Predicting local malaria exposure using a Lasso-based two-level cross validation algorithm.
Ontology highlight
ABSTRACT:
SUBMITTER: Kouwaye B
PROVIDER: S-EPMC5663424 | biostudies-literature | 2017
REPOSITORIES: biostudies-literature
Publications
Predicting local malaria exposure using a Lasso-based two-level cross validation algorithm.
Kouwaye Bienvenue B Rossi Fabrice F Fonton Noël N Garcia André A Dossou-Gbété Simplice S Hounkonnou Mahouton Norbert MN Cottrell Gilles G
PloS one 20171031 10
Recent studies have highlighted the importance of local environmental factors to determine the fine-scale heterogeneity of malaria transmission and exposure to the vector. In this work, we compare a classical GLM model with backward selection with different versions of an automatic LASSO-based algorithm with 2-level cross-validation aiming to build a predictive model of the space and time dependent individual exposure to the malaria vector, using entomological and environmental data from a cohor ...[more]
Similar Datasets
Project description:Background and Objectives: Falls account for the highest proportion of preventable injury among older adults. Thus, the United States' Centers for Disease Control and Prevention (CDC) developed the Stopping Elderly Accidents, Deaths, and Injuries (STEADI) algorithm to screen for fall risk. We referred to our STEADI algorithm adaptation as "Quick-STEADI" and compared the predictive abilities of the three-level (low, moderate, and high risk) and two-level (at-risk and not at-risk) Quick-STEADI algorithms. We additionally assessed the qualitative implementation of the Quick-STEADI algorithm in clinical settings. Research Design and Methods: We followed a prospective cohort (N = 200) of adults (65+ years) in the Bassett Healthcare Network (Cooperstown, NY) for 6 months in 2019. We conducted a generalized linear mixed model, adjusting for sociodemographic variables, to determine how baseline fall risk predicted subsequent daily falls. We plotted receiver operating characteristic (ROC) curves and measured the area under the curve (AUC) to determine the predictive ability of the Quick-STEADI algorithm. We identified a participant sample (N = 8) to gauge the experience of the screening process and a screener sample (N = 3) to evaluate the screening implementation. Results: For the three-level Quick-STEADI algorithm, participants at low and moderate risk for falls had a reduced likelihood of daily falls compared to those at high risk (-1.09, p = 0.04; -0.99, p = 0.04). For the two-level Quick-STEADI algorithm, participants not at risk for falls were not associated with a reduced likelihood of daily falls compared to those at risk (-0.89, p = 0.13). The discriminatory ability of the three-level and two-level Quick-STEADI algorithm demonstrated similar predictability of daily falls, based on AUC (0.653; 0.6570). Furthermore, participants and screeners found the Quick-STEADI algorithm to be efficient and viable. Discussion and Implications: The Quick-STEADI is a suitable, alternative fall risk screening algorithm. Qualitative assessments of the Quick-STEADI algorithm demonstrated feasibility in integrating a falls screening program in a clinical setting. Future research should address the validation and the implementation of the Quick-STEADI algorithm in community health settings to determine if falls screening and prevention can be streamlined in these settings. This may increase engagement in fall prevention programs and decrease overall fall risk among older adults.
| S-EPMC7438745 | biostudies-literature
Project description:Chronic lead (Pb) exposure causes long term health effects. While recent exposure can be assessed by measuring blood lead (half-life 30 days), chronic exposures can be assessed by measuring lead in bone (half-life of many years to decades). Bone lead measurements, in turn, have been measured non-invasively in large population-based studies using x-ray fluorescence techniques, but the method remains limited due to technical availability, expense, and the need for licensing radioactive materials used by the instruments. Thus, we developed prediction models for bone lead concentrations using a flexible machine learning approach--Super Learner, which combines the predictions from a set of machine learning algorithms for better prediction performance. The study population included 695 men in the Normative Aging Study, aged 48 years and older, whose bone (patella and tibia) lead concentrations were directly measured using K-shell-X-ray fluorescence. Ten predictors (blood lead, age, education, job type, weight, height, body mass index, waist circumference, cumulative cigarette smoking (pack-year), and smoking status) were selected for patella lead and 11 (the same 10 predictors plus serum phosphorus) for tibia lead using the Boruta algorithm. We implemented Super Learner to predict bone lead concentrations by calculating a weighted combination of predictions from 8 algorithms. In the nested cross-validation, the correlation coefficients between measured and predicted bone lead concentrations were 0.58 for patella lead and 0.52 for tibia lead, which has improved the correlations obtained in previously-published linear regression-based prediction models. We evaluated the applicability of these prediction models to the National Health and Nutrition Examination Survey for the associations between predicted bone lead concentrations and blood pressure, and positive associations were observed. These bone lead prediction models provide reasonable accuracy and can be used to evaluate health effects of cumulative lead exposure in studies where bone lead is not measured.
| S-EPMC10160242 | biostudies-literature
Project description:Genome-wide association study (GWAS) has turned out to be an essential technology for exploring the genetic mechanism of complex traits. To reduce the complexity of computation, it is well accepted to remove unrelated single nucleotide polymorphisms (SNPs) before GWAS, e.g., by using iterative sure independence screening expectation-maximization Bayesian Lasso (ISIS EM-BLASSO) method. In this work, a modified version of ISIS EM-BLASSO is proposed, which reduces the number of SNPs by a screening methodology based on Pearson correlation and mutual information, then estimates the effects via EM-Bayesian Lasso (EM-BLASSO), and finally detects the true quantitative trait nucleotides (QTNs) through likelihood ratio test. We call our method a two-stage mutual information based Bayesian Lasso (MBLASSO). Under three simulation scenarios, MBLASSO improves the statistical power and retains the higher effect estimation accuracy when comparing with three other algorithms. Moreover, MBLASSO performs best on model fitting, the accuracy of detected associations is the highest, and 21 genes can only be detected by MBLASSO in Arabidopsis thaliana datasets.
| S-EPMC7516787 | biostudies-literature
Project description:BackgroundHeterogeneity in malaria exposure complicates survival analyses of vaccine efficacy trials and confounds the association between immune correlates of protection and malaria infection in longitudinal studies. Analysis may be facilitated by taking into account the variability in individual exposure levels, but it is unclear how exposure can be estimated at an individual level.Method and findingsWe studied three cohorts (Chonyi, Junju and Ngerenya) in Kilifi District, Kenya to assess measures of malaria exposure. Prospective data were available on malaria episodes, geospatial coordinates, proximity to infected and uninfected individuals and residence in predefined malaria hotspots for 2,425 individuals. Antibody levels to the malaria antigens AMA1 and MSP1(142) were available for 291 children from Junju. We calculated distance-weighted local prevalence of malaria infection within 1 km radius as a marker of individual's malaria exposure. We used multivariable modified Poisson regression model to assess the discriminatory power of these markers for malaria infection (i.e. asymptomatic parasitaemia or clinical malaria). The area under the receiver operating characteristic (ROC) curve was used to assess the discriminatory power of the models. Local malaria prevalence within 1 km radius and AMA1 and MSP1(142) antibodies levels were independently associated with malaria infection. Weighted local malaria prevalence had an area under ROC curve of 0.72 (95%CI: 0.66-0.73), 0.71 (95%CI: 0.69-0.73) and 0.82 (95%CI: 0.80-0.83) among cohorts in Chonyi, Junju and Ngerenya respectively. In a small subset of children from Junju, a model incorporating weighted local malaria prevalence with AMA1 and MSP1(142) antibody levels provided an AUC of 0.83 (95%CI: 0.79-0.88).ConclusionWe have proposed an approach to estimating the intensity of an individual's malaria exposure in the field. The weighted local malaria prevalence can be used as individual marker of malaria exposure in malaria vaccine trials and longitudinal studies of natural immunity to malaria.
| S-EPMC3315550 | biostudies-literature
Project description:BackgroundTreatment with atezolizumab and bevacizumab has become standard of care for advanced unresectable hepatocellular carcinoma (HCC) but carries an increased gastrointestinal bleeding risk. Therefore, patients are often required to undergo esophagogastroduodenoscopy (EGD) to rule out esophageal varices (EV) prior to initiating therapy, which can delay care and lead to unnecessary procedural risks and health care costs. In 2019, the EVendo score was created and validated as a noninvasive tool to accurately screen out patients who were at low risk for having EV that required treatment. We sought to validate whether the EVendo score could be used to accurately predict the presence of EV and varices needing treatment (VNT) in patients with HCC.MethodsThis was a retrospective multicenter cohort study of patients with HCC from 9/2004 to 12/2021. We included patients who underwent EGDs within 1 year after their HCC diagnosis. We collected clinical parameters needed to calculate an EVendo score at the time of EGD and compared the EVendo model prediction to the gold standard endoscopic report in predicting presence of VNT.Results112 with HCC were recruited to this study, with 117 qualifying EGDs. VNT occurred in 39 (33.3%) patients. The EVendo score had a sensitivity of 97.4% and a negative predictive value of 96.9%, supporting the validity in applying EVendo in predicting VNT in HCC.ConclusionIn this study, we validated the use of the EVendo score in ruling out VNT in patients with HCC. The application of the EVendo score could safely defer about 30% of EGDs for EV screening in HCC patients. Although additional validation cohorts are needed, this suggests that EVendo score can potentially be applied in patients with HCC to avoid unnecessary EGDs, which can ultimately mitigate healthcare costs and delays in initiating HCC treatment with atezolizumab and bevacizumab.
| S-EPMC11341647 | biostudies-literature
Project description:BackgroundAlgorithms of sequence alignment are the key instruments for computer-assisted studies of biopolymers. Obviously, it is important to take into account the "quality" of the obtained alignments, i.e. how closely the algorithms manage to restore the "gold standard" alignment (GS-alignment), which superimposes positions originating from the same position in the common ancestor of the compared sequences. As an approximation of the GS-alignment, a 3D-alignment is commonly used not quite reasonably. Among the currently used algorithms of a pair-wise alignment, the best quality is achieved by using the algorithm of optimal alignment based on affine penalties for deletions (the Smith-Waterman algorithm). Nevertheless, the expedience of using local or global versions of the algorithm has not been studied.ResultsUsing model series of amino acid sequence pairs, we studied the relative "quality" of results produced by local and global alignments versus (1) the relative length of similar parts of the sequences (their "cores") and their nonhomologous parts, and (2) relative positions of the core regions in the compared sequences. We obtained numerical values of the average quality (measured as accuracy and confidence) of the global alignment method and the local alignment method for evolutionary distances between homologous sequence parts from 30 to 240 PAM and for the core length making from 10% to 70% of the total length of the sequences for all possible positions of homologous sequence parts relative to the centers of the sequences.ConclusionWe revealed criteria allowing to specify conditions of preferred applicability for the local and the global alignment algorithms depending on positions and relative lengths of the cores and nonhomologous parts of the sequences to be aligned. It was demonstrated that when the core part of one sequence was positioned above the core of the other sequence, the global algorithm was more stable at longer evolutionary distances and larger nonhomologous parts than the local algorithm. On the contrary, when the cores were positioned asymmetrically, the local algorithm was more stable at longer evolutionary distances and larger nonhomologous parts than the global algorithm. This opens a possibility for creation of a combined method allowing generation of more accurate alignments.
| S-EPMC3223492 | biostudies-literature
Project description:BackgroundObjective and early identification of hospitalized patients, and particularly those with novel coronavirus disease 2019 (COVID-19), who may require mechanical ventilation (MV) may aid in delivering timely treatment.Research questionCan a transparent deep learning (DL) model predict the need for MV in hospitalized patients and those with COVID-19 up to 24 h in advance?Study design and methodsWe trained and externally validated a transparent DL algorithm to predict the future need for MV in hospitalized patients, including those with COVID-19, using commonly available data in electronic health records. Additionally, commonly used clinical criteria (heart rate, oxygen saturation, respiratory rate, Fio2, and pH) were used to assess future need for MV. Performance of the algorithm was evaluated using the area under receiver operating characteristic curve (AUC), sensitivity, specificity, and positive predictive value.ResultsWe obtained data from more than 30,000 ICU patients (including more than 700 patients with COVID-19) from two academic medical centers. The performance of the model with a 24-h prediction horizon at the development and validation sites was comparable (AUC, 0.895 vs 0.882, respectively), providing significant improvement over traditional clinical criteria (P < .001). Prospective validation of the algorithm among patients with COVID-19 yielded AUCs in the range of 0.918 to 0.943.InterpretationA transparent deep learning algorithm improves on traditional clinical criteria to predict the need for MV in hospitalized patients, including in those with COVID-19. Such an algorithm may help clinicians to optimize timing of tracheal intubation, to allocate resources and staff better, and to improve patient care.
| S-EPMC8027289 | biostudies-literature
Project description:To evaluate the risks of pesticides for pollinators, we must not only evaluate their toxicity but also understand how pollinators are exposed to these xenobiotics in the field. We focused on this last point and modeled honey bee exposure to pesticides at the landscape level. Pollen pellet samples (n = 60) from 40 Belgian apiaries were collected from late July to October 2011 and underwent palynological and pesticide residue analyses. Areas of various crops around each apiary were measured at 4 spatial scales. The most frequently detected pesticides were the fungicides boscalid (n = 19, 31.7%) and pyrimethanil (n = 10, 16.7%) and the insecticide dimethoate (n = 10, 16.7%). We were able to predict exposure probability for boscalid and dimethoate by using broad indicators of cropping intensity, but it remained difficult to identify the precise source of contamination (e.g. specific crops in which the use of the pesticide is authorized). For pyrimethanil, we were not able to build any convincing landscape model that could explain the contamination. Our results, combined with the late sampling period, strongly suggest that pesticides applied to crops unattractive to pollinators, and therefore considered of no risk for them, may be sources of exposure through weeds, drift to neighboring plants, or succeeding crops.
| S-EPMC5476569 | biostudies-literature
Project description:The Lasso shrinkage procedure achieved its popularity, in part, by its tendency to shrink estimated coefficients to zero, and its ability to serve as a variable selection procedure. Using data-adaptive weights, the adaptive Lasso modified the original procedure to increase the penalty terms for those variables estimated to be less important by ordinary least squares. Although this modified procedure attained the oracle properties, the resulting models tend to include a large number of "false positives" in practice. Here, we adapt the concept of local false discovery rates (lFDRs) so that it applies to the sequence, λn, of smoothing parameters for the adaptive Lasso. We define the lFDR for a given λn to be the probability that the variable added to the model by decreasing λn to λn-δ is not associated with the outcome, where δ is a small value. We derive the relationship between the lFDR and λn, show lFDR =1 for traditional smoothing parameters, and show how to select λn so as to achieve a desired lFDR. We compare the smoothing parameters chosen to achieve a specified lFDR and those chosen to achieve the oracle properties, as well as their resulting estimates for model coefficients, with both simulation and an example from a genetic study of prostate specific antigen.
| S-EPMC3769997 | biostudies-literature
Project description:ObjectivesTo present the COVID Local Risk Index (CLRI), a measure of city- and neighborhood-level risk for SARS COV-2 infection and poor outcomes, and validate it using sub-city SARS COV-2 outcome data from 47 large U.S. cities.MethodsCross-sectional validation analysis of CLRI against SARS COV-2 incidence, percent positivity, hospitalization, and mortality. CLRI scores were validated against ZCTA-level SARS COV-2 outcome data gathered in 2020-2021 from public databases or through data use agreements using a negative binomial model.ResultsCLRI was associated with each SARS COV-2 outcome in pooled analysis. In city-level models, CLRI was positively associated with positivity in 11/14 cities for which data were available, hospitalization in 6/6 cities, mortality in 13/14 cities, and incidence in 33/47 cities.ConclusionsCLRI is a valid tool for assessing sub-city risk of SARS COV-2 infection and illness severity. Stronger associations with positivity, hospitalization and mortality may reflect differential testing access, greater weight on components associated with poor outcomes than transmission, omitted variable bias, or other reasons. City stakeholders can use the CLRI, publicly available on the City Health Dashboard (www.cityhealthdashboard.com), to guide SARS COV-2 resource allocation.
| S-EPMC9128556 | biostudies-literature