Lasso Penalization Research Articles

Stock Forecasting Using an Improved Version of Adaptive Group Lasso

Stock is one of the few things in the world that influence the economy of the society, hence its worth a shot and very desirable to predict a stock price in the future. Hence in this paper we propose a stock prediction model based on linear regression model. The database of the training is based on the Goldman Sachs database of stocks found from the Google. Here we choose Lasso penalization technique cause the, this performs well with the sparsity of the network, meaning when the network has less features and more observations. Here we have proposed an improved version of lasso function and have proposed an algorithm to improve the performance of the model.

Nonlinear sparse partial least squares: an investigation of the effect of nonlinearity and sparsity on the decoding of intracranial data

Objective. Partial Least Squares (PLS) regression is a suitable linear decoder model for correlated and high dimensional neural data. This algorithm has been widely used in the application of brain–computer interface (BCI) for the decoding of motor parameters. PLS does not consider nonlinear relations between brain signal features. The nonlinear version of PLS that considers a nonlinear relationship between the latent variables has not been proposed for the decoding of intracranial data. This nonlinear model may cause overfitting in some cases due to a larger number of free parameters. In this paper, we develop a new version of nonlinear PLS, namely nonlinear sparse PLS (NLS PLS) and test it in BCI applications. Approach. In motor related BCI systems, improving the decoding accuracy of both kinetic and kinematic parameters of movement is crucial. To do this, two BCI datasets were chosen to decode the force amplitude and position of hand trajectory using the nonlinear and sparse versions of PLS algorithm. In our new NLS PLS method, we considered a polynomial relationship between the latent variables and used the lasso penalization in the latent space to avoid overfitting and to improve the decoding accuracy. Main results. Some linear and nonlinear based PLS models and our new proposed method, NLS PLS, were applied to the two datasets. According to our results, significant improvement from the NLS PLS method is confirmed over other methods. Our results show that nonlinear PLS outperforms generic PLS in the force decoding but it has lower accuracy in the hand trajectory decoding because of high dimensional feature space. By using lasso penalization, we presented a sparse nonlinear PLS-based model that outperforms generic PLS in both datasets and improves the coefficient of determination, 34% in the force decoding and 10% in the hand trajectory decoding. Significance. We constructed a simple PLS-based model that considers a nonlinear relationship between features and it is also robust to overfitting because of using the lasso penalty in the latent space. This model is suitable for a high dimensional and correlated datasets, like intracranial data and can improve the accuracy of estimation.

Severe Cytokine Release Syndrome Is Associated with Impaired Hematopoietic Recovery after CD19-Targeted CAR-T Cell Therapy

Background Chimeric antigen receptor therapy (CAR-T) directed against CD19 has demonstrated efficacy in patients with relapsed/refractory (R/R) B-cell malignancies. Delayed hematopoietic recovery with grade 3/4 neutropenia and thrombocytopenia, requiring extended growth factor administration or transfusions, has been observed in patients undergoing CAR-T cell therapy, although the factors influencing recovery are poorly understood. In this study, we performed multivariable analyses to identify factors associated with hematopoietic recovery in patients undergoing CD19 CAR-T cell therapy. Methods We retrospectively analyzed 125 patients with R/R acute lymphoblastic leukemia (ALL), non-Hodgkin lymphoma (NHL), and chronic lymphocytic leukemia (CLL), treated with CD19-targeted CAR-T cells on a phase 1/2 clinical trial in our institution (NCT01865617). Patients receiving more than one CAR-T infusion were excluded. Criteria for neutropenia, thrombocytopenia, and recovery were defined as per the Center for International Blood and Marrow Transplant Research (CIBMTR) reporting guidelines: neutropenia, absolute neutrophil count (ANC) ≤ 500/mm3; thrombocytopenia, platelet (Plt) count ≤ 20 x 109/L; neutrophil recovery, ANC > 0.5 x 109/L for three consecutive laboratory values obtained on different days irrespective of growth factor administration; platelet recovery, Plt > 20 x 109/L for three consecutive values obtained on different days in the absence of platelet transfusion for seven days. For competing risk analysis, an event was defined as having achieved ANC or Plt recovery, with the following considered as competing events: death, new cytotoxic therapy, relapse with marrow involvement in the absence of ANC or platelet recovery. Patients who never met the CIBMTR criteria for neutropenia of thrombocytopenia were considered as having recovered at time = 0. To identify factors associated with impaired hematopoietic recovery after CD19 CAR-T cell therapy, patient-, disease- and CAR-T cell therapy-related variables were included in a multivariable Fine and Gray model prior to variable selection using LASSO penalization (Table 2 footnote). Results We included 125 patients (ALL, n=44; CLL, n=37; NHL, n=44) with a median age of 55 (range, 20-76). Patients were heavily pre-treated with a median of 4 prior therapies (range, 1-10); 31% had undergone prior autologous or allogeneic hematopoietic cell transplantation (HCT). Median ANC and Plt prior to lymphodepletion were 2 x 109/L (range 0-23) and 112 x 109/L, range 3-425), respectively. Patient and treatment characteristics are summarized in Table 1. ANC and Plt recovery after CD19 CAR-T cell therapy were observed in 91% (ALL, 86%; CLL, 92%; NHL, 95%) and 86% (ALL, 86%; CLL, 86%; NHL, 84%) of patients, respectively. Median time to ANC recovery was 9 days and the probability of ANC recovery at day 28, 60, and 90 was 80% (95%CI, 73-87), 86% (95%CI, 80-92) and 89% (95%CI, 83-94), respectively. The probability of platelet recovery on the day of CAR-T cell infusion was 55% (95%CI, 46-64); rising to 74% (95%CI, 67-82), 83% (95%CI, 76-90), and 84% (95%CI, 77-90) at day 28, 60, and 90, respectively. A competing event was always observed in patients without ANC or Plt recovery. In multivariable analysis, higher pre-lymphodepletion Plt count (HR=1.08 per 25 x 109/L increase, p=0.006) and higher peak CD8+ CAR-T cells in blood (HR=1.47 per log10 cells/µL increase, p<0.001) were associated with faster ANC recovery. ALL diagnosis and higher cytokine release syndrome (CRS) grade were associated with slower ANC recovery (CLL vs ALL, HR=1.60, p=0.02; NHL vs ALL, HR=2.07, p=0.007). Higher CRS grade was also associated with slower Plt recovery (HR=0.67 per grade increase, p<0.001). Higher pre-lymphodepletion platelet count and higher peak CD8+ CAR-T cell in blood were associated with faster platelet recovery (HR=1.08 per 25 x 109/L increase, p=0.001; HR=1.41 per log10 cells/µL increase, p<0.001). Of note, lymphodepletion intensity did not seem to affect hematopoietic recovery. Table 2 summarizes the results of the multivariable analysis. Figure 1 shows ANC and Plt recovery by CRS grade. Conclusion We identified CRS grade as independently associated with impaired hematopoietic recovery after CD19 CAR-T cell therapy. Our findings suggest that the prevention of CRS may improve hematopoietic recovery after CD19 CAR-T cell therapy. Figure Disclosures Hirayama: DAVA Oncology: Honoraria. Maloney:Celgene,Kite Pharma: Honoraria, Research Funding; BioLine RX, Gilead,Genentech,Novartis: Honoraria; Juno Therapeutics: Honoraria, Patents & Royalties: patients pending , Research Funding; A2 Biotherapeutics: Honoraria, Other: Stock options . Turtle:Caribou Biosciences: Equity Ownership, Membership on an entity's Board of Directors or advisory committees; Nektar Therapeutics: Other: Ad hoc advisory board member, Research Funding; Allogene: Other: Ad hoc advisory board member; Novartis: Other: Ad hoc advisory board member; Juno Therapeutics: Patents & Royalties: Co-inventor with staff from Juno Therapeutics; pending, Research Funding; Precision Biosciences: Equity Ownership, Membership on an entity's Board of Directors or advisory committees; Eureka Therapeutics: Equity Ownership, Membership on an entity's Board of Directors or advisory committees; T-CURX: Membership on an entity's Board of Directors or advisory committees; Kite/Gilead: Other: Ad hoc advisory board member; Humanigen: Other: Ad hoc advisory board member.

Robust methods in Mendelian randomization via penalization of heterogeneous causal estimates.

Methods have been developed for Mendelian randomization that can obtain consistent causal estimates under weaker assumptions than the standard instrumental variable assumptions. The median-based estimator and MR-Egger are examples of such methods. However, these methods can be sensitive to genetic variants with heterogeneous causal estimates. Such heterogeneity may arise from over-dispersion in the causal estimates, or specific variants with outlying causal estimates. In this paper, we develop three extensions to robust methods for Mendelian randomization with summarized data: 1) robust regression (MM-estimation); 2) penalized weights; and 3) Lasso penalization. Methods using these approaches are considered in two applied examples: one where there is evidence of over-dispersion in the causal estimates (the causal effect of body mass index on schizophrenia risk), and the other containing outliers (the causal effect of low-density lipoprotein cholesterol on Alzheimer’s disease risk). Through an extensive simulation study, we demonstrate that robust regression applied to the inverse-variance weighted method with penalized weights is a worthwhile additional sensitivity analysis for Mendelian randomization to provide robustness to variants with outlying causal estimates. The results from the applied examples and simulation study highlight the importance of using methods that make different assumptions to assess the robustness of findings from Mendelian randomization investigations with multiple genetic variants.

Predicting Overall Survival in Patients with Metastatic Rectal Cancer: a Machine Learning Approach

BackgroundA significant proportion of patients with rectal cancer will present with synchronous metastasis at the time of diagnosis. Overall survival (OS) for these patients are highly variable and previous attempts to build predictive models often have low predictive power, with concordance indexes (c-index) less than 0.70. MethodsUsing the National Cancer Database (2010–2014), we identified patients with synchronous metastatic rectal cancer. The data was split into a training dataset (diagnosis years 2010–2012), which was used to build the machine learning model, and a testing dataset (diagnosis years 2013–2014), which was used to externally validate the model. A nomogram predicting 3-year OS was created using Cox proportional hazard regression with lasso penalization. Predictors were selected based on clinical significance and availability in NCDB. Performance of the machine learning model was assessed by c-index. ResultsA total of 4098 and 3107 patients were used to construct and validate the nomogram, respectively. Internally validated c-indexes at 1, 2, and 3 years were 0.816 (95% CI 0.813–0.818), 0.789 (95% CI 0.786–0.790), and 0.778 (95% CI 0.775–0.780), respectively. External validated c-indexes at 1, 2, and 3 years were 0.811, 0.779, and 0.778, respectively. ConclusionsThere is wide variability in the OS for patients with metastatic rectal cancer, making accurate predictions difficult. However, using machine learning techniques, more accurate models can be built. This will aid patients and clinicians in setting expectations and making clinical decisions in this group of challenging patients.

Accounting for covariate distributions in slope-unit-based landslide susceptibility models. A case study in the alpine environment

Thousands or even million of pixels can be contained in a single Slope Unit. Hence, each covariate used in spatial predictive models is characterized by a distribution of values for each Slope Unit.Here, we model the whole covariates' distribution within Slope Units for landslide susceptibility purposes. This is done by finely dissecting each covariate into quantiles and then modeling the susceptibility via a LASSO penalized Binary Logistic Regression. We choose a LASSO penalization because the common Stepwise procedure is not selective enough to shrink a large number of covariates to an interpretable subset (which we also demonstrate here). LASSO mostly selects 6 covariates out of 372 to explain the spatial distribution of shallow landslides in the Upper Badia valley, Italian Alps. This allows us to verify that the selection does not include any quantile close to the median hence, nor to the mean. The latter is the common representation of the covariates' distribution within Slope Units, which we also test and report in the supplements. Overall, we suggest to always investigate the whole distribution because the mean may not be the most informative nor the most performing way to generate Slope-Unit-based susceptibility models. In this general context, we generate our landslide inventory by combining semi-automated (OBIA) and manual mapping procedures. Our inventory, quantile covariates' representation and LASSO penalization produce excellent performances and interpretable relations between covariates and landslides.

Time-dependent Poisson reduced rank models for political text data analysis

We consider Poisson reduced rank models where parameters depend on time. Our main motivation comes from studies in comparative politics where one wants to locate party positions in a certain political space. For this purpose, several empirical methods have been proposed using text as data sources. As the data structure of texts is quite complex, its analysis to extract information is generally a difficult task. Furthermore, political texts usually contain a large number of words such that a simultaneous analysis of word counts becomes challenging. In this paper, we consider Poisson models for each word count simultaneously and provide a statistical method suitable to analyze political text data. We consider a novel model which allows the political lexicon to change over time and develop an estimation procedure based on LASSO and fused LASSO penalization techniques to address high-dimensionality via significant dimension reduction. This model gives insights into the potentially changing use of words by left and right-wing parties over time. The procedure allows to identify automatically those words having a discriminating effect between party positions. To address the dependence structure of word counts over time, we propose integer-valued time series processes to implement a suitable bootstrap method for constructing confidence intervals for the model parameters. We apply our approach to party manifesto data from German parties over seven federal elections after German reunification. The approach does not require any a priori information nor expert knowledge to process the data.

Explanation of somatic symptoms by mental health and personality traits: application of Bayesian regularized quantile regression in a large population study

BackgroundSomatic syndrome is one of the remarkably prevalent issues in primary health care and subspecialty settings. We aimed to elucidate multidimensional associations between somatic symptoms with major mental problems and personality traits in the framework of the quantile regression model with a Bayesian approach.MethodsA total of 4763 employees at Isfahan University of Medical Sciences and Health Services in Isfahan province, Iran, filled out four validated questionnaires including Hospital Anxiety and Depression Scale (HADS), NEO Questionnaire, General Health Questionnaire (GHQ) and PHQ-15 for somatic symptom severity. In addition, Functional Gastrointestinal Disorders (FGIDs) were determined using Rome IV criteria. Exploratory Factor Analysis (EFA) and Bayesian regularized quantile regression with adaptive LASSO penalization were applied for reduced dimension of somatic symptoms and variable selection and parameter estimation, respectively.ResultsThe 25 major somatic symptoms were grouped into four factors including general, upper gastrointestinal, lower gastrointestinal and respiratory by EFA. Stress, depression, and anxiety had significant effects on all of the four extracted factors. The effect of anxiety in each four extracted factors was more than stress and depression. Neuroticism and agreeableness had significant effects on all of the four extracted factors, generally (p < 0.05).ConclusionsGiven the high prevalence of somatic symptoms and psychosomatic complaints in correlation with the diverse range of mental co-morbidities, developing more detailed diagnostic tools and methods is crucial; nonetheless, it seems that providing better interdisciplinary approaches in general medical practice is groundwork.

Abstract 668: A recommender system to enhance drug development in oncology from phase I trials: The RESOLVED2 project

Abstract The industry and investigators’ communities are currently facing a conjunction of increasing number of anti-neoplastic agents (ANA) candidate for phase 1 clinical trials (P1CT) and an important attrition rate for final approval. Improvement of the drug development strategies in oncology could be achieved by machine learning supporting go/no-go decisions after phase I trial. PubMed abstracts of P1CT reporting ANA were used together with pharmacological classification and molecular targets of ANA from DrugBank5.0 database. To account for heterogeneous follow-up we computed a new metric entitled Food &amp; Drug Administration approval-free-survival (FDA-aFS). FDA-aFS is a right-censored variable defined by the time between first P1CT publication and the FDA approval date or censoring for compounds not approved up to 2018, July. A training set and a test set were randomly generated from the whole dataset with a 70%/30% ratio. A regression and a classification version of RESOLVED2 have been developed in the training set and evaluated independently on the test set. The RESOLVED2 regression model has been optimized using a Cox logistic regression on FDA-aFS, penalized by Lasso with cross validation on the training set. RESOLVED2 predictions have been compared to a model including common variables used for go/no-go decisions (toxicity, clinical activity, observation of complete tumor response). The RESOLVED2 classification model (predicted approved ANA vs predicted non-approved ANA) was computed in the training set using predicted probabilities derived from the regression model, by the identification of the cut-off reaching minimal log-rank derived p-value when comparing FDA-aFS of predicted approved versus predicted non-approved ANA. We identified 462 ANA matching with DrugBank entries (P1CT publication interval: 1972-2017). Among 1411 variables, 28 remained in the model after Lasso penalization. The relation of the model to the FDA-aFS had a weighted concordance index of 0.89 on the independent test set, versus 0.69 for a classical model based on efficacy/toxicity. Classified approval predictions achieved a p&amp;lt;10e-10 for association with observed FDA-aFS in the test set: at 6 years of follow-up, 73% (95%CI [49%; 86%]) of predicted approved drugs were indeed approved and 92% (95%CI [87%; 98%]) of predicted non-approved drugs were still not approved. A predicted approved drug was 16 times more likely to be approved compared to a predicted non-approved drug (Hazard ratio=16.4; 95%CI [8.40; 32.2]; p&amp;lt;10e-10). RESOLVED2 is an applicable machine learning model that could support early go/no-go decision as soon as P1CT completion based on accurate prediction of FDA approval for drugs in oncology. Citation Format: Guillaume Beinse, Virgile Tellier, Valentin Charvet, Isabelle Borget, Christophe Massard, Antoine Hollebecque, Loic Verlingue. A recommender system to enhance drug development in oncology from phase I trials: The RESOLVED2 project [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 668.

Dynamic semi-parametric factor model for functional expectiles

High-frequency data can provide us with a quantity of information for forecasting and help to calculate and prevent the future risk based on extremes. This tail behaviour is very often driven by exogenous components and may be modelled conditionally on other variables. However, many of these phenomena are observed over time, exhibiting non-trivial dynamics and dependencies. We propose a functional dynamic factor model to study the dynamics of expectile curves. The complexity of the model and the number of dependent variables are reduced by lasso penalization. The functional factors serve as a low-dimensional representation of the conditional tail event, while the time-variation is captured by factor loadings. We illustrate the model with an application to climatology, where daily data over years on temperature, rainfalls or strength of wind are available.

D-Trace estimation of a precision matrix with eigenvalue control

The estimation of a precision matrix has an important role in several research fields. In high dimensional settings, one of the most prominent approaches to estimate the precision matrix is the Lasso norm penalized convex optimization. This framework guarantees the sparsity of the estimated precision matrix. However, it does not control the eigenspectrum of the obtained estimator. Moreover, Lasso penalization shrinks the largest eigenvalues of the estimated precision matrix. In this article, we focus on D-trace estimation methodology of a precision matrix. We propose imposing a negative trace penalization on the objective function of the D-trace approach, aimed to control the eigenvalues of the estimated precision matrix. Through extensive numerical analysis, using simulated and real datasets, we show the advantageous performance of our proposed methodology.

Machine learning for predictive analyses in health: an example of an application to predict death in the elderly in São Paulo, Brazil

This study aims to present the stages related to the use of machine learning algorithms for predictive analyses in health. An application was performed in a database of elderly residents in the city of São Paulo, Brazil, who participated in the Health, Well-Being, and Aging Study (SABE) (n = 2,808). The outcome variable was the occurrence of death within five years of the elder's entry into the study (n = 423), and the predictors were 37 variables related to the elder's demographic, socioeconomic, and health profile. The application was organized according to the following stages: division of data in training (70%) and testing (30%), pre-processing of the predictors, learning, and assessment of the models. The learning stage used 5 algorithms to adjust the models: logistic regression with and without penalization, neural networks, gradient boosted trees, and random forest. The algorithms' hyperparameters were optimized by 10-fold cross-validation to select those corresponding to the best models. For each algorithm, the best model was assessed in test data via area under the ROC curve (AUC) and related measures. All the models presented AUC ROC greater than 0.70. For the three models with the highest AUC ROC (neural networks and logistic regression with LASSO penalization and without penalization, respectively), quality measures of the predicted probability were also assessed. The expectation is that with the increased availability of data and trained human capital, it will be possible to develop predictive machine learning models with the potential to help health professionals make the best decisions.

Bayesian MIDAS Penalized Regressions: Estimation, Selection, and Prediction

We propose a new approach to mixed-frequency regressions in a high-dimensional environment that resorts to Group Lasso penalization and Bayesian techniques for estimation and inference. To improve the sparse recovery ability of the model, we also consider a Group Lasso with a spike-and-slab prior. Penalty hyper-parameters governing the model shrinkage are automatically tuned via an adaptive MCMC algorithm. Simulations show that the proposed models have good selection and forecasting performance, even when the design matrix presents high cross-correlation. When applied to U.S. GDP data, the results suggest that financial variables may have some, although limited, short-term predictive content.

Analyzing Large Workers’ Compensation Claims Using Generalized Linear Models and Monte Carlo Simulation

Insurance practitioners rely on statistical models to predict future claims in order to provide financial protection. Proper predictive statistical modeling is more challenging when analyzing claims with lower frequency, but high costs. The paper investigated the use of predictive generalized linear models (GLMs) to address this challenge. Workers’ compensation claims with costs equal to or more than US$100,000 were analyzed in agribusiness industries in the Midwest of the USA from 2008 to 2016. Predictive GLMs were built with gamma, Weibull, and lognormal distributions using the lasso penalization method. Monte Carlo simulation models were developed to check the performance of predictive models in cost estimation. The results show that the GLM with gamma distribution has the highest predictivity power (R2 = 0.79). Injury characteristics and worker’s occupation were predictive of large claims’ occurrence and costs. The conclusions of this study are useful in modifying and estimating insurance pricing within high-risk agribusiness industries. The approach of this study can be used as a framework to forecast workers’ compensation claims amounts with rare, high-cost events in other industries. This work is useful for insurance practitioners concerned with statistical and predictive modeling in financial risk analysis.

Dynamic Semi-Parametric Factor Model for Functional Expectiles

High-frequency data can provide us with a quantity of information for forecasting, help to calculate and prevent the future risk based on extremes. This tail behaviour is very often driven by exogenous components and may be modelled conditional on other variables. However, many of these phenomena are observed over time, exhibiting non-trivial dynamics and dependencies. We propose a functional dynamic factor model to study the dynamics of expectile curves. The complexity of the model and the number of dependent variables are reduced by lasso penalization. The functional factors serve as a low-dimensional representation of the conditional tail event, while the time-variation is captured by factor loadings. We illustrate the model with an application to climatology, where daily data over years on temperature, rainfalls or strength of wind are available.

Multinomial functional regression with wavelets and LASSO penalization

A classification problem with a functional predictor is studied, and it is suggested to use a multinomial functional regression (MFR) model for the analysis. The discrete wavelet transform and LASSO penalization are combined for estimation, and the fitted model is used for classification of new curves with unknown class membership. The MFR approach is applied to two datasets, one regarding lameness detection for horses and another regarding speech recognition. In the applications, as well as in a simulation study, the performance of the MFR approach is compared to that of other methods for supervised classification of functional data, and MFR performs as well or better than the other methods.

Uncertainty quantification for functional dependent random variables

This paper proposes a new methodology to model uncertainties associated with functional random variables. This methodology allows to deal simultaneously with several dependent functional variables and to address the specific case where these variables are linked to a vectorial variable, called covariate. In this case, the proposed uncertainty modelling methodology has two objectives: to retain both the most important features of the functional variables and their features which are the most correlated to the covariate. This methodology is composed of two steps. First, the functional variables are decomposed on a functional basis. To deal simultaneously with several dependent functional variables, a Simultaneous Partial Least Squares algorithm is proposed to estimate this basis. Second, the joint probability density function of the coefficients selected in the decomposition is modelled by a Gaussian mixture model. A new sparse method based on a Lasso penalization algorithm is proposed to estimate the Gaussian mixture model parameters and reduce their number. Several criteria are introduced to assess the methodology performance: its ability to approximate the functional variables probability distribution, their dependence structure and their features which explain the covariate. Finally, the whole methodology is applied on a simulated example and on a nuclear reliability test case.

Air pollution and temperature on associations with DNA methylation in a mother-child cohort: an epigenome-wide analysis using penalization methods.

Introduction: Epidemiological studies have suggested that air pollution, and possibly temperature, are associated with birth weight. Changes in DNA methylation are possible mediators of part of this association, a question so far little considered using shrinkage statistical methods. Throughout in utero development, the placenta plays a key role in controlling growth and development. We studied the association between air pollution and temperature with placental DNA methylation during various pregnancy exposure windows. Methods: Placenta samples were collected at delivery in 668 women from a French mother-child cohort. DNA methylation was assayed using the HumanMethylation450 Illumina BeadChip (450K). Average NO2 and PM10 levels over the whole pregnancy were estimated at the home address by a dispersion model. Temperature was daily recorded by meteorological stations located across the entire French territory. We assessed whether these environmental exposures are related to 426,049 CpGs using LASSO penalization (shrinkage) method. Results: PM10 and NO2 concentrations were associated with methylation of 2 genes (CCNA1 and ADARB2) and 3 genes (ADORA2B, ADARB2 and EPB41L2), respectively. PM10 levels were associated with a decrease in DNA methylation on CCNA1 and ADARB2. NO2 levels were associated with a decrease in DNA methylation on ADORA2B and ADARB2 gene but with an increase in DNA methylation on EPB41L2. For example, ADORA2B gene codes for adenosine receptor A2B, which has a role in placental development and possibly in the pathophysiology of preeclampsia. Conclusion: We found evidence of associations between air pollution and DNA methylation in pregnant women. These identified genes may reflect biological of air pollution and should be examined in future studies.

Consistent model identification of varying coefficient quantile regression with BIC tuning parameter selection

ABSTRACTQuantile regression provides a flexible platform for evaluating covariate effects on different segments of the conditional distribution of response. As the effects of covariates may change with quantile level, contemporaneously examining a spectrum of quantiles is expected to have a better capacity to identify variables with either partial or full effects on the response distribution, as compared to focusing on a single quantile. Under this motivation, we study a general adaptively weighted LASSO penalization strategy in the quantile regression setting, where a continuum of quantile index is considered and coefficients are allowed to vary with quantile index. We establish the oracle properties of the resulting estimator of coefficient function. Furthermore, we formally investigate a Bayesian information criterion (BIC)-type uniform tuning parameter selector and show that it can ensure consistent model selection. Our numerical studies confirm the theoretical findings and illustrate an application of the new variable selection procedure.

New improvements in the use of dependence measures for sensitivity analysis and screening

ABSTRACTPhysical phenomena are commonly modelled by time consuming numerical simulators, function of many uncertain parameters whose influences can be measured via a global sensitivity analysis. The usual variance-based indices require too many simulations, especially as the inputs are numerous. To address this limitation, we consider recent advances in dependence measures, focusing on the distance correlation and the Hilbert–Schmidt independence criterion. We study and use these indices for a screening purpose. Numerical tests reveal differences between variance-based indices and dependence measures. Then, two approaches are proposed to use the latter for a screening purpose. The first approach uses independence tests, with existing asymptotic versions and spectral extensions; bootstrap versions are also proposed. The second considers a linear model with dependence measures, coupled to a bootstrap selection method or a Lasso penalization. Numerical experiments show their potential in the presence of many non-influential inputs and give successful results for a nuclear reliability application.