GLMMLasso: An Algorithm for High-Dimensional Generalized Linear Mixed Models Using ℓ1-Penalization

Exploring the contribution of inter-individual factors to the development of physiological heat strain in older adults exposed to simulated indoor overheating

Older adults are at elevated risk of heat-related mortality due to age-associated declines in thermoregulatory and cardiovascular function. However, the inter-individual factors that exacerbate physiological heat strain during heat exposure remain unclear, making it challenging to identify more heat-vulnerable subgroups. We therefore explored factors contributing to inter-individual variability in physiological responses of older adults exposed to simulated hot weather. Thirty-seven older adults (61-80 years, 16 females) rested for 8 h in 31 and 36 °C (45% relative humidity). Core (rectal) temperature, heart rate (HR), HR variability, mean arterial pressure (MAP), and cardiac autonomic responses to standing were measured at baseline and end-exposure. Bootstrapped least absolute shrinkage and selection operator regression was used to evaluate whether variation in these responses was related to type 2 diabetes (T2D, n = 10), hypertension (n = 18), age, sex, body morphology, habitual physical activity levels, and/or heat-acclimatization. T2D was identified as a predictor of end-exposure HR (with vs. without: 13 beats/min (bootstrap 95% confidence interval: 6, 23)), seated MAP (-7 mmHg (-18, 1)), and the systolic pressure response to standing (20 mmHg (4, 36)). HR was also influenced by sex (female vs. male: 8 beats/min (1, 16)). No other predictors were identified. The inter-individual factors explored did not meaningfully contribute to the variation in body temperature responses in older adults exposed to simulated indoor overheating. By contrast, cardiovascular responses were exacerbated in females and individuals with T2D. These findings improve understanding of how inter-individual differences contribute to heat-induced physiological strain in older persons.

The plasma proteome is linked with left ventricular and left atrial function parameters in patients with chronic heart failure

AIMS

Examining the systemic biological processes in the heterogeneous syndrome of heart failure with reduced ejection fraction (HFrEF), as reflected by circulating proteins, in relation to echocardiographic characteristics, may provide insights into heart failure pathophysiology. We investigated the link of 4210 repeatedly measured circulating proteins with repeatedly measured echocardiographic parameters as well as with elevated left atrial pressure (LAP), in patients with HFrEF, to provide insights into underlying mechanisms.

METHODS AND RESULTS

In 173 patients with HFrEF, we performed 6-monthly echocardiography and trimonthly blood sampling during a median follow-up of 2.7 (inter-quartile range: 2.5-2.8) years. We investigated circulating proteins in relation to echocardiographic parameters of left ventricular [left ventricular ejection fraction (LVEF), global longitudinal strain (GLS)] and left atrial function [left atrial reservoir strain (LASr)] and elevated LAP (E/e' ratio >15) and used gene enrichment analyses to identify underlying pathophysiological processes. We found 723, 249, 792, and 427 repeatedly measured proteins, with significant associations with LVEF, GLS, LASr, and E/e' ratio, respectively. Proteins associated with LASr reflected pathophysiological mechanisms mostly related to the extracellular matrix. Proteins associated with GLS reflected cardiovascular biological processes and diseases, whereas those associated with LVEF reflected processes involved in the sympathetic nervous system. Moreover, 49 proteins were associated with elevated LAP; after correction for LVEF, three proteins remained: cystatin-D, fibulin-5, and HSP40.

CONCLUSION

Circulating proteins show varying associations with different echocardiographic parameters in patients with HFrEF. These findings suggest that pathways involved in atrial and ventricular dysfunction, as reflected by the plasma proteome, are distinct.

Mental workload evaluation using weighted phase lag index and coherence features extracted from EEG data

Electroencephalogram (EEG) represents an effective, non-invasive technology to study mental workload. However, volume conduction, a common EEG artifact, influences functional connectivity analysis of EEG data. EEG coherence has been used traditionally to investigate functional connectivity between brain areas associated with mental workload, while weighted Phase Lag Index (wPLI) is a measure that improves on coherence by reducing susceptibility to volume conduction, a common EEG artifact. The goal of this study was to compare two methods of functional connectivity analysis, wPLI and coherence, in the context of mental workload evaluation. The study involved model development for mental workload domains and comparing their performance using coherence-based features, wPLI-based features, and a combination of both. Generalized linear mixed-effects model (GLMM) with the least absolute shrinkage and selection operator (LASSO) feature selection method was used for model development. Results indicated that the model developed using a combination of both feature types demonstrated improved predictive performance across all mental workload domains, compared to models that used each feature type individually. The R2 values were 0.82 for perceived task complexity, 0.71 for distraction, 0.91 for mental demand, 0.85 for physical demand, 0.74 for situational stress, and 0.74 for temporal demand. Furthermore, task complexity and functional connectivity patterns in different brain areas were identified as significant contributors to perceived mental workload (p-value<0.05). Findings showed the potential of using EEG data for mental workload evaluation which suggests that combination of coherence and wPLI can improve the accuracy of mental workload domains prediction. Future research should aim to validate these results on larger, diverse datasets to confirm their generalizability and refine the predictive models.

Prediction of Robotic Anastomosis Competency Evaluation (RACE) metrics during vesico-urethral anastomosis using electroencephalography, eye-tracking, and machine learning

Residents learn the vesico-urethral anastomosis (VUA), a key step in robot-assisted radical prostatectomy (RARP), early in their training. VUA assessment and training significantly impact patient outcomes and have high educational value. This study aimed to develop objective prediction models for the Robotic Anastomosis Competency Evaluation (RACE) metrics using electroencephalogram (EEG) and eye-tracking data. Data were recorded from 23 participants performing robot-assisted VUA (henceforth 'anastomosis') on plastic models and animal tissue using the da Vinci surgical robot. EEG and eye-tracking features were extracted, and participants' anastomosis subtask performance was assessed by three raters using the RACE tool and operative videos. Random forest regression (RFR) and gradient boosting regression (GBR) models were developed to predict RACE scores using extracted features, while linear mixed models (LMM) identified associations between features and RACE scores. Overall performance scores significantly differed among inexperienced, competent, and experienced skill levels (P value < 0.0001). For plastic anastomoses, R2 values for predicting unseen test scores were: needle positioning (0.79), needle entry (0.74), needle driving and tissue trauma (0.80), suture placement (0.75), and tissue approximation (0.70). For tissue anastomoses, the values were 0.62, 0.76, 0.65, 0.68, and 0.62, respectively. The models could enhance RARP anastomosis training by offering objective performance feedback to trainees.

Statistical and computational methods for integrating microbiome, host genomics, and metabolomics data

Large-scale microbiome studies are progressively utilizing multiomics designs, which include the collection of microbiome samples together with host genomics and metabolomics data. Despite the increasing number of data sources, there remains a bottleneck in understanding the relationships between different data modalities due to the limited number of statistical and computational methods for analyzing such data. Furthermore, little is known about the portability of general methods to the metagenomic setting and few specialized techniques have been developed. In this review, we summarize and implement some of the commonly used methods. We apply these methods to real data sets where shotgun metagenomic sequencing and metabolomics data are available for microbiome multiomics data integration analysis. We compare results across methods, highlight strengths and limitations of each, and discuss areas where statistical and computational innovation is needed.

What can hospital emergency admissions prior to cancer diagnosis tell us about socio-economic inequalities in cancer diagnosis? Evidence from population-based data in England

BACKGROUND

More deprived cancer patients are at higher risk of Emergency Presentation (EP) with most studies pointing to lower symptom awareness and increased comorbidities to explain those patterns. With the example of colon cancer, we examine patterns of hospital emergency admissions (HEAs) history in the most and least deprived patients as a potential precursor of EP.

METHODS

We analysed the rates of hospital admissions and their admission codes (retrieved from Hospital Episode Statistics) in the two years preceding cancer diagnosis by sex, deprivation and route to diagnosis (EP, non-EP). To select the conditions (grouped admission codes) that best predict emergency admission, we adapted the purposeful variable selection to mixed-effects logistic regression.

RESULTS

Colon cancer patients diagnosed through EP had the highest number of HEAs than all the other routes to diagnosis, especially in the last 7 months before diagnosis. Most deprived patients had an overall higher rate and higher probability of HEA but fewer conditions associated with it.

CONCLUSIONS

Our findings point to higher use of emergency services for non-specific symptoms and conditions in the most deprived patients, preceding colon cancer diagnosis. Health system barriers may be a shared factor of socio-economic inequalities in EP and HEAs.

Ultrasound-based steatosis grading system using 2D-attenuation imaging: An individual patient data meta-analysis with external validation

BACKGROUND AND AIMS

Noninvasive tools assessing steatosis, such as ultrasonography-based 2D-attenuation imaging (ATI), are needed to tackle the worldwide burden of steatotic liver disease. This one-stage individual patient data (IPD) meta-analysis aimed to create an ATI-based steatosis grading system.

APPROACH AND RESULTS

A systematic review (EMBASE + MEDLINE, 2018-2022) identified studies, including patients with histologically or magnetic resonance imaging proton-density fat fraction (MRI-PDFF)-verified ATI for grading steatosis (S0 to S3). One-stage IPD meta-analyses were conducted using generalized mixed models with a random study-specific intercept. Created ATI-based steatosis grading system (aS0 to aS3) was externally validated on a prospective cohort of patients with type 2 diabetes and metabolic dysfunction-associated steatotic liver disease (n=174, histologically and MRI-PDFF-verified steatosis). Eleven enrolled studies included 1374 patients, classified into S0, S1, S2, and S3 in 45.4%, 35.0%, 9.3%, and 10.3% of the cases. ATI was correlated with histological steatosis ( r = 0.60; 95% CI: 0.52, 0.67; p < 0.001) and MRI-PDFF ( r = 0.70; 95% CI: 0.66, 0.73; p < 0.001) but not with liver stiffness ( r = 0.03; 95% CI: -0.04, 0.11, p = 0.343). Steatosis grade was an independent factor associated with ATI (coefficient: 0.24; 95% CI: [0.22, 0.26]; p < 0.001). ATI marginal means within S0, S1, S2, and S3 subpopulations were 0.59 (95% CI: [0.58, 0.61]), 0.69 (95% CI [0.67, 0.71]), 0.78 (95% CI: [0.76, 0.81]), and 0.85 (95% CI: [0.83, 0.88]) dB/cm/MHz; all contrasts between grades were significant ( p < 0.0001). Three ATI thresholds were calibrated to create a new ATI-based steatosis grading system (aS0 to aS3, cutoffs: 0.66, 0.73, and 0.81 dB/cm/MHz). Its external validation showed Obuchowski measures of 0.84 ± 0.02 and 0.82 ± 0.02 with histologically based and MRI-PDFF-based references.

CONCLUSIONS

ATI is a reliable, noninvasive marker of steatosis. This validated ATI-based steatosis grading system could be valuable in assessing patients with metabolic dysfunction-associated steatotic liver disease.

Efficient computation of high-dimensional penalized generalized linear mixed models by latent factor modeling of the random effects

Modern biomedical datasets are increasingly high-dimensional and exhibit complex correlation structures. Generalized linear mixed models (GLMMs) have long been employed to account for such dependencies. However, proper specification of the fixed and random effects in GLMMs is increasingly difficult in high dimensions, and computational complexity grows with increasing dimension of the random effects. We present a novel reformulation of the GLMM using a factor model decomposition of the random effects, enabling scalable computation of GLMMs in high dimensions by reducing the latent space from a large number of random effects to a smaller set of latent factors. We also extend our prior work to estimate model parameters using a modified Monte Carlo Expectation Conditional Minimization algorithm, allowing us to perform variable selection on both the fixed and random effects simultaneously. We show through simulation that through this factor model decomposition, our method can fit high-dimensional penalized GLMMs faster than comparable methods and more easily scale to larger dimensions not previously seen in existing approaches.

A boosting method to select the random effects in linear mixed models

This paper proposes a novel likelihood-based boosting method for the selection of the random effects in linear mixed models. The nonconvexity of the objective function to minimize, which is the negative profile log-likelihood, requires the adoption of new solutions. In this respect, our optimization approach also employs the directions of negative curvature besides the usual Newton directions. A simulation study and a real-data application show the good performance of the proposal.

Fixed and random effect selections in generalized linear mixed models

Generalized linear mixed models are commonly used to describe relationships between correlated responses and covariates in medical research. In this paper, we propose a simple and easily implementable regularized estimation approach to select both fixed and random effects in generalized linear mixed model. Specifically, we propose to construct and optimize the objective functions using the confidence distributions of model parameters, as opposed to using the observed data likelihood functions, to perform effect selections. Two estimation methods are developed. The first one is to use the joint confidence distribution of model parameters to perform simultaneous fixed and random effect selections. The second method is to use the marginal confidence distributions of model parameters to perform the selections of fixed and random effects separately. With a proper choice of regularization parameters in the adaptive LASSO framework, we show the consistency and oracle properties of the proposed regularized estimators. Simulation studies have been conducted to assess the performance of the proposed estimators and demonstrate computational efficiency. Our method has also been applied to two longitudinal cancer studies to identify demographic and clinical factors associated with patient health outcomes after cancer therapies.

Graph Regression Model for Spatial and Temporal Environmental Data-Case of Carbon Dioxide Emissions in the United States

We develop a new model for spatio-temporal data. More specifically, a graph penalty function is incorporated in the cost function in order to estimate the unknown parameters of a spatio-temporal mixed-effect model based on a generalized linear model. This model allows for more flexible and general regression relationships than classical linear ones through the use of generalized linear models (GLMs) and also captures the inherent structural dependencies or relationships of the data through this regularization based on the graph Laplacian. We use a publicly available dataset from the National Centers for Environmental Information (NCEI) in the United States of America and perform statistical inferences of future CO2 emissions in 59 counties. We empirically show how the proposed method outperforms widely used methods, such as the ordinary least squares (OLS) and ridge regression for this challenging problem.

The Impacts of HIV-Related Service Interruptions During the COVID-19 Pandemic: Protocol of a Mixed Methodology Longitudinal Study

The global COVID-19 pandemic has imposed unprecedented pressure on health systems and has interrupted public health efforts for other major health conditions, including HIV. It is critical to comprehensively understand how the pandemic has affected the delivery and utilization of HIV-related services and what are the effective strategies that may mitigate the negative impacts of COVID-19 and resultant interruptions. The current study thus aims to comprehensively investigate HIV service interruptions during the pandemic following a socioecological model, to assess their impacts on various outcomes of the HIV prevention and treatment cascade and to identify resilience resources for buffering impacts of interruptions on HIV treatment cascade outcomes. We will assess HIV service interruptions in South Carolina (SC) since 2020 using operational report data from Ryan White HIV clinics and HIV service utilization data (including telehealth use) based on statewide electronic health records (EHR) and cellphone-based place visitation data. We will further explore how HIV service interruptions affect HIV prevention and treatment cascade outcomes at appropriate geospatial units based on the integration of multi-type, multi-source datasets (e.g., EHR, geospatial data). Finally, we will identify institutional-, community-, and structural-level factors (e.g., resilience resources) that may mitigate the adverse impacts of HIV service interruptions based on the triangulation of quantitative (i.e., EHR data, geospatial data, online survey data) and qualitative (i.e., in-depth interviews with clinic leaders, healthcare providers, people living with HIV, and HIV clinic operational reports) data regarding health system infrastructure, social capital, and organizational preparedness. Our proposed research can lead to a better understanding of complicated HIV service interruptions in SC and resilience factors that can mitigate the negative effects of such interruptions on various HIV treatment cascade outcomes. The multilevel resilience resources identified through data triangulation will assist SC health departments and communities in developing strategic plans in response to this evolving pandemic and other future public health emergencies (e.g., monkeypox, disasters caused by climate change). The research findings can also inform public health policymaking and the practices of other Deep South states with similar sociocultural contexts in developing resilient healthcare systems and communities and advancing epidemic preparedness.

Deep learning and differential equations for modeling changes in individual-level latent dynamics between observation periods

When modeling longitudinal biomedical data, often dimensionality reduction as well as dynamic modeling in the resulting latent representation is needed. This can be achieved by artificial neural networks for dimension reduction and differential equations for dynamic modeling of individual-level trajectories. However, such approaches so far assume that parameters of individual-level dynamics are constant throughout the observation period. Motivated by an application from psychological resilience research, we propose an extension where different sets of differential equation parameters are allowed for observation subperiods. Still, estimation for intra-individual subperiods is coupled for being able to fit the model also with a relatively small dataset. We subsequently derive prediction targets from individual dynamic models of resilience in the application. These serve as outcomes for predicting resilience from characteristics of individuals, measured at baseline and a follow-up time point, and selecting a small set of important predictors. Our approach is seen to successfully identify individual-level parameters of dynamic models that allow to stably select predictors, that is, resilience factors. Furthermore, we can identify those characteristics of individuals that are the most promising for updates at follow-up, which might inform future study design. This underlines the usefulness of our proposed deep dynamic modeling approach with changes in parameters between observation subperiods.

Application of multilevel models for predicting pain following root canal treatment

OBJECTIVES

This study developed predictive models for one-week acute and six-month persistent pain following root canal treatment (RCT). An additional aim was to study the gain in predictive efficacy of models containing clinical factors only, over models containing sociodemographic characteristics.

METHODS

A secondary data analysis of 708 patients who received RCTs was conducted. Three sets of predictors were used: (1) combined set, containing all predictors in the data set; (2) clinical set and (3) sociodemographic set. Missing data were handled by multiple imputation using the missing indicator method. The multilevel least absolute selection and shrinkage operator (LASSO) regression was used to select predictors into the final multilevel logistic models. Three measures, the area under the receiver operating characteristic curve (AUROC) and precision-recall curve (AUPRC) and calibration curves, were used to assess the predictive performance of the models.

RESULTS

The selected-in factors in the final models, using LASSO regression, are related to pre- and intra-treatment clinical symptoms and pain experience. Predictive performance of the models remained the same with the inclusion (exclusion) of the socio-demographic factors. For predicting one-week outcome, the model built with combined set of predictors yielded the highest AUROC and AUPRC of 0.85 and 0.72, followed by the models built with clinical factors (AUROC = 0.82, AUPRC = 0.66). The lowest predictive ability was found in models with only sociodemographic characteristics (AUROC = 0.68, AUPRC = 0.40). Similar patterns were observed in predicting six-month outcome, where the AUROC for models with combined, clinical and sociodemographic sets of predictors were 0.85, 0.89 and 0.66, respectively, and the AUPRC were 0.48, 0.53 and 0.22, respectively.

CONCLUSIONS

Clinical factors such as the severity and experience of pre-operative and intra-operative pain were discovered important to the subsequent development of pain following RCTs. Adding sociodemographic characteristics to the models with clinical factors did not change the models' predictive performance or the proportion of explained variance.

Classification prediction of early pulmonary nodes based on weighted gene correlation network analysis and machine learning

OBJECTIVE

To use weighted gene correlation network analysis (WGCNA) and machine learning algorithm to predict classification of early pulmonary nodes with public databases.

METHODS

The expression data and clinical data of lung cancer patients were firstly extracted from public database (GTEx and TCGA) to study the differentially expressed genes (DEGs) of lung adenocarcinoma (LUAD). The intersection of three R packages (Dseq2, Limma, EdgeR) methods were selected as candidate DEGs for further study. WGCNA was used to obtain relevant modules and key genes of lung cancer classification, GO and KEGG enrichment analysis was performed. The model was built using two machine learning methods, Least Absolute Shrinkage and Selection Operator (LASSO) regression and tumor classification was also predicted with extreme Gradient Boosting (XGBoost) algorithm.

RESULTS

DEGs analysis revealed that there were 1306 LUAD genes. WGCNA module analysis showed that a total of 116 genes were significantly related to classification, and module genes were mainly related to 14 KEGG pathways. The machine learning algorithm identified 10 target genes by LASSO regression analysis of differential genes, and 18 genes were identified by XGBoost model. A total of 6 genes were found from the intersection of the above methods as classification signatures of early pulmonary nodules, including "HMGB3" "ARHGAP6" "TCF21" "FCN3" "COL6A6" "GOLM1".

CONCLUSION

Using DEGs analysis, WGCNA method and machine learning algorithm, six gene signatures related to early stage of LUAD, which can assist clinicians in disease classification prediction.

Stability selection for mixed effect models with large numbers of predictor variables: A simulation study

Covariate selection when the number of available variables is large relative to the number of observations is problematic in epidemiology and remains the focus of continued research. Whilst a variety of statistical methods have been developed to attempt to overcome this issue, at present very few methods are available for wide data that include a clustered outcome. The purpose of this research was to make an empirical evaluation of a new method for covariate selection in wide data settings when the dependent variable is clustered. We used 3300 simulated datasets with a variety of defined structures and known sets of true predictor variables to conduct an empirical evaluation of a mixed model stability selection procedure. Comparison was made with an alternative method based on regularisation using the least absolute shrinkage and selection operator (Lasso) penalty. Model performance was assessed using several metrics including the true positive rate (proportion of true covariates selected in a final model) and false discovery rate (proportion of variables selected in a final model that were non-true (false) variables). For stability selection, the false discovery rate was consistently low, generally remaining ≤ 0.02 indicating that on average fewer than 1 in 50 of the variables selected in a final model were false variables. This was in contrast to the Lasso-based method in which the false discovery rate was between 0.59 and 0.72, indicating that generally more than 60% of variables selected in a final model were false variables. In contrast however, the Lasso method attained higher true positive rates than stability selection, although both methods achieved good results. For the Lasso method, true positive rates remained ≥ 0.93 whereas for stability selection the true positive rate was 0.73-0.97. Our results suggest both methods may be of value for covariate selection with high dimensional data with a clustered outcome. When high specificity is needed for identification of true covariates, stability selection appeared to offer the better solution, although with a slight loss of sensitivity. Conversely when high sensitivity is needed, the Lasso approach may be useful, even if accompanied by a substantial loss of specificity. Overall, the results indicated the loss of sensitivity when employing stability selection is relatively small compared to the loss of specificity when using the Lasso and therefore stability selection may provide the better option for the analyst when evaluating data of this type.

A Lasso and a Regression Tree Mixed-Effect Model with Random Effects for the Level, the Residual Variance, and the Autocorrelation

Research in psychology is experiencing a rapid increase in the availability of intensive longitudinal data. To use such data for predicting feelings, beliefs, and behavior, recent methodological work suggested combinations of the longitudinal mixed-effect model with Lasso regression or with regression trees. The present article adds to this literature by suggesting an extension of these models that-in addition to a random effect for the mean level-also includes a random effect for the within-subject variance and a random effect for the autocorrelation. After introducing the extended mixed-effect location scale (E-MELS), the extended mixed-effect location-scale Lasso model (Lasso E-MELS), and the extended mixed-effect location-scale tree model (E-MELS trees), we show how its parameters can be estimated using a marginal maximum likelihood approach. Using real and simulated example data, we illustrate how to use E-MELS, Lasso E-MELS, and E-MELS trees for building prediction models to forecast individuals' daily nervousness. The article is accompanied by an R package (called mels) and functions that support users in the application of the suggested models.

Characteristics and Outcomes of US Patients Hospitalized With COVID-19

BACKGROUND

Understanding COVID-19 epidemiology is crucial to clinical care and to clinical trial design and interpretation.

OBJECTIVE

To describe characteristics, treatment, and outcomes among patients hospitalized with COVID-19 early in the pandemic.

METHODS

A retrospective cohort study of consecutive adult patients with laboratory-confirmed, symptomatic SARS-CoV-2 infection admitted to 57 US hospitals from March 1 to April 1, 2020.

RESULTS

Of 1480 inpatients with COVID-19, median (IQR) age was 62.0 (49.4-72.9) years, 649 (43.9%) were female, and 822 of 1338 (61.4%) were non-White or Hispanic/Latino. Intensive care unit admission occurred in 575 patients (38.9%), mostly within 4 days of hospital presentation. Respiratory failure affected 583 patients (39.4%), including 284 (19.2%) within 24 hours of hospital presentation and 413 (27.9%) who received invasive mechanical ventilation. Median (IQR) hospital stay was 8 (5-15) days overall and 15 (9-24) days among intensive care unit patients. Hospital mortality was 17.7% (n = 262). Risk factors for hospital death identified by penalized multivariable regression included older age; male sex; comorbidity burden; symptoms-to-admission interval; hypotension; hypoxemia; and higher white blood cell count, creatinine level, respiratory rate, and heart rate. Of 1218 survivors, 221 (18.1%) required new respiratory support at discharge and 259 of 1153 (22.5%) admitted from home required new health care services.

CONCLUSIONS

In a geographically diverse early-pandemic COVID-19 cohort with complete hospital folllow-up, hospital mortality was associated with older age, comorbidity burden, and male sex. Intensive care unit admissions occurred early and were associated with protracted hospital stays. Survivors often required new health care services or respiratory support at discharge.

Penalized Poisson model for network meta-analysis of individual patient time-to-event data

Network meta-analysis (NMA) allows the combination of direct and indirect evidence from a set of randomized clinical trials. Performing NMA using individual patient data (IPD) is considered as a "gold standard" approach as it provides several advantages over NMA based on aggregate data. For example, it allows to perform advanced modeling of covariates or covariate-treatment interactions. An important issue in IPD NMA is the selection of influential parameters among terms that account for inconsistency, covariates, covariate-by-treatment interactions or nonproportionality of treatments effect for time to event data. This issue has not been deeply studied in the literature yet and in particular not for time-to-event data. A major difficulty is to jointly account for between-trial heterogeneity which could have a major influence on the selection process. The use of penalized generalized mixed effect model is a solution, but existing implementations have several shortcomings and an important computational cost that precludes their use for complex IPD NMA. In this article, we propose a penalized Poisson regression model to perform IPD NMA of time-to-event data. It is based only on fixed effect parameters which improve its computational cost over the use of random effects. It could be easily implemented using existing penalized regression package. Computer code is shared for implementation. The methods were applied on simulated data to illustrate the importance to take into account between trial heterogeneity during the selection procedure. Finally, it was applied to an IPD NMA of overall survival of chemotherapy and radiotherapy in nasopharyngeal carcinoma.

Predicting fracture outcomes from clinical registry data using artificial intelligence supplemented models for evidence-informed treatment (PRAISE) study protocol

BACKGROUND

Distal radius (wrist) fractures are the second most common fracture admitted to hospital. The anatomical pattern of these types of injuries is diverse, with variation in clinical management, guidelines for management remain inconclusive, and the uptake of findings from clinical trials into routine practice limited. Robust predictive modelling, which considers both the characteristics of the fracture and patient, provides the best opportunity to reduce variation in care and improve patient outcomes. This type of data is housed in unstructured data sources with no particular format or schema. The "Predicting fracture outcomes from clinical Registry data using Artificial Intelligence (AI) Supplemented models for Evidence-informed treatment (PRAISE)" study aims to use AI methods on unstructured data to describe the fracture characteristics and test if using this information improves identification of key fracture characteristics and prediction of patient-reported outcome measures and clinical outcomes following wrist fractures compared to prediction models based on standard registry data.

METHODS AND DESIGN

Adult (16+ years) patients presenting to the emergency department, treated in a short stay unit, or admitted to hospital for >24h for management of a wrist fracture in four Victorian hospitals will be included in this study. The study will use routine registry data from the Victorian Orthopaedic Trauma Outcomes Registry (VOTOR), and electronic medical record (EMR) information (e.g. X-rays, surgical reports, radiology reports, images). A multimodal deep learning fracture reasoning system (DLFRS) will be developed that reasons on EMR information. Machine learning prediction models will test the performance with/without output from the DLFRS.

DISCUSSION

The PRAISE study will establish the use of AI techniques to provide enhanced information about fracture characteristics in people with wrist fractures. Prediction models using AI derived characteristics are expected to provide better prediction of clinical and patient-reported outcomes following distal radius fracture.

LVS-HARMED Risk Score for Incident Heart Failure in Patients With Atrial Fibrillation Who Present to the Emergency Department: Data from a World-Wide Registry

Background Heart failure (HF) is a common complication to atrial fibrillation (AF), leading to rehospitalization and death. Early identification of patients with AF at risk for HF might improve outcomes. We aimed to derive a score to predict 1-year risk of new-onset HF after an emergency department (ED) visit with AF. Methods and Results The RE-LY AF (Randomized Evaluation of Long-Term Anticoagulant Therapy) registry enrolled patients with AF presenting to an ED in 47 countries, and followed them for a year. The end point was HF hospitalization and/or HF death. Among 15 400 ED patients, 9765 had no prior HF (mean age, 64.9±14.9 years). Within 1 year, new-onset HF developed in 6.8% of patients, of whom 21% died of HF. Independent predictors of HF included left ventricular hypertrophy (odds ratio [OR], 1.47; 95% CI, 1.19-1.82), valvular heart disease (OR, 1.55; 95% CI, 1.18-2.04), smoking (OR, 1.42; 95% CI, 1.12-1.78), height (OR, 0.93; 95% CI, 0.90-0.95 per 3 cm), age (OR, 1.11; 95% CI, 1.07-1.15 per 5 years), rheumatic heart disease (OR, 1.77, 95% CI, 1.24-2.51), prior myocardial infarction (OR, 1.85; 95% CI, 1.45-2.36), remaining in AF at ED discharge (OR, 1.86; 95% CI, 1.46-2.36), and diabetes (OR, 1.33; 95% CI, 1.09-1.64). A continuous risk prediction score (LVS-HARMED [left ventricular, valvular heart disease, smoking or other tobacco use, height, age, rheumatic heart disease, myocardial infarction, emergency department discharge rhythm, and diabetes]) had good discrimination (C statistic, 0.735; 95% CI, 0.716-0.755). Validation was conducted internally using bootstrapping (optimism-corrected C statistic, 0.705) and externally (C statistic, 0.699). The 1-year incidence of HF hospitalization and/or HF death across quartile groups of the score was 1.1%, 4.5%, 6.9%, and 14.4%, respectively. LVS-HARMED also predicted incident stroke (C statistic, 0.753; 95% CI, 0.728-0.778). Conclusions The LVS-HARMED score predicts new-onset HF after an ED visit for AF. Preventative strategies should be considered in patients with high LVS-HARMED HF risk.

A regularized hidden Markov model for analyzing the ‘hot shoe’ in football

We propose a penalized likelihood approach in hidden Markov models (HMMs) to perform automated variable selection. To account for a potential large number of covariates, which also may be substantially correlated, we consider the elastic net penalty containing LASSO and ridge as special cases. By quadratically approximating the non-differentiable penalty, we ensure that the likelihood can be maximized numerically. The feasibility of our approach is assessed in simulation experiments. As a case study, we examine the ‘hot hand’ effect, whose existence is highly debated in different fields, such as psychology and economics. In the present work, we investigate a potential ‘hot shoe’ effect for the performance of penalty takers in (association) football, where the (latent) states of the HMM serve for the underlying form of a player.

A random forest method with feature selection for developing medical prediction models with clustered and longitudinal data

BACKGROUND

Machine learning methodologies are gaining popularity for developing medical prediction models for datasets with a large number of predictors, particularly in the setting of clustered and longitudinal data. Binary Mixed Model (BiMM) forest is a promising machine learning algorithm which may be applied to develop prediction models for clustered and longitudinal binary outcomes. Although machine learning methods for clustered and longitudinal methods such as BiMM forest exist, feature selection has not been analyzed via data simulations. Feature selection improves the practicality and ease of use of prediction models for clinicians by reducing the burden of data collection. Thus, feature selection procedures are not only beneficial, but are often necessary for development of medical prediction models. In this study, we aim to assess feature selection within the BiMM forest setting for modeling clustered and longitudinal binary outcomes.

METHODS

We conducted a simulation study to compare BiMM forest with feature selection (backward elimination or stepwise selection) to standard generalized linear mixed model feature selection methods (shrinkage and backward elimination). We also evaluated feature selection methods to develop models predicting mobility disability in older adults using the Health, Aging and Body Composition Study dataset as an example utilization of the proposed methodology.

RESULTS

BiMM forest with backward elimination generally offered higher computational efficiency, similar or higher predictive performance (accuracy and area under the receiver operating curve), and similar or higher ability to identify correct features compared to linear methods for the different simulated scenarios. For predicting mobility disability in older adults, methods generally performed similarly in terms of accuracy, area under the receiver operating curve, and specificity; however, BiMM forest with backward elimination had the highest sensitivity.

CONCLUSIONS

This study is novel because it is the first investigation of feature selection for developing random forest prediction models for clustered and longitudinal binary outcomes. Results from the simulation study reveal that BiMM forest with backward elimination has the highest accuracy (performance and identification of correct features) and lowest computation time compared to other feature selection methods in some scenarios and similar performance in other scenarios. Many informatics datasets have clustered and longitudinal outcomes and results from this study suggest that BiMM forest with backward elimination may be beneficial for developing medical prediction models.

County-level predictors of retention in care status among people living with HIV in South Carolina from 2010 to 2016: a data-driven approach

OBJECTIVE

The aim of this study was to examine the geospatial variation of retention in care (RIC) across the counties in South Carolina (SC) from 2010 to 2016 and identify the relevant county-level predictors.

DESIGN

Aggregated data on county-level RIC among HIV patients from 2010 to 2016 were retrieved from an electronic HIV/AIDS reporting system in SC Department of Health and Environmental Control. Sociological framework of health was used to select potential county-level predictors from multiple public datasets.

METHODS

Geospatial mapping was used to display the spatial heterogeneity of county-level RIC rate in SC. Generalized linear mixed effect regression with least absolute shrinkage and selection operator (LASSO) was employed to identify county-level predictors related to the change of RIC status over time. Confusion matrix and area under the curve statistics were used to evaluate model performance.

RESULTS

More than half of the counties had their RIC rates lower than the national average. The change of county-level RIC rate from 2010 to 2016 was not significant, and spatial heterogeneity in RIC rate was identified. A total of 22 of the 31 county-level predictors were selected by LASSO for predicting county-level RIC status. Counties with lower collective efficacy, larger proportions of men and/or persons with high education were more likely to have their RIC rates lower than the national average. In contrast, numbers of accessible mental health centres were positively related to county-level RIC status.

CONCLUSION

Spatial variation in RIC could be identified, and county-level factors associated with accessible healthcare facilities and social capital significantly contributed to these variations. Structural and individual interventions targeting these factors are needed to improve the county-level RIC and reduce the spatial variation in HIV care.

Characteristics of Caves Used by Wintering Bats in a Subtropical Environment

Many cave-roosting bats are declining in number throughout their range because of multiple threats, including disease, disturbance by humans, and habitat loss. Successful conservation of cave-roosting bats requires an understanding of the roles of microclimate, cave structure, and the surrounding landscape on the distribution and abundance of bats, and in particular, the use of winter cave roosts. Cave bats have been well studied in temperate climates, but we know little about these bats in subtropical climates. From 2015 to 2017, we conducted 399 winter surveys of 162 caves in Florida to evaluate factors affecting the presence and abundance of bats at cave roosts. We also determined whether temperatures at these cave roosts were suitable for Pseudogymnoascus destructans, the fungus that causes white-nose syndrome. Across all 3 y, we detected three species of bats: tricolored bat Perimyotis subflavus in 126 (77.8%) caves, southeastern myotis Myotis austroriparius in 51 (31.5%) caves, and a single Rafinesque's big-eared bat Corynorhinus rafinesquii. We analyzed the presence of both tricolored bats and southeastern myotis in caves by using mixed-effects models with a binomial distribution and the number of tricolored bats by using generalized linear mixed-effects models with a Poisson error distribution. Tricolored bat presence was significantly and positively influenced by a cooler cave surface temperature, larger cave entrance size, a single entrance, and the presence of solution holes. Tricolored bat abundance increased in longer, cooler caves that had solution holes, a single, unobstructed entrance, and had no signs of flooding. Southeastern myotis presence was positively associated with longer, domed caves. To be most effective, conservation efforts in subtropical climates should focus on caves with the identified characteristics. More than 90% of caves surveyed each year had a cave interior surface temperature below the upper critical growth limit for P. destructans (19.8°C), indicating that most Florida caves are suitable for growth of P. destructans. Therefore, regular monitoring of winter bat populations in subtropical caves may be important for detecting P. destructans and white-nose syndrome.

Ensemble Linear Subspace Analysis of High-Dimensional Data

Regression models provide prediction frameworks for multivariate mutual information analysis that uses information concepts when choosing covariates (also called features) that are important for analysis and prediction. We consider a high dimensional regression framework where the number of covariates (p) exceed the sample size (n). Recent work in high dimensional regression analysis has embraced an ensemble subspace approach that consists of selecting random subsets of covariates with fewer than p covariates, doing statistical analysis on each subset, and then merging the results from the subsets. We examine conditions under which penalty methods such as Lasso perform better when used in the ensemble approach by computing mean squared prediction errors for simulations and a real data example. Linear models with both random and fixed designs are considered. We examine two versions of penalty methods: one where the tuning parameter is selected by cross-validation; and one where the final predictor is a trimmed average of individual predictors corresponding to the members of a set of fixed tuning parameters. We find that the ensemble approach improves on penalty methods for several important real data and model scenarios. The improvement occurs when covariates are strongly associated with the response, when the complexity of the model is high. In such cases, the trimmed average version of ensemble Lasso is often the best predictor.

Generalized multi-SNP mediation intersection-union test

To elucidate the molecular mechanisms underlying genetic variants identified from genome-wide association studies (GWAS) for a variety of phenotypic traits encompassing binary, continuous, count, and survival outcomes, we propose a novel and flexible method to test for mediation that can simultaneously accommodate multiple genetic variants and different types of outcome variables. Specifically, we employ the intersection-union test approach combined with the likelihood ratio test to detect mediation effect of multiple genetic variants via some mediator (e.g., the expression of a neighboring gene) on outcome. We fit high-dimensional generalized linear mixed models under the mediation framework, separately under the null and alternative hypothesis. We leverage Laplace approximation to compute the marginal likelihood of outcome and use coordinate descent algorithm to estimate corresponding parameters. Our extensive simulations demonstrate the validity of our proposed methods and substantial, up to 97%, power gains over alternative methods. Applications to real data for the study of Chlamydia trachomatis infection further showcase advantages of our methods. We believe our proposed methods will be of value and general interest in this post-GWAS era to disentangle the potential causal mechanism from DNA to phenotype for new drug discovery and personalized medicine.

Predicting spatial and temporal responses to non-pharmaceutical interventions on COVID-19 growth rates across 58 counties in New York State: A prospective event-based modeling study on county-level sociological predictors

BACKGROUND

Non-pharmaceutical interventions (NPIs) have been implemented in the New York State since the COVID-19 outbreak on March 1, 2020 to control the transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Socioeconomic heterogeneity across counties closely manifests differences in the post-NPIs growth rate of incidence, which is a crucial indicator to guide future infectious control policy making. Few studies, however, examined the geospatial and sociological variations in the epidemic growth across different time points of NPIs.

OBJECTIVE

To guide a more effective reopening plan while controlling the transmission, the current study aims at 1) identifying hotspots of the growth rate of COVID-19 incidence among the 57 counties and New York City in NYS over time, and 2) examining the association of COVID-19 growth rates after eight critical NPIs time points and most relevant county-level sociological predictors.

METHODS

County-level COVID-19 incidence rates were retrieved from the Social Explorer Website between March 7, 2020 to June 22, 2020. 5-day moving average growth rates of COVID-19 incidence were calculated for 16 selected time points, including the dates of eight NPIs and their respective 14-day-lag-behind time points. A total of 36 county-level indicators were extracted from multiple public datasets. Geospatial mapping and heatmap were used to analyze spatial and temporal heterogeneity of county-level COVID-19 outbreak over selected NPIs-related dates. Generalized mixed effect least absolute shrinkage and selection operator (LASSO) regression, controlling for the 5-day moving average growth rates of COVID-19 testing rates, was employed to identify significant county-level predictors related to the changes of county-level COVID-19 growth rates over time.

RESULTS

COVID-19 infection increased and peaked by the end of March (η=22.50%). Growth rates of COVID-19 decreased by 50.48% after implementing NPIs such as closures of schools, non-essential businesses, parks, and subways. There was a geospatial shift in the region with the highest growth rates from New York metropolitan area towards Western and Northern regions over time. Proportions of population aged 45 years and above (β=3.25 [0.17-6.32]), living alone at residential houses (β=3.31 [0.39--6.22]), and proportion of crowd residential houses (β=6.15 [2.15-10.14]) were positively associated with the growth rate of COVID-19 infection. In contrast, living alone at rental houses (β=-2.47 [-4.83--0.12]) and rate of mental health providers (β=-1.11 [-1.95--0.28]) were negatively associated with COVID-19 growth rate across all 16 time points.

CONCLUSIONS

There are geospatial differences in COVID-19 incidence after implementing different NPIs. Socioeconomic, racial/ethnic, and healthcare resource disparities at the structural and historical levels across counties need to be considered in infection control policymaking to narrow the unequal health impact on vulnerable populations effectively.

CLINICALTRIAL

An iterative algorithm for joint covariate and random effect selection in mixed effects models

We consider joint selection of fixed and random effects in general mixed-effects models. The interpretation of estimated mixed-effects models is challenging since changing the structure of one set of effects can lead to different choices of important covariates in the model. We propose a stepwise selection algorithm to perform simultaneous selection of the fixed and random effects. It is based on Bayesian Information criteria whose penalties are adapted to mixed-effects models. The proposed procedure performs model selection in both linear and nonlinear models. It should be used in the low-dimension setting where the number of ovariates and the number of random effects are moderate with respect to the total number of observations. The performance of the algorithm is assessed via a simulation study, which includes also a comparative study with alternatives when available in the literature. The use of the method is illustrated in the clinical study of an antibiotic agent kinetics.

A generalization of functional clustering for discrete multivariate longitudinal data

This paper presents a new model-based generalized functional clustering method for discrete longitudinal data, such as multivariate binomial and Poisson distributed data. For this purpose, we propose a multivariate functional principal component analysis (MFPCA)-based clustering procedure for a latent multivariate Gaussian process instead of the original functional data directly. The main contribution of this study is two-fold: modeling of discrete longitudinal data with the latent multivariate Gaussian process and developing of a clustering algorithm based on MFPCA coupled with the latent multivariate Gaussian process. Numerical experiments, including real data analysis and a simulation study, demonstrate the promising empirical properties of the proposed approach.

Individual and collective encoding of risk in animal groups

The need to make fast decisions under risky and uncertain conditions is a widespread problem in the natural world. While there has been extensive work on how individual organisms dynamically modify their behavior to respond appropriately to changing environmental conditions (and how this is encoded in the brain), we know remarkably little about the corresponding aspects of collective information processing in animal groups. For example, many groups appear to show increased "sensitivity" in the presence of perceived threat, as evidenced by the increased frequency and magnitude of repeated cascading waves of behavioral change often observed in fish schools and bird flocks under such circumstances. How such context-dependent changes in collective sensitivity are mediated, however, is unknown. Here we address this question using schooling fish as a model system, focusing on 2 nonexclusive hypotheses: 1) that changes in collective responsiveness result from changes in how individuals respond to social cues (i.e., changes to the properties of the "nodes" in the social network), and 2) that they result from changes made to the structural connectivity of the network itself (i.e., the computation is encoded in the "edges" of the network). We find that despite the fact that perceived risk increases the probability for individuals to initiate an alarm, the context-dependent change in collective sensitivity predominantly results not from changes in how individuals respond to social cues, but instead from how individuals modify the spatial structure, and correspondingly the topology of the network of interactions, within the group. Risk is thus encoded as a collective property, emphasizing that in group-living species individual fitness can depend strongly on coupling between scales of behavioral organization.

Penalized negative binomial models for modeling an overdispersed count outcome with a high-dimensional predictor space: Application predicting micronuclei frequency

Chromosomal aberrations, such as micronuclei (MN), have served as biomarkers of genotoxic exposure and cancer risk. Guidelines for the process of scoring MN have been presented by the HUman MicroNucleus (HUMN) project. However, these guidelines were developed for assay performance but do not address how to statistically analyze the data generated by the assay. This has led to the application of various statistical methods that may render different interpretations and conclusions. By combining MN with data from other high-throughput genomic technologies such as gene expression microarray data, we may elucidate molecular features involved in micronucleation. Traditional methods that can model discrete (synonymously, count) data, such as MN frequency, require that the number of explanatory variables (P) is less than the number of samples (N). Due to this limitation, penalized models have been developed to enable model fitting for such over-parameterized datasets. Because penalized models in the discrete response setting are lacking, particularly when the count outcome is over-dispersed, herein we present our penalized negative binomial regression model that can be fit when P > N. Using simulation studies we demonstrate the performance of our method in comparison to commonly used penalized Poisson models when the outcome is over-dispersed and applied it to MN frequency and gene expression data collected as part of the Norwegian Mother and Child Cohort Study. Our countgmifs R package is available for download from the Comprehensive R Archive Network and can be applied to datasets having a discrete outcome that is either Poisson or negative binomial distributed and a high-dimensional covariate space.

Detection of Bladder Contractions From the Activity of the External Urethral Sphincter in Rats Using Sparse Regression

Bladder overactivity and incontinence and dysfunction can be mitigated by electrical stimulation of the pudendal nerve applied at the onset of a bladder contraction. Thus, it is important to predict accurately both bladder pressure and the onset of bladder contractions. We propose a novel method for prediction of bladder pressure using a time-dependent spectrogram representation of external urethral sphincter electromyographic (EUS EMG) activity and a least absolute shrinkage and selection operator regression model. There was a statistically significant improvement in prediction of bladder pressure compared with methods based on the firing rate of EUS EMG activity. This approach enabled prediction of the onset of bladder contractions with 91% specificity and 96% sensitivity and may be suitable for closed-loop control of bladder continence.

Measuring associations between the microbiota and repeated measures of continuous clinical variables using a lasso-penalized generalized linear mixed model

Background

Human microbiome studies in clinical settings generally focus on distinguishing the microbiota in health from that in disease at a specific point in time. However, microbiome samples may be associated with disease severity or continuous clinical health indicators that are often assessed at multiple time points. While the temporal data from clinical and microbiome samples may be informative, analysis of this type of data can be problematic for standard statistical methods.

Results

To identify associations between microbiota and continuous clinical variables measured repeatedly in two studies of the respiratory tract, we adapted a statistical method, the lasso-penalized generalized linear mixed model (LassoGLMM). LassoGLMM can screen for associated clinical variables, incorporate repeated measures of individuals, and address the large number of species found in the microbiome. As is common in microbiome studies, when the number of variables is an order of magnitude larger than the number of samples LassoGLMM can be imperfect in its variable selection. We overcome this limitation by adding a pre-screening step to reduce the number of variables evaluated in the model. We assessed the use of this adapted two-stage LassoGLMM for its ability to determine which microbes are associated with continuous repeated clinical measures.

We found associations (retaining a non-zero coefficient in the LassoGLMM) between 10 laboratory measurements and 43 bacterial genera in the oral microbiota, and between 2 cytokines and 3 bacterial genera in the lung. We compared our associations with those identified by the Wilcoxon test after dichotomizing our outcomes and identified a non-significant trend towards differential abundance between high and low outcomes. Our two-step LassoGLMM explained more of the variance seen in the outcome of interest than other variants of the LassoGLMM method.

Conclusions

We demonstrated a method that can account for the large number of genera detected in microbiome studies and repeated measures of clinical or longitudinal studies, allowing for the detection of strong associations between microbes and clinical measures. By incorporating the design strengths of repeated measurements and a prescreening step to aid variable selection, our two-step LassoGLMM will be a useful analytic method for investigating relationships between microbes and repeatedly measured continuous outcomes.

Electronic supplementary material

The online version of this article (10.1186/s13040-018-0173-9) contains supplementary material, which is available to authorized users.

Pharmacometabolomics applied to zonisamide pharmacokinetic parameter prediction

INTRODUCTION

Zonisamide is a new-generation anticonvulsant antiepileptic drug metabolized primarily in the liver, with subsequent elimination via the renal route.

OBJECTIVES

Our objective was to evaluate the utility of pharmacometabolomics in the detection of zonisamide metabolites that could be related to its disposition and therefore, to its efficacy and toxicity.

METHODS

This study was nested to a bioequivalence clinical trial with 28 healthy volunteers. Each participant received a single dose of zonisamide on two separate occasions (period 1 and period 2), with a washout period between them. Blood samples of zonisamide were obtained from all patients at baseline for each period, before volunteers were administered any medication, for metabolomics analysis.

RESULTS

After a Lasso regression was applied, age, height, branched-chain amino acids, steroids, triacylglycerols, diacyl glycerophosphoethanolamine, glycerophospholipids susceptible to methylation, phosphatidylcholines with 20:4 FA (arachidonic acid) and cholesterol ester and lysophosphatidylcholine were obtained in both periods.

CONCLUSION

To our knowledge, this is the only research study to date that has attempted to link basal metabolomic status with pharmacokinetic parameters of zonisamide.

Dopaminergic and Opioid Pathways Associated with Impulse Control Disorders in Parkinson's Disease

Introduction

Impulse control disorders (ICDs) are frequent non-motor symptoms in Parkinson’s disease (PD), with potential negative effects on the quality of life and social functioning. ICDs are closely associated with dopaminergic therapy, and genetic polymorphisms in several neurotransmitter pathways may increase the risk of addictive behaviors in PD. However, clinical differentiation between patients at risk and patients without risk of ICDs is still troublesome. The aim of this study was to investigate if genetic polymorphisms across several neurotransmitter pathways were associated with ICD status in patients with PD.

Methods

Whole-exome sequencing data were available for 119 eligible PD patients from the Norwegian ParkWest study. All participants underwent comprehensive neurological, neuropsychiatric, and neuropsychological assessments. ICDs were assessed using the self-report short form version of the Questionnaire for Impulsive-Compulsive Disorders in PD. Single-nucleotide polymorphisms (SNPs) from 17 genes were subjected to regression with elastic net penalization to identify candidate variants associated with ICDs. The area under the curve of receiver-operating characteristic curves was used to evaluate the level of ICD prediction.

Results

Among the 119 patients with PD included in the analysis, 29% met the criteria for ICD and 63% were using dopamine agonists (DAs). Eleven SNPs were associated with ICDs, and the four SNPs with the most robust performance significantly increased ICD predictability (AUC = 0.81, 95% CI 0.73–0.90) compared to clinical data alone (DA use and age; AUC = 0.65, 95% CI 0.59–0.78). The strongest predictive factors were rs5326 in DRD1, which was associated with increased odds of ICDs, and rs702764 in OPRK1, which was associated with decreased odds of ICDs.

Conclusion

Using an advanced statistical approach, we identified SNPs in nine genes, including a novel polymorphism in DRD1, with potential application for the identification of PD patients at risk for ICDs.

Feature selection for high-dimensional temporal data

BACKGROUND

Feature selection is commonly employed for identifying collectively-predictive biomarkers and biosignatures; it facilitates the construction of small statistical models that are easier to verify, visualize, and comprehend while providing insight to the human expert. In this work we extend established constrained-based, feature-selection methods to high-dimensional "omics" temporal data, where the number of measurements is orders of magnitude larger than the sample size. The extension required the development of conditional independence tests for temporal and/or static variables conditioned on a set of temporal variables.

RESULTS

The algorithm is able to return multiple, equivalent solution subsets of variables, scale to tens of thousands of features, and outperform or be on par with existing methods depending on the analysis task specifics.

CONCLUSIONS

The use of this algorithm is suggested for variable selection with high-dimensional temporal data.

Longitudinal data analysis for rare variants detection with penalized quadratic inference function

Longitudinal genetic data provide more information regarding genetic effects over time compared with cross-sectional data. Coupled with next-generation sequencing technologies, it becomes reality to identify important genes containing both rare and common variants in a longitudinal design. In this work, we adopted a weighted sum statistic (WSS) to collapse multiple variants in a gene region to form a gene score. When multiple genes in a pathway were considered together, a penalized longitudinal model under the quadratic inference function (QIF) framework was applied for efficient gene selection. We evaluated the estimation accuracy and model selection performance under different model settings, then applied the method to a real dataset from the Genetic Analysis Workshop 18 (GAW18). Compared with the unpenalized QIF method, the penalized QIF (pQIF) method achieved better estimation accuracy and higher selection efficiency. The pQIF remained optimal even when the working correlation structure was mis-specified. The real data analysis identified one important gene, angiotensin II receptor type 1 (AGTR1), in the Ca2+/AT-IIR/α-AR signaling pathway. The estimated effect implied that AGTR1 may have a protective effect for hypertension. Our pQIF method provides a general tool for longitudinal sequencing studies involving large numbers of genetic variants.

Adjusted adaptive Lasso for covariate model-building in nonlinear mixed-effect pharmacokinetic models

One important aim in population pharmacokinetics (PK) and pharmacodynamics is identification and quantification of the relationships between the parameters and covariates. Lasso has been suggested as a technique for simultaneous estimation and covariate selection. In linear regression, it has been shown that Lasso possesses no oracle properties, which means it asymptotically performs as though the true underlying model was given in advance. Adaptive Lasso (ALasso) with appropriate initial weights is claimed to possess oracle properties; however, it can lead to poor predictive performance when there is multicollinearity between covariates. This simulation study implemented a new version of ALasso, called adjusted ALasso (AALasso), to take into account the ratio of the standard error of the maximum likelihood (ML) estimator to the ML coefficient as the initial weight in ALasso to deal with multicollinearity in non-linear mixed-effect models. The performance of AALasso was compared with that of ALasso and Lasso. PK data was simulated in four set-ups from a one-compartment bolus input model. Covariates were created by sampling from a multivariate standard normal distribution with no, low (0.2), moderate (0.5) or high (0.7) correlation. The true covariates influenced only clearance at different magnitudes. AALasso, ALasso and Lasso were compared in terms of mean absolute prediction error and error of the estimated covariate coefficient. The results show that AALasso performed better in small data sets, even in those in which a high correlation existed between covariates. This makes AALasso a promising method for covariate selection in nonlinear mixed-effect models.

A Proteome-Derived Longitudinal Pharmacodynamic Biomarker for Diffuse Systemic Sclerosis Skin

In this study we systematically investigated alterations in the serum proteome of patients with diffuse cutaneous systemic sclerosis and identified differentially expressed proteins that correlated with disease severity. Our goal was to identify a combination of serum proteins that would provide a biological measure for the extent of skin disease and that could be combined into a longitudinal pharmacodynamic biomarker. We found that 16% of the sera proteins analyzed by SOMAscan aptamer technology, from two cohorts of patients with diffuse cutaneous systemic sclerosis, were identified as differentially regulated between diffuse cutaneous systemic sclerosis and controls and correlated with modified Rodnan skin score. This dataset showed tumor necrosis factor-α, IFN-γ, transforming growth factor-β, and IL-13 as potential upstream regulators of the serum protein patterns in the sera of patients with diffuse cutaneous systemic sclerosis. By ELISA, two analytes (ST2 and Spondin-1) best described longitudinal change in modified Rodnan skin score, using linear mixed models. This model was then validated in three independent cohorts. In this study we discovered a large array of proteins not previously associated with systemic sclerosis that provide insight into pathogenesis and potential targets for therapeutic intervention. Furthermore, we show that two of these proteins can be combined to form a robust longitudinal biomarker that might be used in clinical trials to assess changes in diffuse cutaneous systemic sclerosis skin disease over time.