the risk stratification system developed could guide clinical and surgical decision making, reduce healthcare costs, and improve the quality of patient care by better identifying high-risk patients for closer perioperative management.

Assessing the geographical distribution of 76 Dendrobium species and impacts of climate change on their potential suitable distribution area in China

The geographical distribution of plant resources is of great significance for studying the origin, distribution, and evolution of species. Climate and geographical factors help shape the distribution of plant species. Dendrobium is a commonly used traditional medicine and a precious economic crop in China. Owing to the over-exploitation and increasing medicinal demand of Dendrobium species plants, systematic investigation of the geographical distribution of the plants and analysis of their potential distribution under climate change are important for protecting Dendrobium plants. We adopted DIVA-GIS to analyze the georeferenced records of 76 species of the Dendrobium species collected from 2166 herbarium records. We analyzed the eco-geographical distribution and species richness of the genus Dendrobium to simulate the distribution of current and future scenarios using MaxEnt. The results revealed the distribution of Dendrobium in 30 provinces of China, with species abundance in Yunnan, Guangxi, Guangdong, and Hainan. Our model identified the following bioclimatic variables: precipitation in the driest months and the warmest seasons, isothermality, and range of annual temperature. Among them, annual precipitation is the most crucial bioclimatic variable affecting the distribution of 16 selected Dendrobium species. The change of climate in the future will lead to an increase in habitat suitability for some Dendrobium species as follows: D. officinal 2.12%, D. hancockii by 6.00%, D. hercoglossum by 8.25%, D. devonianum by 7.71%, D. henryi by 9.40%, and D. hainanense by 13.70%. By contrast, habitat suitability will dramatically decrease for other Dendrobium species: D. chrysotoxum by 0.89%, D. chrysanthum by 12.68%, D. fimbriatum by 5.07%, D. aduncum by 11.44%, D. densiflorum by 18.47%, D. aphyllum by 8.05%, D. loddigesii by 16.45%, D. nobile by 5.41%, D. falconeri by 8.73%, and D. moniliforme by 10.61%. The reduction of these species will be detrimental to the medicinal and economic value of the genus Dendrobium. Therefore, targeted development and reasonable management strategies should be adopted to conserve these valuable resources.

Covid19-Mexican-Patients' Dataset (Covid19MPD) Classification and Prediction Using Feature Importance

Coronavirus disease, Covid19, pandemic has a great effect on human heath worldwide since it was first detected in late 2019. A clear understanding of the structure of the available Covid19 datasets might give the healthcare provider a better understanding of identifying some of the cases at an early stage. In this article, we will be looking into a Covid19 Mexican Patients' Dataset (Covid109MPD), and we will apply number of machine learning algorithms on the dataset to select the best possible classification algorithm for the death and survived cases in Mexico, then we will study the performance of the enhancement of the specified classifiers in term of their features selection in order to be able to predict sever, and or death, cases from the available dataset. Results show that J48 classifier gives the best classification accuracy with 94.41% and RMSE = 0.2028 and ROC = 0.919, compared to other classifiers, and when using feature selection method, J48 classifier can predict a surviving Covid19MPD case within 94.88% accuracy, and by using only 10 out of the total 19 features.

Industrial data science – a review of machine learning applications for chemical and process industries

Understand and optimize industrial processes via machine learning and chemical engineering principles.

Changes of public risk perception in China: 2008-2018

This paper characterizes the risk acceptance of the Chinese public based on a psychometric paradigm and documents its change by conducting a nationally representative longitudinal survey spanning 10 years. We explore key factors that influence the acceptance of seven typical risks: drinking water pollution, interior decoration, electromagnetic radiation, air pollution, chemical plants, public transportation, and natural hazards, reflecting the general and referential changes in risk perception. The results show a general decrease in the acceptance of all of these risks in the examined decade, especially in economically developed areas. Different types of risk perception varied, but environmental risks had similar trends of perception. The perceived benefits from these risks and local GDP had the greatest impact on risk acceptance. The interaction between the changing perspectives of the emerging middle class and the evolving hazard risk landscape may be the reasons for the reduction in risk acceptance. The main findings offer insights for effective risk education and communication as well as sustainable risk management strategy.

Changes in the International Classification of Functioning, Disability, and Health Components "Activity/Participation" as Predicted Through Patient-Reported Outcomes Along With Comprehensive Back Pain Rehabilitation

OBJECTIVES

The World Health Organization (WHO) recommended the International Classification of Functioning, Disability and Health (ICF) but its use in clinical practice is sparse. This study investigated the limitations and restrictions in the most relevant brief ICF core set categories for chronic low back pain (cLBP) as automatically predicted from routinely measured outcomes using a novel, validated mapping algorithm.

MATERIALS AND METHODS

Of 2718 cLBP patients recruited, data from 1541 (64% females) were available from before and at the end of 6 months comprehensive outpatient rehabilitation. Assessments included the Roland Morris Disability Questionnaire (RMDQ) and Pain Disability Index (PDI) questionnaires, the percentage of patients with predicted limitations and restrictions in important activity and participation ICF categories, bodily functional measurements, pain intensity, and anxiety/depression (EQ-5D).

RESULTS

At baseline, both the RMDQ and the PDI measures were within the third of the lowest disability scores whilst 80% of the patients had limitations with "maintaining a body position" and 30% with "walking" ICF categories. Intervention-associated gains in the maximum isometric lumbar extension and flexion strength and the lumbar range of motion were significant overall, but improvements in patients' ICF limitations/restrictions varied. Anxiety/depression, lumbar range of motion, and extension strength all had a significant impact on the majority of the ICF categories, whereas flexion strength had none.

DISCUSSION

The rate of patients with predicted limitations/restrictions in activity/participation ICF core categories for cLBP partly mirrored disability levels and the impact of the body function scores on these limitations/restrictions in ICF categories was varied. Thus, assessing problems in the ICF activity/participation core categories is of relevance to clinical practice for both treatment goal setting and intervention planning. This may be achieved by computer-generated mapping without additional time burden.

Novel Drivers of Virulence in Clostridioides difficile Identified via Context-Specific Metabolic Network Analysis

The pathogen Clostridioides difficile causes toxin-mediated diarrhea and is the leading cause of hospital-acquired infection in the United States. Due to growing antibiotic resistance and recurrent infection, targeting C. difficile metabolism presents a new approach to combat this infection. Genome-scale metabolic network reconstructions (GENREs) have been used to identify therapeutic targets and uncover properties that determine cellular behaviors. Thus, we constructed C. difficile GENREs for a hypervirulent isolate (strain [str.] R20291) and a historic strain (str. 630), validating both with in vitro and in vivo data sets. Growth simulations revealed significant correlations with measured carbon source usage (positive predictive value [PPV] ≥ 92.7%), and single-gene deletion analysis showed >89.0% accuracy. Next, we utilized each GENRE to identify metabolic drivers of both sporulation and biofilm formation. Through contextualization of each model using transcriptomes generated from in vitro and infection conditions, we discovered reliance on the pentose phosphate pathway as well as increased usage of cytidine and N-acetylneuraminate when virulence expression is reduced, which was subsequently supported experimentally. Our results highlight the ability of GENREs to identify novel metabolite signals in higher-order phenotypes like bacterial pathogenesis. IMPORTANCE Clostridioides difficile has become the leading single cause of hospital-acquired infections. Numerous studies have demonstrated the importance of specific metabolic pathways in aspects of C. difficile pathophysiology, from initial colonization to regulation of virulence factors. In the past, genome-scale metabolic network reconstruction (GENRE) analysis of bacteria has enabled systematic investigation of the genetic and metabolic properties that contribute to downstream virulence phenotypes. With this in mind, we generated and extensively curated C. difficile GENREs for both a well-studied laboratory strain (str. 630) and a more recently characterized hypervirulent isolate (str. R20291). In silico validation of both GENREs revealed high degrees of agreement with experimental gene essentiality and carbon source utilization data sets. Subsequent exploration of context-specific metabolism during both in vitro growth and infection revealed consistent patterns of metabolism which corresponded with experimentally measured increases in virulence factor expression. Our results support that differential C. difficile virulence is associated with distinct metabolic programs related to use of carbon sources and provide a platform for identification of novel therapeutic targets.

Nkx3-1 and Fech genes might be switch genes involved in pituitary non-functioning adenoma invasiveness

Non-functioning pituitary adenomas (NFPAs) are typical pituitary macroadenomas in adults associated with increased mortality and morbidity. Although pituitary adenomas are commonly considered slow-growing benign brain tumors, numerous of them possess an invasive nature. Such tumors destroy sella turcica and invade the adjacent tissues such as the cavernous sinus and sphenoid sinus. In these cases, the most critical obstacle for complete surgical removal is the high risk of damaging adjacent vital structures. Therefore, the development of novel therapeutic strategies for either early diagnosis through biomarkers or medical therapies to reduce the recurrence rate of NFPAs is imperative. Identification of gene interactions has paved the way for decoding complex molecular mechanisms, including disease-related pathways, and identifying the most momentous genes involved in a specific disease. Currently, our knowledge of the invasion of the pituitary adenoma at the molecular level is not sufficient. The current study aimed to identify critical biomarkers and biological pathways associated with invasiveness in the NFPAs using a three-way interaction model for the first time. In the current study, the Liquid association method was applied to capture the statistically significant triplets involved in NFPAs invasiveness. Subsequently, Random Forest analysis was applied to select the most important switch genes. Finally, gene set enrichment (GSE) and gene regulatory network (GRN) analyses were applied to trace the biological relevance of the statistically significant triplets. The results of this study suggest that "mRNA processing" and "spindle organization" biological processes are important in NFAPs invasiveness. Specifically, our results suggest Nkx3-1 and Fech as two switch genes in NFAPs invasiveness that may be potential biomarkers or target genes in this pathology.

Toward insights on antimicrobial selectivity of host defense peptides via machine learning model interpretation

Host defense peptides are promising candidates for the development of novel antibiotics. To realize their therapeutic potential, high levels of target selectivity is essential. This study aims to identify factors governing selectivity via the use of the random forest algorithm for correlating peptide sequence information with their bioactivity data. Satisfactory predictive models were achieved from out-of-bag prediction that yielded accuracies and Matthew's correlation coefficients in excess of 0.80 and 0.57, respectively. Model interpretation through the use of variable importance metrics and partial dependence plots indicated that the selectivity was heavily influenced by the composition and distribution patterns of molecular charge and solubility related parameters. Furthermore, the three investigated bacterial target species (Escherichia coli, Pseudomonas aeruginosa and Staphylococcus aureus) likely had a significant influence on how selectivity was realized as there appears to be a similar underlying selectivity mechanism on the basis of charge-solubility properties (i.e. but which is tailored according to the target in question).

Common irrigation drivers of freshwater salinisation in river basins worldwide

Freshwater salinisation is a growing problem, yet cross-regional assessments of freshwater salinity status and the impact of agricultural and other sectoral uses are lacking. Here, we assess inland freshwater salinity patterns and evaluate its interactions with irrigation water use, across seven regional river basins (401 river sub-basins) around the world, using long-term (1980-2010) salinity observations. While a limited number of sub-basins show persistent salinity problems, many sub-basins temporarily exceeded safe irrigation water-use thresholds and 57% experience increasing salinisation trends. We further investigate the role of agricultural activities as drivers of salinisation and find common contributions of irrigation-specific activities (irrigation water withdrawals, return flows and irrigated area) in sub-basins of high salinity levels and increasing salinisation trends, compared to regions without salinity issues. Our results stress the need for considering these irrigation-specific drivers when developing management strategies and as a key human component in water quality modelling and assessment.

Earthquake-Induced Building-Damage Mapping Using Explainable AI (XAI)

Building-damage mapping using remote sensing images plays a critical role in providing quick and accurate information for the first responders after major earthquakes. In recent years, there has been an increasing interest in generating post-earthquake building-damage maps automatically using different artificial intelligence (AI)-based frameworks. These frameworks in this domain are promising, yet not reliable for several reasons, including but not limited to the site-specific design of the methods, the lack of transparency in the AI-model, the lack of quality in the labelled image, and the use of irrelevant descriptor features in building the AI-model. Using explainable AI (XAI) can lead us to gain insight into identifying these limitations and therefore, to modify the training dataset and the model accordingly. This paper proposes the use of SHAP (Shapley additive explanation) to interpret the outputs of a multilayer perceptron (MLP)-a machine learning model-and analyse the impact of each feature descriptor included in the model for building-damage assessment to examine the reliability of the model. In this study, a post-event satellite image from the 2018 Palu earthquake was used. The results show that MLP can classify the collapsed and non-collapsed buildings with an overall accuracy of 84% after removing the redundant features. Further, spectral features are found to be more important than texture features in distinguishing the collapsed and non-collapsed buildings. Finally, we argue that constructing an explainable model would help to understand the model's decision to classify the buildings as collapsed and non-collapsed and open avenues to build a transferable AI model.

Models that combine transcriptomic with spatial protein information exceed the predictive value for either single modality

Immunotherapy has reshaped the field of cancer therapeutics but the population that benefits are small in many tumor types, warranting a companion diagnostic test. While immunohistochemistry (IHC) for programmed death-ligand 1 (PD-L1) or mismatch repair (MMR) and polymerase chain reaction (PCR) for microsatellite instability (MSI) are the only approved companion diagnostics others are under consideration. An optimal companion diagnostic test might combine the spatial information of IHC with the quantitative information from RNA expression profiling. Here, we show proof of concept for combination of spatially resolved protein information acquired by the NanoString GeoMx® Digital Spatial Profiler (DSP) with transcriptomic information from bulk mRNA gene expression acquired using NanoString nCounter® PanCancer IO 360™ panel on the same cohort of immunotherapy treated melanoma patients to create predictive models associated with clinical outcomes. We show that the combination of mRNA and spatially defined protein information can predict clinical outcomes more accurately (AUC 0.97) than either of these factors alone.

Buoys with looming eyes deter seaducks and could potentially reduce seabird bycatch in gillnets

Bycatch of seabirds in gillnet fisheries is a global conservation issue with an estimated 400 000 seabirds killed each year. To date, no underwater deterrents trialled have consistently reduced seabird bycatch across operational fisheries. Using a combination of insights from land-based strategies, seabirds' diving behaviours and their cognitive abilities, we developed a floating device exploring the effect of large eyespots and looming movement to prevent vulnerable seabirds from diving into gillnets. Here, we tested whether this novel above-water device called 'Looming eyes buoy' (LEB) would consistently deter vulnerable seaducks from a focal area. We counted the number of birds present in areas with and without LEBs in a controlled experimental setting. We show that long-tailed duck Clangula hyemalis abundance declined by approximately 20-30% within a 50 m radius of the LEB and that the presence of LEBs was the most important variable explaining this decline. We found no evidence for a memory effect on long-tailed ducks but found some habituation to the LEB within the time frame of the project (62 days). While further research is needed, our preliminary trials indicate that above-water visual devices could potentially contribute to reduce seabird bycatch if appropriately deployed in coordination with other management measures.

Complement C3 identified as a unique risk factor for disease severity among young COVID-19 patients in Wuhan, China

Given that a substantial proportion of the subgroup of COVID-19 patients that face a severe disease course are younger than 60 years, it is critical to understand the disease-specific characteristics of young COVID-19 patients. Risk factors for a severe disease course for young COVID-19 patients and possible non-linear influences remain unknown. Data were analyzed from COVID-19 patients with clinical outcome in a single hospital in Wuhan, China, collected retrospectively from Jan 24th to Mar 27th. Clinical, demographic, treatment and laboratory data were collected from patients' medical records. Uni- and multivariable analysis using logistic regression and random forest, with the latter allowing the study of non-linear influences, were performed to investigate the clinical characteristics of a severe disease course. A total of 762 young patients (median age 47 years, interquartile range [IQR] 38–55, range 18–60; 55.9% female) were included, as well as 714 elderly patients as a comparison group. Among the young patients, 362 (47.5%) had a severe/critical disease course and the mean age was statistically significantly higher in the severe subgroup than in the mild subgroup (59.3 vs. 56.0, Student's t-test: p < 0.001). The uni- and multivariable analysis suggested that several covariates such as elevated levels of serum amyloid A (SAA), C-reactive protein (CRP) and lactate dehydrogenase (LDH), and decreased lymphocyte counts influence disease severity independently of age. Elevated levels of complement C3 (odds ratio [OR] 15.6, 95% CI 2.41–122.3; p = 0.039) are particularly associated with the risk of developing severe COVID-19 specifically in young patients, whereas no such influence seems to exist for elderly patients. Additional analysis suggests that the influence of complement C3 in young patients is independent of age, gender, and comorbidities. Variable importance values and partial dependence plots obtained using random forests delivered additional insights, in particular indicating non-linear influences of risk factors on disease severity. This study identified increased levels of complement C3 as a unique risk factor for adverse outcomes specific to young COVID-19 patients.

Machine Learning-Derived Fractal Features of Shape and Texture of the Left Atrium and Pulmonary Veins From Cardiac Computed Tomography Scans Are Associated With Risk of Recurrence of Atrial Fibrillation Postablation

[Figure: see text].

Taxonomic revision of the genus Xenopholis Peters, 1869 (Serpentes: Dipsadidae): Integrating morphology with ecological niche

A reliable identification and delimitation of species is an essential pre-requisite for many fields of science and conservation. The Neotropical herpetofauna is the world's most diverse, including many taxa of uncertain or debated taxonomy. Here we tackle one such species complex, by evaluating the taxonomic status of species currently allocated in the snake genus Xenopholis (X. scalaris, X. undulatus, and X. werdingorum). We base our conclusions on concordance between quantitative (meristic and morphometric) and qualitative (color pattern, hemipenes and skull features) analyses of morphological characters, in combination with ecological niche modeling. We recognize all three taxa as valid species and improve their respective diagnosis, including new data on color in life, pholidosis, bony morphology, and male genitalia. We find low overlap among the niches of each species, corroborating the independent source of phenotypic evidence. Even though all three species occur in the leaf litter of distinct forested habitats, Xenopholis undulatus is found in the elevated areas of the Brazilian Shield (Caatinga, Cerrado and Chaco), whereas X. scalaris occurs in the Amazon and Atlantic rainforests, and X. werdingorum in the Chiquitanos forest and Pantanal wetlands. We discuss the disjunct distribution between Amazonian and Atlantic Forest snake species in the light of available natural history and ecological aspects. This study shows the advantages of combining multiple data sources for reliable identification and circumscription of ecologically similar species.

Principles of mRNA control by human PUM proteins elucidated from multimodal experiments and integrative data analysis

The human PUF-family proteins, PUM1 and PUM2, posttranscriptionally regulate gene expression by binding to a PUM recognition element (PRE) in the 3'-UTR of target mRNAs. Hundreds of PUM1/2 targets have been identified from changes in steady-state RNA levels; however, prior studies could not differentiate between the contributions of changes in transcription and RNA decay rates. We applied metabolic labeling to measure changes in RNA turnover in response to depletion of PUM1/2, showing that human PUM proteins regulate expression almost exclusively by changing RNA stability. We also applied an in vitro selection workflow to precisely identify the binding preferences of PUM1 and PUM2. By integrating our results with prior knowledge, we developed a "rulebook" of key contextual features that differentiate functional versus nonfunctional PREs, allowing us to train machine learning models that accurately predict the functional regulation of RNA targets by the human PUM proteins.

Predicting the Potential Distribution of Two Varieties of Litsea coreana (Leopard-Skin Camphor) in China under Climate Change

Climate change considerably affects vegetation growth and may lead to changes in vegetation distribution. Leopard-skin camphor is an endangered species, and the main raw material for hawk tea, and has various pharmacodynamic functions. Studying the potential distribution of two leopard-skin camphor varieties under climate change should assist in the effective protection of these species. We collected the distribution point data for 130 and 89 Litsea coreana Levl. var. sinensis and L. coreana Levl. var. lanuginosa, respectively, and data for 22 environmental variables. We also predicted the potential distribution of the two varieties in China using the maximum entropy (MaxEnt) model and analyzed the key environmental factors affecting their distribution. Results showed that the two varieties are mainly located in the subtropical area south of the Qinling Mountains–Huai River line in the current and future climate scenarios, and the potentially suitable area for L. coreana Levl. var. lanuginosa is larger than that of L. coreana Levl. var. sinensis. Compared with current climatic conditions, the potentially suitable areas of the two leopard-skin camphor varieties will move to high-latitude and -altitude areas and the total suitable area will increase slightly, while moderately and highly suitable areas will be significantly reduced under future climatic scenarios. For example, under a 2070-RCP8.5 (representative of a high greenhouse gas emission scenario in the 2070s) climatic scenario, the highly suitable areas of L. coreana Levl. var. sinensis and L. coreana Levl. var. lanuginosa are 6900 and 300 km2, and account for only 10.27% and 0.21% of the current area, respectively. Temperature is the key environmental factor affecting the potential distribution of the two varieties, especially the mean daily diurnal range (Bio2) and the min temperature of the coldest month (Bio6). The results can provide a reference for relevant departments in taking protective measures to prevent the decrease or extinction of the species under climate change.

Identification of primary effecters of N2O emissions from full-scale biological nitrogen removal systems using random forest approach

Wastewater treatment plants (WWTPs) have long been recognized as point sources of N₂O, a potent greenhouse gas and ozone-depleting agent. Multiple mechanisms, both biotic and abiotic, have been suggested to be responsible for N₂O production from WWTPs, with basis on extrapolation from laboratory results and statistical analyses of metadata collected from operational full-scale plants. In this study, random forest (RF) analysis, a machine-learning approach for feature selection from highly multivariate datasets, was adopted to investigate N₂O production mechanism in activated sludge tanks of WWTPs from a novel perspective. Standardized measurements of N₂O effluxes coupled with exhaustive metadata collection were performed at activated sludge tanks of three biological nitrogen removal WWTPs at different times of the year. The multivariate datasets were used as inputs for RF analyses. Computation of the permutation variable importance measures returned biomass-normalized dissolved inorganic carbon concentration (DIC·VSS^-1) and specific ammonia oxidation activity (sOUR_AOB) as the most influential parameters determining N₂O emissions from the aerated zones (or phases) of activated sludge bioreactors. For the anoxic tanks, dissolved-organic-carbon-to-NO₂^-/NO₃^- ratio (DOC·(NO₂^--N + NO₃^--N)^-1) was singled out as the most influential. These data analysis results clearly indicate disparate mechanisms for N₂O generation in the oxic and anoxic activated sludge bioreactors, and provide evidences against significant contributions of N₂O carryover across different zones or phases or niche-specific microbial reactions, with aerobic NH₃/NH₄⁺ oxidation to NO₂^- and anoxic denitrification predominantly responsible from aerated and anoxic zones or phases of activated sludge bioreactors, respectively.

Machine-learning-based detection of adaptive divergence of the stream mayfly Ephemera strigata populations

Adaptive divergence is a key mechanism shaping the genetic variation of natural populations. A central question linking ecology with evolutionary biology is how spatial environmental heterogeneity can lead to adaptive divergence among local populations within a species. In this study, using a genome scan approach to detect candidate loci under selection, we examined adaptive divergence of the stream mayfly Ephemera strigata in the Natori River Basin in northeastern Japan. We applied a new machine-learning method (i.e., random forest) besides traditional distance-based redundancy analysis (dbRDA) to examine relationships between environmental factors and adaptive divergence at non-neutral loci. Spatial autocorrelation analysis based on neutral loci was employed to examine the dispersal ability of this species. We conclude the following: (a) E. strigata show altitudinal adaptive divergence among the populations in the Natori River Basin; (b) random forest showed higher resolution for detecting adaptive divergence than traditional statistical analysis; and (c) separating all markers into neutral and non-neutral loci could provide full insight into parameters such as genetic diversity, local adaptation, and dispersal ability.

Bacterial communities associated with cell phones and shoes

Background

Every human being carries with them a collection of microbes, a collection that is likely both unique to that person, but also dynamic as a result of significant flux with the surrounding environment. The interaction of the human microbiome (i.e., the microbes that are found directly in contact with a person in places such as the gut, mouth, and skin) and the microbiome of accessory objects (e.g., shoes, clothing, phones, jewelry) is of potential interest to both epidemiology and the developing field of microbial forensics. Therefore, the microbiome of personal accessories are of interest because they serve as both a microbial source and sink for an individual, they may provide information about the microbial exposure experienced by an individual, and they can be sampled non-invasively.

Findings

We report here a large-scale study of the microbiome found on cell phones and shoes. Cell phones serve as a potential source and sink for skin and oral microbiome, while shoes can act as sampling devices for microbial environmental experience. Using 16S rRNA gene sequencing, we characterized the microbiome of thousands of paired sets of cell phones and shoes from individuals at sporting events, museums, and other venues around the United States.

Conclusions

We place this data in the context of previous studies and demonstrate that the microbiome of phones and shoes are different. This difference is driven largely by the presence of “environmental” taxa (taxa from groups that tend to be found in places like soil) on shoes and human-associated taxa (taxa from groups that are abundant in the human microbiome) on phones. This large dataset also contains many novel taxa, highlighting the fact that much of microbial diversity remains uncharacterized, even on commonplace objects.

Repeated measures random forests (RMRF): Identifying factors associated with nocturnal hypoglycemia

Nocturnal hypoglycemia is a common phenomenon among patients with diabetes and can lead to a broad range of adverse events and complications. Identifying factors associated with hypoglycemia can improve glucose control and patient care. We propose a repeated measures random forest (RMRF) algorithm that can handle nonlinear relationships and interactions and the correlated responses from patients evaluated over several nights. Simulation results show that our proposed algorithm captures the informative variable more often than naïvely assuming independence. RMRF also outperforms standard random forest and extremely randomized trees algorithms. We demonstrate scenarios where RMRF attains greater prediction accuracy than generalized linear models. We apply the RMRF algorithm to analyze a diabetes study with 2524 nights from 127 patients with type 1 diabetes. We find that nocturnal hypoglycemia is associated with HbA1c, bedtime blood glucose (BG), insulin on board, time system activated, exercise intensity, and daytime hypoglycemia. The RMRF can accurately classify nights at high risk of nocturnal hypoglycemia.

An efficient and adaptive test of auditory mental imagery

The ability to silently hear music in the mind has been argued to be fundamental to musicality. Objective measurements of this subjective imagery experience are needed if this link between imagery ability and musicality is to be investigated. However, previous tests of musical imagery either rely on self-report, rely on melodic memory, or do not cater in range of abilities. The Pitch Imagery Arrow Task (PIAT) was designed to address these shortcomings; however, it is impractically long. In this paper, we shorten the PIAT using adaptive testing and automatic item generation. We interrogate the cognitive processes underlying the PIAT through item response modelling. The result is an efficient online test of auditory mental imagery ability (adaptive Pitch Imagery Arrow Task: aPIAT) that takes 8 min to complete, is adaptive to participant's individual ability, and so can be used to test participants with a range of musical backgrounds. Performance on the aPIAT showed positive moderate-to-strong correlations with measures of non-musical and musical working memory, self-reported musical training, and general musical sophistication. Ability on the task was best predicted by the ability to maintain and manipulate tones in mental imagery, as well as to resist perceptual biases that can lead to incorrect responses. As such, the aPIAT is the ideal tool in which to investigate the relationship between pitch imagery ability and musicality.

Exploring the Potential of High-Resolution Satellite Imagery for the Detection of Soybean Sudden Death Syndrome

Sudden death syndrome (SDS) is one of the major yield-limiting soybean diseases in the Midwestern United States. Effective management for SDS requires accurate detection in soybean fields. Since traditional scouting methods are time-consuming, labor-intensive, and often destructive, alternative methods to monitor SDS in large soybean fields are needed. This study explores the potential of using high-resolution (3 m) PlanetScope satellite imagery for detection of SDS using the random forest classification algorithm. Image data from blue, green, red, and near-infrared (NIR) spectral bands, the calculated normalized difference vegetation index (NDVI), and crop rotation information were used to detect healthy and SDS-infected quadrats in a soybean field experiment with different rotation treatments, located in Boone County, Iowa. Datasets collected during the 2016, 2017, and 2018 soybean growing seasons were analyzed. The results indicate that spectral features, when combined with ground-based information, can detect areas in soybean plots that are at risk for disease, even before foliar symptoms develop. The classification of healthy and diseased soybean quadrats was >75% accurate and the area under the receiver operating characteristic curve (AUROC) was >70%. Our results indicate that high-resolution satellite imagery and random forest analyses have the potential to detect SDS in soybean fields, and that this approach may facilitate large-scale monitoring of SDS (and possibly other economically important soybean diseases). It may also be useful for guiding recommendations for site-specific management in current and future seasons.

Screening and Identification of Potential Biomarkers for Hepatocellular Carcinoma: An Analysis of TCGA Database and Clinical Validation

Introduction

Hepatocellular carcinoma (HCC) is the fifth most common cancer in the world. Up to now, many genes associated with HCC have not yet been identified. In this study, we screened the HCC-related genes through the integrated analysis of the TCGA database, of which the potential biomarkers were also further validated by clinical specimens. The discovery of potential biomarkers for HCC provides more opportunities for diagnostic indicators or gene-targeted therapies.

Methods

Cancer-related genes in The Cancer Genome Atlas (TCGA) HCC database were screened by a random forest (RF) classifier based on the RF algorithm. Proteins encoded by the candidate genes and other associated proteins obtained via protein–protein interaction (PPI) analysis were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses. The newly identified genes were further validated in the HCC cell lines and clinical tissue specimens by Western blotting, immunofluorescence, and immunohistochemistry (IHC). Survival analysis verified the clinical value of genes.

Results

Ten genes with the best feature importance in the RF classifier were screened as candidate genes. By comprehensive analysis of PPI, GO and KEGG, these genes were confirmed to be closely related to HCC tumors. Representative NOX4 and FLVCR1 were selected for further validation by biochemical analysis which showed upregulation in both cancer cell lines and clinical tumor tissues. High expression of NOX4 or FLVCR1 in cancer cells predicts low survival.

Conclusion

Herein, we report that NOX4 and FLVCR1 are promising biomarkers for HCC that may be used as diagnostic indicators or therapeutic targets.

Mapping patient reported outcome measures for low back pain to the International Classification of Functioning, Disability and Health using random forests

BACKGROUND

There is need for feasible and efficient concepts to document patients functioning impairment according to the International Classification of Functioning, Disability and Health (ICF) without imposing additional burden to clinical practice.

AIM

The aim of this study was to develop and validate an automatic linking approach that translates information derived from patient reported outcome measures (PROMs) into the ICF.

DESIGN

Proof-of-concept study.

SETTING

Participants completed both the Roland-Morris disability questionnaire and the Pain Disability Index and were interviewed using the activity and participation component of the ICF brief core set for low back pain.

POPULATION

A total of 244 patients with light to moderate chronic low back pain (cLBP); additionally, 19 patients with higher levels of pain were recruited and assessed for validation purposes.

METHODS

Based on information extracted from the PROMs and considering the factors age and gender, random forest models that predicted the presence or absence of an impairment at the specific ICF category were computed and validated.

RESULTS

Accuracy of the models was found to be acceptable for the most relevant ICF brief core set categories for low back pain if applied at the population level.

CONCLUSIONS

The presented approach can be assumed valid if applied at large on population level. The results are of relevance for the further development of automatic linking programs that would allow the ICF-based classification of functioning properties within the International Classification of Diseases (ICD-11) for any health condition.

CLINICAL REHABILITATION IMPACT

The presented approach eases the documentation of patients' functioning impairment according to the standardized ICF.

Complex Interactions Between Weather, and Microbial and Physicochemical Water Quality Impact the Likelihood of Detecting Foodborne Pathogens in Agricultural Water

Agricultural water is an important source of foodborne pathogens on produce farms. Managing water-associated risks does not lend itself to one-size-fits-all approaches due to the heterogeneous nature of freshwater environments. To improve our ability to develop location-specific risk management practices, a study was conducted in two produce-growing regions to (i) characterize the relationship between Escherichia coli levels and pathogen presence in agricultural water, and (ii) identify environmental factors associated with pathogen detection. Three AZ and six NY waterways were sampled longitudinally using 10-L grab samples (GS) and 24-h Moore swabs (MS). Regression showed that the likelihood of Salmonella detection (Odds Ratio [OR] = 2.18), and eaeA-stx codetection (OR = 6.49) was significantly greater for MS compared to GS, while the likelihood of detecting L. monocytogenes was not. Regression also showed that eaeA-stx codetection in AZ (OR = 50.2) and NY (OR = 18.4), and Salmonella detection in AZ (OR = 4.4) were significantly associated with E. coli levels, while Salmonella detection in NY was not. Random forest analysis indicated that interactions between environmental factors (e.g., rainfall, temperature, turbidity) (i) were associated with likelihood of pathogen detection and (ii) mediated the relationship between E. coli levels and likelihood of pathogen detection. Our findings suggest that (i) environmental heterogeneity, including interactions between factors, affects microbial water quality, and (ii) E. coli levels alone may not be a suitable indicator of food safety risks. Instead, targeted methods that utilize environmental and microbial data (e.g., models that use turbidity and E. coli levels to predict when there is a high or low risk of surface water being contaminated by pathogens) are needed to assess and mitigate the food safety risks associated with preharvest water use. By identifying environmental factors associated with an increased likelihood of detecting pathogens in agricultural water, this study provides information that (i) can be used to assess when pathogen contamination of agricultural water is likely to occur, and (ii) facilitate development of targeted interventions for individual water sources, providing an alternative to existing one-size-fits-all approaches.

Machine learning-based prediction of response to growth hormone treatment in Turner syndrome: the LG Growth Study

Background Growth hormone (GH) treatment has become a common practice in Turner syndrome (TS). However, there are only a few studies on the response to GH treatment in TS. The aim of this study is to predict the responsiveness to GH treatment and to suggest a prediction model of height outcome in TS. Methods The clinical parameters of 105 TS patients registered in the LG Growth Study (LGS) were retrospectively reviewed. The prognostic factors for the good responders were identified, and the prediction of height response was investigated by the random forest (RF) method, and also, multiple regression models were applied. Results In the RF method, the most important predictive variable for the increment of height standard deviation score (SDS) during the first year of GH treatment was chronologic age (CA) at start of GH treatment. The RF method also showed that the increment of height SDS during the first year was the most important predictor in the increment of height SDS after 3 years of treatment. In a prediction model by multiple regression, younger CA was the significant predictor of height SDS gain during the first year (32.4% of the variability). After 3 years of treatment, mid-parental height (MPH) and the increment of height SDS during the first year were identified as significant predictors (76.6% of the variability). Conclusions Both the machine learning approach and the multiple regression model revealed that younger CA at the start of GH treatment was the most important factor related to height response in patients with TS.

Role of Discourse Information in Urdu Sentiment Classification

In computational linguistics, sentiment analysis refers to the classification of opinions in a positive class or a negative class. There exist a lot of different methods for sentiment analysis of the English language, but the literature lacks the availability of methods and techniques for Urdu, which is the largely spoken language in the South Asian sub-continent and the national language of Pakistan. The currently available techniques, such as adjective count method known as Bag of Words (BoW), is not sufficient for classification of complex sentiment written in the Urdu language. Also, the performance of available machine-learning techniques (with legacy features), for classification of Urdu sentiments, are not comparable with the achieved accuracy of other languages. In the case of the English language, the discourse information (sub-sentence-level information) boosts the performance of both the BoW method and machine-learning techniques, but there are very few works available that have tested the context-level information for the sentiment analysis of the Urdu language. This research aims to extract the discourse information from the Urdu sentiments and utilise the discourse information to improve the performance and reduce the error rate of existing techniques for Urdu Sentiment classification. The proposed solution extracts the discourse information, suggests a new set of features for machine-learning techniques, and introduces a set of rules to extend the capabilities of the BoW model. The results show that the task has been enhanced significantly and the performance metrics such as recall, precision, and accuracy are increased by 31.25%, 8.46%, and 21.6%, respectively. In future, the proposed technique can be extended to sentiments with more than two sub-opinions, such as for blogs, reviews, and TV talk shows.

PET and CSF amyloid-β status are differently predicted by patient features: information from discordant cases

BACKGROUND

Amyloid-β PET and CSF Aβ42 yield discordant results in 10-20% of memory clinic patients, possibly providing unique information. Although the predictive power of demographic, clinical, genetic, and imaging features for amyloid positivity has previously been investigated, it is unknown whether these features differentially predict amyloid-β status based on PET or CSF or whether this differs by disease stage.

METHODS

We included 768 patients (subjective cognitive decline (SCD, n = 194), mild cognitive impairment (MCI, n = 127), dementia (AD and non-AD, n = 447) with amyloid-β PET and CSF Aβ42 measurement within 1 year. Ninety-seven (13%) patients had discordant PET/CSF amyloid-β status. We performed parallel random forest models predicting separately PET and CSF status using 17 patient features (demographics, APOE4 positivity, CSF (p)tau, cognitive performance, and MRI visual ratings) in the total patient group and stratified by syndrome diagnosis. Thereafter, we selected features with the highest variable importance measure (VIM) as input for logistic regression models, where amyloid status on either PET or CSF was predicted by (i) the selected patient feature and (ii) the patient feature adjusted for the status of the other amyloid modality.

RESULTS

APOE4, CSF tau, and p-tau had the highest VIM for PET and CSF in all groups. In the amyloid-adjusted logistic regression models, p-tau was a significant predictor for PET-amyloid in SCD (OR = 1.02 [1.01-1.04], pFDR = 0.03), MCI (OR = 1.05 [1.02-1.07], pFDR < 0.01), and dementia (OR = 1.04 [1.03-1.05], pFDR < 0.001), but not for CSF-amyloid. APOE4 (OR = 3.07 [1.33-7.07], punc < 0.01) was associated with CSF-amyloid in SCD, while it was only predictive for PET-amyloid in MCI (OR = 9.44 [2.93, 30.39], pFDR < 0.01). Worse MMSE scores (OR = 1.21 [1.03-1.41], punc = 0.02) were associated to CSF-amyloid status in SCD, whereas worse memory (OR = 1.17 [1.05-1.31], pFDR = 0.02) only predicted PET positivity in dementia.

CONCLUSION

Amyloid status based on either PET or CSF was predicted by different patient features, and this varied by disease stage, suggesting that PET-CSF discordance yields unique information. The stronger associations of both APOE4 carriership and worse memory z-scores with CSF-amyloid in SCD suggest that CSF-amyloid is more sensitive early in the disease course. The higher predictive value of CSF p-tau for a positive PET scan suggests that PET is more specific to AD pathology.

Exploring the association between problem drinking and language use on Facebook in young adults

Recent literature suggests that variations in both formal and content aspects of texts shared on social media tend to reflect user-level differences in demographic, psychosocial, and behavioral characteristics. In the present study, we examined associations between language use on Facebook and problematic alcohol use. We collected texts shared on Facebook by a sample of 296 adult social media users (66.9% females; mean age = 28.44 years (SD = 7.38)). Texts were mined using the closed-vocabulary approach based on the Linguistic Inquiry Word Count (LIWC) semantic dictionary, and an open-vocabulary approach performed via Latent Dirichlet Allocation (LDA). Then, we examined associations between emerging textual features and alcohol-drinking scores as assessed using the AUDIT-C questionnaire. As a final aim, we employed the Random Forest machine-learning algorithm to determine and compare the predictive accuracy of closed- and open-vocabulary features over users' AUDIT-C scores. We found use of words about family, school, and positive feelings and emotions to be negatively associated with alcohol use and problematic drinking, while words suggesting interest in sport events, politics and economics, nightlife, and use of coarse language were more frequent among problematic drinkers. Results coming from LIWC and LDA analyses were quite similar, but LDA added information that could not be retrieved only with LIWC analysis. Furthermore, open-vocabulary features outperformed closed-vocabulary features in terms of predictive power over participants’ AUDIT-C scores (r = .46 vs. r = .28, respectively). Emerging relationships between text features and offline behaviors may have important implications for alcohol screening purposes in the online environment.