ThrEEBoost: Thresholded Boosting for Variable Selection and Prediction via Estimating Equations

Most variable selection techniques for high-dimensional models are designed to be used in settings, where observations are independent and completely observed. At the same time, there is a rich literature on approaches to estimation of low-dimensional parameters in the presence of correlation, missingness, measurement error, selection bias, and other characteristics of real data. In this article, we present ThrEEBoost (Thresholded EEBoost), a general-purpose variable selection technique which can accommodate such problem characteristics by replacing the gradient of the loss by an estimating function. ThrEEBoost generalizes the previously proposed EEBoost algorithm (Wolfson 2011 Wolfson, J. (2011), “EEBoost: A General Method for Prediction and Variable Selection Based on Estimating Equations,” Journal of the American Statistical Association, 106, 296–305.[Taylor & Francis Online], [Web of Science ®] , [Google Scholar]) by allowing the number of regression coefficients updated at each step to be controlled by a thresholding parameter. Different thresholding parameter values yield different variable selection paths, greatly diversifying the set of models that can be explored; the optimal degree of thresholding can be chosen by cross-validation. ThrEEBoost was evaluated using simulation studies to assess the effects of different threshold values on prediction error, sensitivity, specificity, and the number of iterations to identify minimum prediction error under both sparse and nonsparse true models with correlated continuous outcomes. We show that when the true model is sparse, ThrEEBoost achieves similar prediction error to EEBoost while requiring fewer iterations to locate the set of coefficients yielding the minimum error. When the true model is less sparse, ThrEEBoost has lower prediction error than EEBoost and also finds the point yielding the minimum error more quickly. The technique is illustrated by applying it to the problem of identifying predictors of weight change in a longitudinal nutrition study. Supplementary materials are available online.     

Efficient and accurate two-scale simulation of non-linear heterogeneous microstructures

This work is concerned with the development of an efficient two-scale numerical scheme for the prediction of the local and effective mechanical response of heterogeneous materials with non-linear constitutive behavior. In order to ensure both, accurate micromechanical fields and feasible overall CPU-times, an efficient but rather simple solution scheme is proposed. Its capabilities are demonstrated using the FE-FFT method [2,3] for the solution of simple two-scale problems. As an example, two-dimensional polycrystalline microstructures with an elasto-viscoplastic constitutive law are considered. (© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Anionic surfactants and surfactant ionic liquids with quaternary ammonium counterions

Small-angle neutron scattering and surface tension have been used to characterize a class of surfactants (SURFs), including surfactant ionic liquids (SAILs). These SURFs and SAILs are based on organic surfactant anions (single-tail dodecyl sulfate, DS, double-chain aerosol-OT, AOT, and the trichain, TC) with substituted quaternary ammonium cations. This class of surfactants can be obtained by straightforward chemistry, being cheaper and more environmentally benign than standard cationic SAILs. A surprising aspect of the results is that, broadly speaking, the physicochemical properties of these SURFs and SAILs are dominated by the nature of the surfactant anion and that the chemical structure of the added cation plays only a secondary role.     

Photoreactive surfactants: a facile and clean route to oxide and metal nanoparticles in reverse micelles

A new class of photoreactive surfactants (PRSs) is presented here, consisting of amphiphiles that can also act as reagents in photochemical reactions. An example PRS is cobalt 2-ethylhexanoate (Co(EH)(2)), which forms reverse micelles (RMs) in a hydrocarbon solvent, as well as mixed reversed micelles with the standard surfactant Aerosol-OT (AOT). Small-angle neutron scattering (SANS) data show that mixed AOT/PRS RMs have a spherical structure and size similar to that of pure AOT micelles. Excitation of the ligand-to-metal charge transfer (LMCT) band in the PRSs promotes electron transfer from PRS to associated metal counterions, leading to the generation of metal and metal-oxide nanoparticles inside the RMs. This work presents proof of concept for employing PRSs as precursors to obtain nearly monodisperse inorganic nanoparticles: here both Co(3)O(4) and Bi nanoparticles have been synthesized at high metal concentration (10(-2) M) by simply irradiating the RMs. These results point toward a new approach of photoreactive self-assembly, which represents a clean and straightforward route to the generation of nanomaterials.     

素食人群尿液1H NMR代谢轮廓的多因素方差分析

结合方差分析(ANOVA)和偏最小二乘法判别分析(PLS-DA)两种分析技术,对素食和普食人群的尿液1H NMR谱进行分析.利用ANOVA方法将数据矩阵分解为几个独立因素矩阵,滤除干扰因素后,再利用PLS-DA对单因素数据进行建模分析.实验结果表明,ANOVA/PLS-DA方法可以有效地减少饮食因素和性别因素之间的相互...     

Sulfonate-grafted porous polymer networks for preferential CO2 adsorption at low pressure

A porous polymer network (PPN) grafted with sulfonic acid (PPN-6-SO(3)H) and its lithium salt (PPN-6-SO(3)Li) exhibit significant increases in isosteric heats of CO(2) adsorption and CO(2)-uptake capacities. IAST calculations using single-component-isotherm data and a 15/85 CO(2)/N(2) ratio at 295 K and 1 bar revealed that the sulfonate-grafted PPN-6 networks show exceptionally high adsorption selectivity for CO(2) over N(2) (155 and 414 for PPN-6-SO(3)H and PPN-6-SO(3)Li, respectively). Since these PPNs also possess ultrahigh physicochemical stability, practical applications in postcombustion capture of CO(2) lie well within the realm of possibility.     

Hydrodynamic fractionation of finite size gold nanoparticle clusters

We demonstrate a high-resolution in situ experimental method for performing simultaneous size classification and characterization of functional gold nanoparticle clusters (GNCs) based on asymmetric-flow field flow fractionation (AFFF). Field emission scanning electron microscopy, atomic force microscopy, multi-angle light scattering (MALS), and in situ ultraviolet-visible optical spectroscopy provide complementary data and imagery confirming the cluster state (e.g., dimer, trimer, tetramer), packing structure, and purity of fractionated populations. An orthogonal analysis of GNC size distributions is obtained using electrospray-differential mobility analysis (ES-DMA). We find a linear correlation between the normalized MALS intensity (measured during AFFF elution) and the corresponding number concentration (measured by ES-DMA), establishing the capacity for AFFF to quantify the absolute number concentration of GNCs. The results and corresponding methodology summarized here provide the proof of concept for general applications involving the formation, isolation, and in situ analysis of both functional and adventitious nanoparticle clusters of finite size.     

Collapse transition of a square-lattice polymer with next nearest-neighbor interaction

We study the collapse transition of a polymer on a square lattice with both nearest-neighbor and next nearest-neighbor interactions, by calculating the exact partition function zeros up to chain length 36. The transition behavior is much more pronounced than that of the model with nearest-neighbor interactions only. The crossover exponent and the transition temperature are estimated from the scaling behavior of the first zeros with increasing chain length. The results suggest that the model is of the same universality class as the usual θ point described by the model with only nearest-neighbor interaction.     

The lipoproteins of cyanobacterial photosystem II

Photosystem II (PSII) complexes from cyanobacteria and plants perform water splitting and plastoquinone reduction and yet have a different complement of lumenal extrinsic proteins. Whereas PSII from all organisms has the PsbO extrinsic protein, crystal structures of PSII from cyanobacteria have PsbV and PsbU while green algae and higher plants instead contain the extrinsic PsbP and PsbQ subunits. Proteomic studies in Synechocystis sp. PCC 6803 identified three further extrinsic proteins in the thylakoid lumen that are associated with cyanobacterial PSII and these are predicted to attach to the thylakoid membrane via a lipidated N-terminus. These proteins are cyanobacterial homologues to the PsbP and PsbQ subunits as well as to Psb27, an additional extrinsic protein associated with "inactive" photosystems that lack the other extrinsic polypeptides. The PsbQ homologue is not present in Prochlorococcus species but otherwise these proteins have been identified in most cyanobacteria although our phylogenetic analyses identified some strains that lack an apparent motif for lipidation in one or other of these subunits. Over the past decade the physiological function of these additional lipoproteins has been investigated in several cyanobacterial strains and recently the structures for each have been solved. This review will evaluate the physiological and structural results obtained for these lipid-attached extrinsic proteins and in silico protein docking of these proteins to PSII centers will be presented.