Large-Scale Elastic Net Regularized Linear Classification SVMs and Logistic Regression

Elastic Net Regularizers have shown much promise in designing sparse classifiers for linear classification. In this work, we propose an alternating optimization approach to solve the dual problems of elastic net regularized linear classification Support Vector Machines (SVMs) and logistic regression (LR). One of the sub-problems turns out to be a simple projection. The other sub-problem can be solved using dual coordinate descent methods developed for non-sparse L2-regularized linear SVMs and LR, without altering their iteration complexity and convergence properties. Experiments on very large datasets indicate that the proposed dual coordinate descent - projection (DCD-P) methods are fast and achieve comparable generalization performance after the first pass through the data, with extremely sparse models.

Correlation between local structural dynamics of proteins inferred from NMR ensembles and evolutionary dynamics of homologues of known structure

Conformational changes in proteins are extremely important for their biochemical functions. Correlation between inherent conformational variations in a protein and conformational differences in its homologues of known structure is still unclear. In this study, we have used a structural alphabet called Protein Blocks (PBs). PBs are used to perform abstraction of protein 3-D structures into a 1-D strings of 16 alphabets (a-p) based on dihedral angles of overlapping pentapeptides. We have analyzed the variations in local conformations in terms of PBs represented in the ensembles of 801 protein structures determined using NMR spectroscopy. In the analysis of concatenated data over all the residues in all the NMR ensembles, we observe that the overall nature of inherent local structural variations in NMR ensembles is similar to the nature of local structural differences in homologous proteins with a high correlation coefficient of .94. High correlation at the alignment positions corresponding to helical and beta-sheet regions is only expected. However, the correlation coefficient by considering only the loop regions is also quite high (.91). Surprisingly, segregated position-wise analysis shows that this high correlation does not hold true to loop regions at the structurally equivalent positions in NMR ensembles and their homologues of known structure. This suggests that the general nature of local structural changes is unique; however most of the local structural variations in loop regions of NMR ensembles do not correlate to their local structural differences at structurally equivalent positions in homologues.

Diffraction tomography from intensity measurements: an evolutionary stochastic search to invert experimental data

We develop iterative diffraction tomography algorithms, which are similar to the distorted Born algorithms, for inverting scattered intensity data. Within the Born approximation, the unknown scattered field is expressed as a multiplicative perturbation to the incident field. With this, the forward equation becomes stable, which helps us compute nearly oscillation-free solutions that have immediate bearing on the accuracy of the Jacobian computed for use in a deterministic Gauss-Newton (GN) reconstruction. However, since the data are inherently noisy and the sensitivity of measurement to refractive index away from the detectors is poor, we report a derivative-free evolutionary stochastic scheme, providing strictly additive updates in order to bridge the measurement-prediction misfit, to arrive at the refractive index distribution from intensity transport data. The superiority of the stochastic algorithm over the GN scheme for similar settings is demonstrated by the reconstruction of the refractive index profile from simulated and experimentally acquired intensity data. (C) 2014 Optical Society of America

Bi-Domain Protein Tyrosine Phosphatases Reveal an Evolutionary Adaptation to Optimize Signal Transduction

Significance: The bi-domain protein tyrosine phosphatases (PTPs) exemplify functional evolution in signaling proteins for optimal spatiotemporal signal transduction. Bi-domain PTPs are products of gene duplication. The catalytic activity, however, is often localized to one PTP domain. The inactive PTP domain adopts multiple functional roles. These include modulation of catalytic activity, substrate specificity, and stability of the bi-domain enzyme. In some cases, the inactive PTP domain is a receptor for redox stimuli. Since multiple bi-domain PTPs are concurrently active in related cellular pathways, a stringent regulatory mechanism and selective cross-talk is essential to ensure fidelity in signal transduction. Recent Advances: The inactive PTP domain is an activator for the catalytic PTP domain in some cases, whereas it reduces catalytic activity in other bi-domain PTPs. The relative orientation of the two domains provides a conformational rationale for this regulatory mechanism. Recent structural and biochemical data reveal that these PTP domains participate in substrate recruitment. The inactive PTP domain has also been demonstrated to undergo substantial conformational rearrangement and oligomerization under oxidative stress. Critical Issues and Future Directions: The role of the inactive PTP domain in coupling environmental stimuli with catalytic activity needs to be further examined. Another aspect that merits attention is the role of this domain in substrate recruitment. These aspects have been poorly characterized in vivo. These lacunae currently restrict our understanding of neo-functionalization of the inactive PTP domain in the bi-domain enzyme. It appears likely that more data from these research themes could form the basis for understanding the fidelity in intracellular signal transduction.

Accurate prediction of interfacial residues in two-domain proteins using evolutionary information: Implications for three-dimensional modeling

With the preponderance of multidomain proteins in eukaryotic genomes, it is essential to recognize the constituent domains and their functions. Often function involves communications across the domain interfaces, and the knowledge of the interacting sites is essential to our understanding of the structure-function relationship. Using evolutionary information extracted from homologous domains in at least two diverse domain architectures (single and multidomain), we predict the interface residues corresponding to domains from the two-domain proteins. We also use information from the three-dimensional structures of individual domains of two-domain proteins to train naive Bayes classifier model to predict the interfacial residues. Our predictions are highly accurate (approximate to 85%) and specific (approximate to 95%) to the domain-domain interfaces. This method is specific to multidomain proteins which contain domains in at least more than one protein architectural context. Using predicted residues to constrain domain-domain interaction, rigid-body docking was able to provide us with accurate full-length protein structures with correct orientation of domains. We believe that these results can be of considerable interest toward rational protein and interaction design, apart from providing us with valuable information on the nature of interactions. Proteins 2014; 82:1219-1234. (c) 2013 Wiley Periodicals, Inc.

Multimodal therapy for complete regression of malignant melanoma using constrained nonlinear optimal dynamic inversion

Using a realistic nonlinear mathematical model for melanoma dynamics and the technique of optimal dynamic inversion (exact feedback linearization with static optimization), a multimodal automatic drug dosage strategy is proposed in this paper for complete regression of melanoma cancer in humans. The proposed strategy computes different drug dosages and gives a nonlinear state feedback solution for driving the number of cancer cells to zero. However, it is observed that when tumor is regressed to certain value, then there is no need of external drug dosages as immune system and other therapeutic states are able to regress tumor at a sufficiently fast rate which is more than exponential rate. As model has three different drug dosages, after applying dynamic inversion philosophy, drug dosages can be selected in optimized manner without crossing their toxicity limits. The combination of drug dosages is decided by appropriately selecting the control design parameter values based on physical constraints. The process is automated for all possible combinations of the chemotherapy and immunotherapy drug dosages with preferential emphasis of having maximum possible variety of drug inputs at any given point of time. Simulation study with a standard patient model shows that tumor cells are regressed from 2 x 107 to order of 105 cells because of external drug dosages in 36.93 days. After this no external drug dosages are required as immune system and other therapeutic states are able to regress tumor at greater than exponential rate and hence, tumor goes to zero (less than 0.01) in 48.77 days and healthy immune system of the patient is restored. Study with different chemotherapy drug resistance value is also carried out. (C) 2014 Elsevier Ltd. All rights reserved.

Optimization of degree of sphericity of primary phase during cooling slope casting of A356 Al alloy: Taguchi method and regression analysis

The present work presents the results of experimental investigation of semi-solid rheocasting of A356 Al alloy using a cooling slope. The experiments have been carried out following Taguchi method of parameter design (orthogonal array of L-9 experiments). Four key process variables (slope angle, pouring temperature, wall temperature, and length of travel of the melt) at three different levels have been considered for the present experimentation. Regression analysis and analysis of variance (ANOVA) has also been performed to develop a mathematical model for degree of sphericity evolution of primary alpha-Al phase and to find the significance and percentage contribution of each process variable towards the final outcome of degree of sphericity, respectively. The best processing condition has been identified for optimum degree of sphericity (0.83) as A(3), B-3, C-2, D-1 i.e., slope angle of 60 degrees, pouring temperature of 650 degrees C, wall temperature 60 degrees C, and 500 mm length of travel of the melt, based on mean response and signal to noise ratio (SNR). ANOVA results shows that the length of travel has maximum impact on degree of sphericity evolution. The predicted sphericity obtained from the developed regression model and the values obtained experimentally are found to be in good agreement with each other. The sphericity values obtained from confirmation experiment, performed at 95% confidence level, ensures that the optimum result is correct and also the confirmation experiment values are within permissible limits. (c) 2014 Elsevier Ltd. All rights reserved.

Luminescence studies and EPR investigation of derived Eu doped ZnO

ZnO:Eu (0.1 mol%) nanopowders have been synthesized by auto ignition based low temperature solution combustion method. Powder X-ray diffraction (PXRD) patterns confirm the nanosized particles which exhibit hexagonal wurtzite structure. The crystallite size estimated from Scherrer's formula was found to be in the range 35-39 nm. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) studies reveal particles are agglomerated with quasi-hexagonal morphology. A blue shift of absorption edge with increase in band gap is observed for Eu doped ZnO samples. Upon 254 nm excitation, ZnO:Eu nanopowders show peaks in regions blue (420-484 nm), green (528 nm) and red (600 nm) which corresponds to both Eu2+ and Eu3+ ions. The electron paramagnetic resonance (EPR) spectrum exhibits a broad resonance signal at g= 4.195 which is attributed to Eu2+ ions. Further, EPR and thermo-luminescence (TL) studies reveal presence of native defects in this phosphor. Using TL glow peaks the trap parameters have been evaluated and discussed. (C) 2014 Elsevier B.V. All rights reserved.

Photoluminescence, thermoluminescence and EPR studies of solvothermally derived Ni2+ doped Y(OH)(3) and Y2O3 multi-particle-chain microrods

Rod like structures of hexagonal Y(OH)(3):Ni2+ and cubic Y2O3:Ni2+ phosphors have been successfully synthesized by solvothermal method. X-ray diffraction studies of as-formed product shows hexagonal phase, whereas the product heat treated at 700 degrees C shows pure cubic phase. Scanning electron micrographs (SEM) of Y(OH)(3):Ni2+ show hexagonal rods while Y2O3:Ni2+ rods were found to consist of many nanoparticles stacked together forming multi-particle-chains. EPR studies suggest that the site symmetry around Ni2+ ions is predominantly octahedral. PL spectra show emission in blue, green and red regions due to the T-3(1)(P-3)->(3)A(2)(F-3), T-1(2)(D-1)->(3)A(2)(F-3) and T-1(2)(D-1)-> T-3(2)(F-3) transitions of Ni2+ ions, respectively. TL studies were carried out for Y(OH)(3):Ni2+ and Y2O3:Ni2+ phosphor upon gamma-dose for 1-6 kGy. A single well resolved glow peaks at 195 and 230 degrees C were recorded for Y(OH)(3):Ni2+ and Y2O3:Ni2+, respectively. The glow peak intensity increases linearly up to 4 kGy and 5 kGy for Y(OH)(3):Ni2+ and Y2O3:Ni2+, respectively. The kinetic parameters such as activation energy (E), frequency factor (s) and order of kinetics (b) were estimated by different methods. The phosphor follows simple glow peak structure, linear response with gamma dose, low fading and simple trap distribution, suggesting that it is quite suitable for radiation dosimetry. (C) 2014 Elsevier B.V. All rights reserved.

Structure, interactions and evolutionary implications of a domain-swapped lectin dimer from Mycobacterium smegmatis

Crystal structure determination of the lectin domain of MSMEG_3662 from Mycobacterium smegmatis and its complexes with mannose and methyl-alpha-mannose, the first effort of its kind on a mycobacterial lectin, reveals a structure very similar to beta-prism II fold lectins from plant sources, but with extensive unprecedented domain swapping in dimer formation. The two subunits in a dimer often show small differences in structure, but the two domains, not always related by 2-fold symmetry, have the same structure. Each domain carries three sugar-binding sites, similar to those in plant lectins, one on each Greek key motif. The occurrence of beta-prism II fold lectins in bacteria, with characteristics similar to those from plants, indicates that this family of lectins is of ancient origin and had evolved into a mature system before bacteria and plants diverged. In plants, the number of binding sites per domain varies between one and three, whereas the number is two in the recently reported lectin domains from Pseudomonas putida and Pseudomonas aeruginosa. An analysis of the sequences of the lectins and the lectin domains shows that the level of sequence similarity among the three Greek keys in each domain has a correlation with the number of binding sites in it. Furthermore, sequence conservation among the lectins from different species is the highest for that Greek key which carries a binding site in all of them. Thus, it would appear that carbohydrate binding influences the course of the evolution of the lectin.

K-plane regression

In this paper, we present a novel algorithm for piecewise linear regression which can learn continuous as well as discontinuous piecewise linear functions. The main idea is to repeatedly partition the data and learn a linear model in each partition. The proposed algorithm is similar in spirit to k-means clustering algorithm. We show that our algorithm can also be viewed as a special case of an EM algorithm for maximum likelihood estimation under a reasonable probability model. We empirically demonstrate the effectiveness of our approach by comparing its performance with that of the state of art algorithms on various datasets. (C) 2014 Elsevier Inc. All rights reserved.

Polynomial Decompositions in Polynomial Time

Fix a prime p. Given a positive integer k, a vector of positive integers Delta = (Delta(1), Delta(2), ... , Delta(k)) and a function Gamma : F-p(k) -> F-p, we say that a function P : F-p(n) -> F-p is (k, Delta, Gamma)-structured if there exist polynomials P-1, P-2, ..., P-k : F-p(n) -> F-p with each deg(P-i) <= Delta(i) such that for all x is an element of F-p(n), P(x) = Gamma(P-1(x), P-2(x), ..., P-k(x)). For instance, an n-variate polynomial over the field Fp of total degree d factors nontrivially exactly when it is (2, (d - 1, d - 1), prod)- structured where prod(a, b) = a . b. We show that if p > d, then for any fixed k, Delta, Gamma, we can decide whether a given polynomial P(x(1), x(2), ..., x(n)) of degree d is (k, Delta, Gamma)-structured and if so, find a witnessing decomposition. The algorithm takes poly(n) time. Our approach is based on higher-order Fourier analysis.

Fruit and Seed Volatiles: Multiple Stage Settings, Actors and Props in an Evolutionary Play

Plants emit volatile organic compounds (VOCs) from most parts of their anatomy. Conventionally, the volatiles of leaves, flowers, fruits and seeds have been investigated separately. This review presents an integrated perspective of volatiles produced by fruits and seeds in the context of selection on the whole plant. It suggests that fruit and seed volatiles may only be understood in the light of the chemistry of the whole plant. Fleshy fruit may be viewed as an ecological arena within which several evolutionary games are being played involving fruit VOCs. Fruit odour and colour may be correlated and interact via multimodal signalling in influencing visits by frugivores. The hypothesis of volatile crypsis in the evolution of hard seeds as protection against volatile diffusion and perception by seed predators is reviewed. Current views on the role of volatiles in ant dispersal of seeds or myrmecochory are summarised, especially the suggestion that ants are being manipulated by plants in the form of a sensory trap while providing this service. Plant VOC production is presented as an emergent phenotype that could result from multiple selection pressures acting on various plant parts; the ``plant'' phenotype and VOC profile may receive significant contributions from symbionts within the plant. Viewing the plant as a holobiont would benefit an understanding of the emergent plant phenotype.

One-Carbon Metabolic Pathway Rewiring in Escherichia coli Reveals an Evolutionary Advantage of 10-Formyltetrahydrofolate Synthetase (Fhs) in Survival under Hypoxia

In cells, N-10-formyltetrahydrofolate (N-10-fTHF) is required for formylation of eubacterial/organellar initiator tRNA and purine nucleotide biosynthesis. Biosynthesis of N-10-fTHF is catalyzed by 5,10-methylene-tetrahydrofolate dehydrogenase/cyclohydrolase (FolD) and/or 10-formyltetrahydrofolate synthetase (Fhs). All eubacteria possess FolD, but some possess both FolD and Fhs. However, the reasons for possessing Fhs in addition to FolD have remained unclear. We used Escherichia coli, which naturally lacks fhs, as our model. We show that in E. coli, the essential function of folD could be replaced by Clostridium perfringens fhs when it was provided on a medium-copy-number plasmid or integrated as a single-copy gene in the chromosome. The fhs-supported folD deletion (Delta folD) strains grow well in a complex medium. However, these strains require purines and glycine as supplements for growth in M9 minimal medium. The in vivo levels of N-10-fTHF in the Delta folD strain (supported by plasmid-borne fhs) were limiting despite the high capacity of the available Fhs to synthesize N-10-fTHF in vitro. Auxotrophy for purines could be alleviated by supplementing formate to the medium, and that for glycine was alleviated by engineering THF import into the cells. The Delta folD strain (harboring fhs on the chromosome) showed a high NADP(+)-to-NADPH ratio and hypersensitivity to trimethoprim. The presence of fhs in E. coli was disadvantageous for its aerobic growth. However, under hypoxia, E. coli strains harboring fhs outcompeted those lacking it. The computational analysis revealed a predominant natural occurrence of fhs in anaerobic and facultative anaerobic bacteria.

An Algorithmic Characterization of Polynomial Functions over Z(pn)

In this paper we consider polynomial representability of functions defined over , where p is a prime and n is a positive integer. Our aim is to provide an algorithmic characterization that (i) answers the decision problem: to determine whether a given function over is polynomially representable or not, and (ii) finds the polynomial if it is polynomially representable. The previous characterizations given by Kempner (Trans. Am. Math. Soc. 22(2):240-266, 1921) and Carlitz (Acta Arith. 9(1), 67-78, 1964) are existential in nature and only lead to an exhaustive search method, i.e. algorithm with complexity exponential in size of the input. Our characterization leads to an algorithm whose running time is linear in size of input. We also extend our result to the multivariate case.