Efficient algorithms for learning kernels from multiple similarity matrices with general convex loss functions

In this paper we consider the problem of learning an n × n kernel matrix from m(1) similarity matrices under general convex loss. Past research have extensively studied the m = 1 case and have derived several algorithms which require sophisticated techniques like ACCP, SOCP, etc. The existing algorithms do not apply if one uses arbitrary losses and often can not handle m > 1 case. We present several provably convergent iterative algorithms, where each iteration requires either an SVM or a Multiple Kernel Learning (MKL) solver for m > 1 case. One of the major contributions of the paper is to extend the well knownMirror Descent(MD) framework to handle Cartesian product of psd matrices. This novel extension leads to an algorithm, called EMKL, which solves the problem in O(m2 log n 2) iterations; in each iteration one solves an MKL involving m kernels and m eigen-decomposition of n × n matrices. By suitably defining a restriction on the objective function, a faster version of EMKL is proposed, called REKL,which avoids the eigen-decomposition. An alternative to both EMKL and REKL is also suggested which requires only an SVMsolver. Experimental results on real world protein data set involving several similarity matrices illustrate the efficacy of the proposed algorithms.

Hydrogen-bond dynamics near a micellar surface: Origin of the universal slow relaxation at complex aqueous interfaces

The dynamics of hydrogen bonds among water molecules themselves and with the polar head groups (PHG) at a micellar surface have been investigated by long molecular dynamics simulations. The lifetime of the hydrogen bond between a PHG and a water molecule is found to be much longer than that between any two water molecules, and is likely to be a general feature of hydrophilic surfaces of organized assemblies. Analyses of individual water trajectories suggest that water molecules can remain bound to the micellar surface for more than 100 ps. The activation energy for such a transition from the bound to a free state for the water molecules is estimated to be about 3.5 kcal/mol.

Nonlinear dynamics of eucaryotic pyruvate dehydrogenase multienzyme complex: Decarboxylation rate, oscillations, and multiplicity

Pyruvate conversion to acetyl-CoA by the pyruvate dehydrogenase (PDH) multienzyme complex is known as a key node in affecting the metabolic fluxes of animal cell culture. However, its possible role in causing possible nonlinear dynamic behavior such as oscillations and multiplicity of animal cells has received little attention. In this work, the kinetic and dynamic behavior of PDH of eucaryotic cells has been analyzed by using both in vitro and simplified in vivo models. With the in vitro model the overall reaction rate (v(1)) of PDH is shown to be a nonlinear function of pyruvate concentration, leading to oscillations under certain conditions. All enzyme components affect v, and the nonlinearity of PDH significantly, the protein X and the core enzyme dihydrolipoamide acyltransferase (E2) being mostly predominant. By considering the synthesis rates of pyruvate and PDH components the in vitro model is expanded to emulate in vivo conditions. Analysis using the in vivo model reveals another interesting kinetic feature of the PDH system, namely, multiple steady states. Depending on the pyruvate and enzyme levels or the operation mode, either a steady state with high pyruvate decarboxylation rate or a steady state with significantly lower decarboxylation rate can be achieved under otherwise identical conditions. In general, the more efficient steady state is associated with a lower pyruvate concentration. A possible time delay in the substrate supply and enzyme synthesis can also affect the steady state to be achieved and lead's to oscillations under certain conditions. Overall, the predictions of multiplicity for the PDH system agree qualitatively well with recent experimental observations in animal cell cultures. The model analysis gives some hints for improving pyruavte metabolism in animal cell culture.

Real-time kinetic analysis of hCG-monoclonal antibody interaction using radiolabeled hCG probe: presence of two forms of Ag-mAb complex as revealed by solid phase dissociation studies

Real-time kinetics of ligand-ligate interaction has predominantly been studied by either fluorescence or surface plasmon resonance based methods. Almost all such studies are based on association between the ligand and the ligate. This paper reports our analysis of dissociation data of monoclonal antibody-antigen (hCG) system using radio-iodinated hCG as a probe and nitrocellulose as a solid support to immobilize mAb. The data was analyzed quantitatively for a one-step and a two-step model. The data fits well into the two-step model. We also found that a fraction of what is bound is non-dissociable (tight-binding portion (TBP)). The TBP was neither an artifact of immobilization nor does it interfere with analysis. It was present when the reaction was carried out in homogeneous solution in liquid phase. The rate constants obtained from the two methods were comparable. The work reported here shows that real-time kinetics of other ligand-ligate interaction can be studied using nitrocellulose as a solid support. (C) 2002 Elsevier Science B.V. All rights reserved.

Structural model for the Cu-B site of dopamine beta-hydroxylase: Crystal structure of a copper(II) complex showing N3OS coordination with an axial sulfur ligation

A copper(II) complex containing a NSO-donor Schiff base and NN-donor 2,2'-bipyridine has been prepared and structurally characterized. The square pyramidal complex with an axial sulfur ligation is a structural model for the CUB site of dopamine-hydroxylase in its oxidized form. The copper(II) complex is catalytically active in the oxidation of ascorbic acid by dioxygen mediated by a copper(I) species which is proposed to have a four-coordinate structure with a N3S coordination geometry.

The first copper(I) complex of tris(hydroxymethyl)phosphine

The first copper(I) complex of tris(hydroxymethyl)phosphine (THP) has been synthesized from a reaction with [Cu(CH3CN)4]ClO4. The crystal structure and spectroscopic properties of this complex showing the tetrahedral coordination around copper and a deprotonated THP coordinated to Cu through P are presented.

Superconductivity of embedded lead nano particles in metallic and amorphous matrices

Nanodispersed lead in metallic and amorphous matrices was synthesized by rapid solidification processing. The optimum microstructure was tailored to avoid percolation of the particles. With these embedded particles it is possible to study quantitatively the effect of size on the superconducting transition temperature by carrying out quantitative microstructural characterization and magnetic measurements. Our results suggest the role of the matrices in enhancement or depression of superconducting transition temperature of lead. The origin of this difference in behavior with respect to different matrices and sizes is discussed.

Microwave spectroscopic and theoretical studies on the phenylacetylene center dot center dot center dot H(2)O complex: C-H center dot center dot center dot O and O-H center dot center dot center dot pi hydrogen bonds as equal partners

Rotational spectra of five isotopologues of the title complex, C(6)H(5)CCH center dot center dot center dot H(2)O, C(6)H(5)CCH center dot center dot center dot HOD, C(6)H(5)CCH center dot center dot center dot D(2)O, C(6)H(5)CCH center dot center dot center dot H(2)(18)O and C(6)H(5)CCD center dot center dot center dot H(2)O, were measured and analyzed. The parent isotopologue is an asymmetric top with kappa = -0.73. The complex is effectively planar (ab inertial plane) and both `a' and `b' dipole transitions have been observed but no c dipole transition could be seen. All the transitions of the parent complex are split into two resulting from an internal motion interchanging the two H atoms in H(2)O. This is confirmed by the absence of such doubling for the C(6)H(5)CCH center dot center dot center dot HOD complex and a significant reduction in the splitting for the D(2)O analog. The rotational spectra, unambiguously, reveal a structure in which H(2)O has both O-H center dot center dot center dot pi (pi cloud of acetylene moiety) and C-H center dot center dot center dot O (ortho C-H group of phenylacetylene) interactions. This is in agreement with the structure deduced by IR-UV double resonance studies (Singh et al., J. Phys. Chem. A, 2008, 112, 3360) and also with the global minimum predicted by advanced electronic structure theory calculations (Sedlack et al., J. Phys. Chem. A, 2009, 113, 6620). Atoms in Molecule (AIM) theoretical analysis of the complex reveals the presence of both O-H center dot center dot center dot pi and C-H center dot center dot center dot O hydrogen bonds. More interestingly, based on the electron densities at the bond critical points, this analysis suggests that both these interactions are equally strong. Moreover, the presence of both these interactions leads to significant deviation from linearity of both hydrogen bonds.

Efficient visible light induced nuclease activity of a ternary mono-1,10-phenanthroline copper(II) complex containing 2-(methylthio)ethylsalicylaidimine

Ternary Schiff base copper(II) complex [CuL(phen)](ClO4), where HL is 2-(methylthio)ethylsalicylaldimine and phen is 1,10-phenanthroline, has been prepared and structurally characterized by X-ray crystallography. The complex shows a CuN3OS coordination in a square-pyramidal (4 + 1) geometry with the sulfur as an equatorial ligand. The complex is an avid binder to double-stranded DNA in the minor groove and exhibits both photonuclease and chemical nuclease activity. When exposed to UV light of 312 nm (96 W) or visible light of 532 nm (125 W) under aerobic conditions, the complex causes significant cleavage of supercoiled pUC19 DNA in the absence of any externally added reducing agent or H2O2.

Novel double insertion of carbon disulfide into two Ru-H bonds of [(dppm)(2)Ru(H)(2)] (dppm = Ph2PCH2PPh2): Synthesis and crystal structure of a methanedithiolate complex

Addition of excess carbon disulfide to cis/trans-[(dPPM)(2)Ru(H)(2)] results in the methanedithiolate complex [(dppm)(2)Ru(eta(2)-S2CH2)] 4 via the intermediacy of cis-[(dppm)(2)Ru(H)(SC(S)H)] 2. The X-ray crystal structure of this species has been determined.

Doubly spectral stochastic finite element method (DSSFEM) for random field problems

Uncertainties in complex dynamic systems play an important role in the prediction of a dynamic response in the mid- and high-frequency ranges. For distributed parameter systems, parametric uncertainties can be represented by random fields leading to stochastic partial differential equations. Over the past two decades, the spectral stochastic finite-element method has been developed to discretize the random fields and solve such problems. On the other hand, for deterministic distributed parameter linear dynamic systems, the spectral finite-element method has been developed to efficiently solve the problem in the frequency domain. In spite of the fact that both approaches use spectral decomposition (one for the random fields and the other for the dynamic displacement fields), very little overlap between them has been reported in literature. In this paper, these two spectral techniques are unified with the aim that the unified approach would outperform any of the spectral methods considered on their own. An exponential autocorrelation function for the random fields, a frequency-dependent stochastic element stiffness, and mass matrices are derived for the axial and bending vibration of rods. Closed-form exact expressions are derived by using the Karhunen-Loève expansion. Numerical examples are given to illustrate the unified spectral approach.