High dimensional model representation constructed by support vector regression. I. Independent variables with known probability distributions

High dimensional model representation (HDMR) is a general set of quantitative model assessment and analysis tools for capturing high dimensional input-output system behavior. In practice, the HDMR component functions are each approximated by an appropriate basis function expansion. This procedure often requires many input-output samples which can restrict the treatment of high dimensional systems. In order to address this problem we introduce svr-based HDMR to efficiently and effectively construct the HDMR expansion by support vector regression (SVR) for a function \(f(\mathbf{x})\). In this paper the results for independent variables sampled over known probability distributions are reported. The theoretical foundation of the new approach relies on the kernel used in SVR itself being an HDMR expansion (referred to as the HDMR kernel ), i.e., an ANOVA kernel whose component kernels are mutually orthogonal and all non-constant component kernels have zero expectation. Several HDMR kernels are constructed as illustrations. While preserving the characteristic properties of HDMR, the svr-based HDMR method enables efficient construction of high dimensional models with satisfactory prediction accuracy from a modest number of samples, which also permits accurate computation of the sensitivity indices. A genetic algorithm is employed to optimally determine all the parameters of the component HDMR kernels and in SVR. The svr-based HDMR introduces a new route to advance HDMR algorithms. Two examples are used to illustrate the capability of the method.     

Starch biosynthesis and modification of starch structure in transgenic plants

Starch is synthesised through the ADP-glucose pathway, involving the three enzymes ADP-glucose pyrophosphorylase, starch synthase and starch branching enzyme. ADP-glucose pyrophosphorylase is the key enzyme of the pathway, determining the flux of carbon into starch. It generates ADP-glucose, which is the substrate for the starch synthases, from glucose-1-phosphate and ATP releasing pyrophosphate. The enzyme is stimulated by 3-phosphoglycerate and inhibited through inorganic phosphate. The starch synthases, which catalyse the transfer of glucose from ADP-glucose to the nonreducing end of a growing α-1,4-glucan, are divided into two classes, the granule-bound starch synthases (GBSS) and the soluble starch synthases (SS). In both classes several isoforms have been described from many different plant species. The branching enzyme, which introduces branchpoints into the amylopectin, can also occur in different isoforms. Other enzymes present in plants, which also act on α-1,4-glucans, such as the starch phosphorylases, disproportionating enzyme and different starch hydrolases, might also be important for dertermining the starch structure and, therefore, its processibility. Many aspects of starch synthesis are not fully understood to date. Starch metabolism can be manipulated through genetic engineering, either by the ectopic expression of different heterologous genes, or through the repression of the expression of endogenous genes using antisense RNA technology. This not only allows the functional analysis of starch biosynthetic proteins, but also the manipulation of starch structure in order to widen its industrial applications. In this way many different potato lines have been generated, containing either different amounts of starch, or which synthesize a structurally modified starch. These structural changes relate to the amylose content, the phosphate content, or the gelatinisation and gelation characteristics of the starch.     

壳聚糖膜固定生物活性短肽RGDS的研究

为改善壳聚糖对细胞的特异性吸附,采用水溶性碳二亚胺将生物活性短肽精氨酸-甘氨酸-天冬氨酸-丝氨酸(RGDS)固定到壳聚糖膜的表面,采用X射线光电子能谱检测固定多肽前后的壳聚糖膜表面,发现反应后壳聚糖膜表面氮元素含量增大,Nls和Cls曲线拟合谱中酰胺键增多,表明RGDS短肽已固定到壳聚糖膜的表面;人角膜缘上皮细胞体外培养实验表明,固定RGDS后壳聚糖膜的细胞黏附率有了明显提高,固定RGDS后的壳聚糖膜在角膜组织工程支架等方面有更好的应用潜力。     

Astrophysical S-factors from asymptotic normalization coefficients

S-factors for direct capture reactions can be found at astrophysical energies from asymptotic normalization coefficients which provide the normalization of the tail of the overlap function. For example the overlap for ⁸B → ⁷Be+p defines the S-factor for ⁷Be (p, γ)⁸B. Peripheral transfer reactions offer a technique to determine these asymptotic normalization coefficients. As a test of the technique, the ¹⁶O(³He, d)¹⁷F reaction has been used to determine asymptotic normalization coefficients for transitions to the ground and first excited states of ¹⁷F. The S-factors for ¹⁶O(p, γ)¹⁷F calculated from these ¹⁷F → ¹⁶O+p asymptotic normalization coefficients are found to be in very good agreement with recent measurements. Following the same technique, the ¹⁰B(⁷Be, ⁸B)⁹Be and ¹⁴N(⁷Be, ⁸B)¹³C reactions have been used to measure the asymptotic normalization coefficient for ⁷Be(p, γ)⁸B. This result provides an indirect determination of S ₁₇(0).     

Micro-electro-mechanical deformable mirrors for aberration control in optical systems

Controlling optical aberrations is one of the enduring problems in optics. Recent advances in adaptive optics for astronomical applications have shown the promise of adaptive optics technology for controlling aberrations. Micro-electro-mechanical deformable mirrors (MEM-DMs) offer an alternative to conventional adaptive optics which, due to the inexpensive nature of MEM-DM technology, will enable a wide range of commercial and scientific applications for optical wave front control. In this paper we describe MEM-DMs, present results of modelling the performance of an MEM-DM for optical aberration control, and present results of experiments to verify that MEM-DMs can control optical aberrations.     

Anionic N-heterocyclic carbenes with N,N'-bis(fluoroaryl) and N,N'-bis(perfluoroaryl) substituents

A series of rhodium complexes, [Rh(cod)(NHC-F(x))(OH(2))] (cod = 1,5-cyclooctadiene; NHC = N-heterocyclic carbene), incorporating anionic N-heterocyclic carbenes with 2-tert-butylmalonyl backbones and 2,6-dimethylphenyl (x = 0), 2,6-difluorophenyl (x = 4), 2,4,6-trifluorophenyl (x = 6), and pentafluorophenyl (x = 10) N,N'-substituents, respectively, has been prepared by deprotonation of the corresponding zwitterionic precursors with potassium hexamethyldisilazide, followed by immediate reaction of the resulting potassium salts with [{RhCl(cod)}(2)]. These complexes could be converted to the related carbonyl derivatives [Rh(CO)(2)(NHC-F(x))(OH(2))] by displacement of the COD ligand with CO. IR and NMR spectroscopy demonstrated that the degree of fluorination of the N-aryl substituents has a considerable influence on the σ-donating and π-accepting properties of the carbene ligands and could be effectively used to tune the electronic properties of the metal center. The carbonyl groups on the carbene ligand backbone provided a particularly sensitive probe for the assessment of the metal-to-ligand π donation. The ortho-fluorine substituents on the N-aryl groups in the carbene ligands interacted with the other ligands on rhodium, determining the conformation of the complexes and creating a pocket suitable for the coordination of water to the metal center. Computational studies were used to explain the influence of the fluorinated N-substituents on the electronic properties of the ligand and evaluate the relative contribution of the σ- and π-interactions to the ligand-metal interaction.     

Logical Analysis of Data in Structure‐Activity Investigation of Polymeric Gene Delivery

To date semi‐empirical or surrogate modeling has demonstrated great success in the prediction of the biologically relevant properties of polymeric materials. For the first time, a correlation between the chemical structures of poly(β‐amino esters) and their efficiency in transfecting DNA was established using the novel technique of logical analysis of data (LAD). Linear combination and explicit representation models were introduced and compared in the framework of the present study. The most successful regression model yielded satisfactory agreement between the predicted and experimentally measured values of transfection efficiency (Pearson correlation coefficient, 0.77; mean absolute error, 3.83). It was shown that detailed analysis of the rules provided by the LAD algorithm offered practical utility to a polymer chemist in the design of new biomaterials.

     

Highly predictive CoMFA and CoMSIA models for two series of stromelysin-1 (MMP-3) inhibitors elucidate S1' and S1-S2' binding modes

Three-dimensional quantitative structure-activity relationship models have been derived using comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) for two training sets of arylsulfonyl isoquinoline-based and thazine/thiazepine-based matrix metalloproteinase inhibitors (MMPIs). The crystal structure of stromelysin-1 (MMP-3) was used to pinpoint areas on the ligands and receptors where steric and electrostatic effects (for CoMFA) and steric, electrostatic, hydrogen-bond donor, hydrogen-bond acceptor, and hydrophobic effects (for CoMSIA) correlate with an increase or decrease in experimental biological activity. The most predictive CoMFA and CoMSIA models were obtained using training-series subsets that sampled a wide range of activities, together with docking and scoring, inertial alignment, investigation of various partial charge formalisms, and manual adjustment of each compound within the active site. The models developed in this study are in agreement with experimentally observed MMP-3 structure-activity relationship data and offer new insights into binding modes involving the partly solvent-exposed S1-S2' subpocket and certain zinc-chelating groups.