2,398 research outputs found
Fast identification of biological pathways associated with a quantitative trait using group lasso with overlaps.
No full textWhere causal SNPs (single nucleotide polymorphisms) tend to accumulate within biological pathways, the incorporation of prior pathways information into a statistical model is expected to increase the power to detect true associations in a genetic association study. Most existing pathways-based methods rely on marginal SNP statistics and do not fully exploit the dependence patterns among SNPs within pathways.We use a sparse regression model, with SNPs grouped into pathways, to identify causal pathways associated with a quantitative trait. Notable features of our "pathways group lasso with adaptive weights" (P-GLAW) algorithm include the incorporation of all pathways in a single regression model, an adaptive pathway weighting procedure that accounts for factors biasing pathway selection, and the use of a bootstrap sampling procedure for the ranking of important pathways. P-GLAW takes account of the presence of overlapping pathways and uses a novel combination of techniques to optimise model estimation, making it fast to run, even on whole genome datasets.In a comparison study with an alternative pathways method based on univariate SNP statistics, our method demonstrates high sensitivity and specificity for the detection of important pathways, showing the greatest relative gains in performance where marginal SNP effect sizes are small
A comparative study of the Lasso-type and heuristic model selection methods
No full textThis study presents a first comparative analysis of Lasso-type (Lasso, adaptive Lasso, elastic net) and heuristic subset selection methods. Although the Lasso has shown success in many situations, it has some limitations. In particular, inconsistent results are obtained for pairwise strongly correlated predictors. An alternative to the Lasso is constituted by model selection based on information criteria (IC), which remains consistent in the situation mentioned. However, these criteria are hard to optimize due to a discrete search space. To overcome this problem, an optimization heuristic (Genetic Algorithm) is applied. Monte-Carlo simulation results are reported to illustrate the performance of the methods.Model selection, Lasso, adaptive Lasso, elastic net, heuristic methods, genetic algorithms
Gene regulatory networks from multifactorial perturbations using graphical lasso: Application to the DREAM4 challenge
No full textA major challenge in the field of systems biology consists of predicting gene regulatory networks based on different training data. Within the DREAM4 initiative, we took part in the multifactorial sub-challenge that aimed to predict gene regulatory networks of size 100 from training data consisting of steady-state levels obtained after applying multifactorial perturbations to the original in silico network. Due to the static character of the challenge data, we tackled the problem via a sparse Gaussian Markov Random Field, which relates network topology with the covariance inverse generated by the gene measurements. As for the computations, we used the Graphical Lasso algorithm which provided a large range of candidate network topologies. The main task was to select the optimal network topology and for that, different model selection criteria were explored. The selected networks were compared with the golden standards and the results ranked using the scoring metrics applied in the challenge, giving a better insight in our submission and the way to improve it.Our approach provides an easy statistical and computational framework to infer gene regulatory networks that is suitable for large networks, even if the number of the observations (perturbations) is greater than the number of variables (genes
Post-l1-penalized estimators in high-dimensional linear regression models
No full textIn this paper we study post-penalized estimators which apply ordinary, unpenalized linear regression to the model selected by first-step penalized estimators, typically LASSO. It is well known that LASSO can estimate the regression function at nearly the oracle rate, and is thus hard to improve upon. We show that post-LASSO performs at least as well as LASSO in terms of the rate of convergence, and has the advantage of a smaller bias. Remarkably, this performance occurs even if the LASSO-based model selection 'fails' in the sense of missing some components of the 'true' regression model. By the 'true' model we mean here the best s-dimensional approximation to the regression function chosen by the oracle. Furthermore, post-LASSO can perform strictly better than LASSO, in the sense of a strictly faster rate of convergence, if the LASSO-based model selection correctly includes all components of the 'true' model as a subset and also achieves a sufficient sparsity. In the extreme case, when LASSO perfectly selects the 'true' model, the post-LASSO estimator becomes the oracle estimator. An important ingredient in our analysis is a new sparsity bound on the dimension of the model selected by LASSO which guarantees that this dimension is at most of the same order as the dimension of the 'true' model. Our rate results are non-asymptotic and hold in both parametric and nonparametric models. Moreover, our analysis is not limited to the LASSO estimator in the first step, but also applies to other estimators, for example, the trimmed LASSO, Dantzig selector, or any other estimator with good rates and good sparsity. Our analysis covers both traditional trimming and a new practical, completely data-driven trimming scheme that induces maximal sparsity subject to maintaining a certain goodness-of-fit. The latter scheme has theoretical guarantees similar to those of LASSO or post-LASSO, but it dominates these procedures as well as traditional trimming in a wide variety of experiments.
Flame pyrolysis prepared catalysts for the steam reforming of ethanol
No full textIntroduction
Among the methods to produce H2, Steam Reforming (SR) is one of the most common and feasible to use [1-7]. One challenge for SR at high temperature is catalyst deactivation by sintering, so that high thermal resistance is a pressing need. By contrast, it is envisaged to operate at lower temperature, to lower the heat input to the reactor. Nevertheless, catalyst deactivation may be impressive by coking, due to the formation of carbon filaments and occurs mainly over big nickel particles [8]. Additional coking may occur over acidic sites of the support. Therefore, an appropriate catalyst formulation should be found, which allows to reach the highest catalytic performance at low operating temperature (i.e. ca. 500°C) together with proper resistance.
The aim of this work is the synthesis of catalysts for the SR of ethanol. A set of catalysts was synthesized by Flame Spray Pyrolysis (FP) and another set was prepared by impregnation of the active phase on the FP-prepared support. This high temperature synthesis was here adopted to impart suitable thermal resistance to the samples and to provide a high metal support interaction, which showed a pivotal importance to improve resistance towards coking.
2 Experimental/methodology
A first set of samples was directly prepared by FP, inserting in one-step the Ni-based active phase into the selected support, TiO2 and La2O3 with different metal loading of 5, 10 and 15 wt%. For comparison, the same formulations have been prepared by impregnation of Ni on the FP prepared supports. The catalysts were reduced for 1h at 800°C in a 20 vol% H2 / N2 gas mixture.
The samples characterization was carried out by conventional methods of XRD, BET, SEM, EDX, TEM and TPR.
Finally the activity test of the samples were performed by means of a micro pilot plant constituted by an Incoloy 800 continuous down flow reactor heated by an electric oven. Catalyst activation was accomplished by feeding 50 cm3/min of a 20 vol% H2/N2 gas mixture, while heating by 10 °C/min up to 800 °C, then kept for 1 hour. For activity testing 0.017 cm3/min of a 3:1 (mol/mol) H2O:CH3CH2OH liquid mixture were feed to the reactor by means of a HPLC pump (Waters, mod. 501). The activity tests were carried out at atmospheric pressure, GHSV=2.500 h-1 (referred to the ethanol + water gaseous mixture) at 500, 625 and 750 °C. The analysis of the out-flowing gas was carried out by a gas chromatograph (Agilent, mod. 7980) ca. 8h.
3 Results and discussion
Nickel-based catalysts at three different loadings (5, 10 and 15wt%), supported over lanthana and titania were synthesized and tested for ethanol steam reforming at 500 and 750 °C. All of them were more active and stable at the latter temperature while at the former the impregnated catalysts with low Ni loading exhibited low H2 productivity, mainly due to unreformed CH4. By contrast, the FP ones demonstrated superior catalytic activity and satisfactory stability, especially with lanthana support, which effectively reduced deactivation by coking at the lowest operating temperature.
The catalytic activity has been correlated to metal dispersion and to the metal-support interaction strength. Both parameters affected also catalyst resistance to coking at 500°C. Overall, lanthana demonstrated and interesting support due to its basic character, which prevented significant coke formation related to the acidic properties of the support. Furthermore, high metal dispersion and proper stabilization on the support allowed to limit the formation of carbon nanofilaments as deactivation mode.
References
[1] Ni, M., Leung, D.Y.C., Leung, M.K.H., Int. J. Hydrogen Energy, 2007, 32, 3238.
[2] Muroyama, H., Nakase, R., Matsui, T., Eguchi, Eguchi, K., Int. J. Hydrogen Energy 2010, 35, 1575.
[3] Llorca, J., Homs, N., Sales, J., De la Piscina, P.R., J. Catal. 2002, 209, 306.
[4] Fatsikostas, A.N., Verykios, X.E., J. Catal. 2004, 225, 439.
[5] Cheekatamarla, P.K., Finnerty, C.M., J. Power Sources 2006, 160, 490.
[6] Frusteri, F., Freni, S., Spadaro, L., Chiodo, V., Bonura, Donato, S., Catal. Commun. 2004, 5, 611.
[7] Wang, C.B., Lee, C.C., Bi, J.L., Siang, Liu, J.Y., Yeh, C.T., Catal. Today 2009, 146, 76.
[8] Yoshida, H. Yamaoka, R., Arci, M., J. Mol. Science, 2015, 16, 350-362
Direct effects testing : a two-stage procedure to test for effect size and variable importance for correlated binary predictors and a binary response.
No full textIn applications such as medical statistics and genetics, we encounter situations where a large number of highly correlated predictors explain a response. For example, the response may be a disease indicator and the predictors may be treatment indicators or single nucleotide polymorphisms (SNPs). Constructing a good predictive model in such cases is well studied. Less well understood is how to recover the 'true sparsity pattern', that is finding which predictors have direct effects on the response, and indicating the statistical significance of the results. Restricting attention to binary predictors and response, we study the recovery of the true sparsity pattern using a two-stage method that separates establishing the presence of effects from inferring their exact relationship with the predictors. Simulations and a real data application demonstrate that the method discriminates well between associations and direct effects. Comparisons with lasso-based methods demonstrate favourable performance of the proposed method. Copyright © 2010 John Wiley & Sons, Ltd
Feature Selection Guided by Structural Information
No full textIn generalized linear regression problems with an abundant number of features, lasso-type regularization which imposes an l1-constraint on the regression coefficients has become a widely established technique. Crucial deficiencies of the lasso were unmasked when Zhou and Hastie (2005) introduced the elastic net. In this paper, we propose to extend the elastic net by admitting general nonnegative quadratic constraints as second form of regularization. The generalized ridge-type constraint will typically make use of the known association structure of features, e.g. by using temporal- or spatial closeness.
We study properties of the resulting ’structured elastic net’ regression estimation procedure, including basic asymptotics and the issue of model selection consistency. In
this vein, we provide an analog to the so-called ’irrepresentable condition’ which holds for the lasso. An oracle property is established by incorporating a scaled l1-constraint. Moreover, we outline algorithmic solutions for the structured elastic net within the generalized linear model family. The rationale and the performance of our approach is illustrated by means of simulated- and real world data
Penalized flexible Bayesian quantile regression
No full textCopyright © 2012 SciResThis article has been made available through the Brunel Open Access Publishing Fund.The selection of predictors plays a crucial role in building a multiple regression model. Indeed, the choice of a suitable subset of predictors can help to improve prediction accuracy and interpretation. In this paper, we propose a flexible Bayesian Lasso and adaptive Lasso quantile regression by introducing a hierarchical model framework approach to en- able exact inference and shrinkage of an unimportant coefficient to zero. The error distribution is assumed to be an infi- nite mixture of Gaussian densities. We have theoretically investigated and numerically compared our proposed methods with Flexible Bayesian quantile regression (FBQR), Lasso quantile regression (LQR) and quantile regression (QR) methods. Simulations and real data studies are conducted under different settings to assess the performance of the pro- posed methods. The proposed methods perform well in comparison to the other methods in terms of median mean squared error, mean and variance of the absolute correlation criterions. We believe that the proposed methods are useful practically
Regularized estimation of large-scale gene association networks using graphical Gaussian models
No full textGraphical Gaussian models are popular tools for the estimation of (undirected) gene association networks from microarray data. A key issue when the number of variables greatly exceeds the number of samples is the estimation of the matrix of partial correlations. Since the (Moore-Penrose) inverse of the sample covariance matrix leads to poor estimates in this scenario, standard methods are inappropriate and adequate regularization techniques are needed. Popular approaches include biased estimates of the covariance matrix and high-dimensional regression schemes, such as the Lasso and Partial Least Squares.
In this article, we investigate a general framework for combining regularized regression methods with the estimation of Graphical Gaussian models. This framework includes various existing methods as well as two new approaches based on ridge regression and adaptive lasso, respectively. These methods are extensively compared both qualitatively and quantitatively within a simulation study and through an application to six diverse real data sets. In addition, all proposed algorithms are implemented in the R package "parcor", available from the R repository CRAN.
In our simulation studies, the investigated non-sparse regression methods, i.e. Ridge Regression and Partial Least Squares, exhibit rather conservative behavior when combined with (local) false discovery rate multiple testing in order to decide whether or not an edge is present in the network.
We confirm the Lasso's well known tendency towards selecting too many edges, whereas the two-stage adaptive Lasso is an interesting alternative that provides sparser solutions.
On six real data sets, we also clearly distinguish the results obtained using the non-sparse methods and those obtained using the sparse methods where specification of the regularization parameter automatically means model selection. Furthermore, for data that violates the assumption of uncorrelated observations (due to replications), the Lasso and the adaptive Lasso yield very complex structures, indicating that they might not be suited under these conditions
On the Choice of Prior in Bayesian Model Averaging
No full textBayesian model averaging attempts to combine parameter estimation and model uncertainty in one coherent framework. The choice of prior is then critical. Within an explicit framework of ignorance we define a ‘suitable’ prior as one which leads to a continuous and suitable analog to the pretest estimator. The normal prior, used in standard Bayesian model averaging, is shown to be unsuitable. The Laplace (or lasso) prior is almost suitable. A suitable prior (the Subbotin prior) is proposed and its properties are investigated.Model averaging;Bayesian analysis;Subbotin prior
- …
