Eliciting and encoding expert knowledge on variable selection into classical or Bayesian species distribution models

The quality of species distribution models (SDMs) relies to a large degree on the quality of the input data, from bioclimatic indices to environmental and habitat descriptors (Austin, 2002). Recent reviews of SDM techniques, have sought to optimize predictive performance e.g. Elith et al., 2006. In general SDMs employ one of three approaches to variable selection. The simplest approach relies on the expert to select the variables, as in environmental niche models Nix, 1986 or a generalized linear model without variable selection (Miller and Franklin, 2002). A second approach explicitly incorporates variable selection into model fitting, which allows examination of particular combinations of variables. Examples include generalized linear or additive models with variable selection (Hastie et al. 2002); or classification trees with complexity or model based pruning (Breiman et al., 1984, Zeileis, 2008). A third approach uses model averaging, to summarize the overall contribution of a variable, without considering particular combinations. Examples include neural networks, boosted or bagged regression trees and Maximum Entropy as compared in Elith et al. 2006. Typically, users of SDMs will either consider a small number of variable sets, via the first approach, or else supply all of the candidate variables (often numbering more than a hundred) to the second or third approaches. Bayesian SDMs exist, with several methods for eliciting and encoding priors on model parameters (see review in Low Choy et al. 2010). However few methods have been published for informative variable selection; one example is Bayesian trees (O’Leary 2008). Here we report an elicitation protocol that helps makes explicit a priori expert judgements on the quality of candidate variables. This protocol can be flexibly applied to any of the three approaches to variable selection, described above, Bayesian or otherwise. We demonstrate how this information can be obtained then used to guide variable selection in classical or machine learning SDMs, or to define priors within Bayesian SDMs.

Information theory and resistance fluctuations in one-dimensional disordered conductors

A novel method is proposed to treat the problem of the random resistance of a strictly one-dimensional conductor with static disorder. For the probability distribution of the transfer matrix R of the conductor we propose a distribution of maximum information entropy, constrained by the following physical requirements: (1) flux conservation, (2) time-reversal invariance, and (3) scaling with the length of the conductor of the two lowest cumulants of ω, where R=exp(iω→⋅Jbhat). The preliminary results discussed in the text are in qualitative agreement with those obtained by sophisticated microscopic theories.

Thermodynamic and kinetic studies on saccharide binding to soya-bean agglutinin.

The fluorescence of N-dansylgalactosamine [N-(5-dimethylaminonaphthalene-1-sulphonyl)galactosamine] was enhanced 11-fold with a 25 nm blue-shift in the emission maximum upon binding to soya-bean agglutinin (SBA). This change was used to determine the association constants and thermodynamic parameters for this interaction. The association constant of 1.51 X 10(6) M-1 at 20 degrees C indicated a very strong binding, which is mainly due to a relatively small entropy value, as revealed by the thermodynamic parameters: delta G = -34.7 kJ X mol-1, delta H = -37.9 kJ X mol-1 and delta S = -10.9 J X mol-1 X K-1. The specific binding of this sugar to SBA shows that the lectin can accommodate a large hydrophobic substituent on the C-2 of galactose. Binding of non-fluorescent ligands, studied by monitoring the fluorescence changes when they are added to a mixture of SBA and N-dansylgalactosamine, indicates that a hydrophobic substituent at the anomeric position increases the affinity of the interaction. The C-6 hydroxy group also stabilizes the binding considerably. Kinetics of binding of N-dansylgalactosamine to SBA studied by stopped-flow spectrofluorimetry are consistent with a single-step mechanism and yielded k+1 = 2.4 X 10(5) M-1 X s-1 and k-1 = 0.2 s-1 at 20 degrees C. The activation parameters indicate an enthalpicly controlled association process.

Thermal studies on melting and crystallization of polyamide fibres

Effect of heating rate on melting and crystallization of polyamide fibres has been examined using differential scanning calorimetric (DSC) technique. Peak temperature for melting (T m) and crystallization (T k) get suppressed with the increase in the heating rate which has been explained on the basis of chain orientation. Heat of melting (DeltaH m) and crystallization (DeltaH k) have been measured.DeltaH m vs. T m shows a nonlinear dependence which has been explained on the basis of entropy change. Quantitative difference inDeltaH m andDeltaH k values has been explained on the basis of orientation and degradation of the polymer.

Potential Distribution and Risk Assessment of an Invasive Plant Species: A Case Study of Hymenachne amplexicaulis in Australia

Given the limited resources available for weed management, a strategic approach is required to give the best bang for your buck. The current study incorporates: (1) a model ensemble approach to identify areas of uncertainty and commonality regarding a species invasive potential, (2) current distribution of the invaded species, and (3) connectivity of systems to identify target regions and focus efforts for more effective management. Uncertainty in the prediction of suitable habitat for H. amplexicaulis (study species) in Australia was addressed in an ensemble-forecasting approach to compare distributional scenarios from four models (CLIMATCH; CLIMEX; boosted regression trees [BRT]; maximum entropy [Maxent]). Models were built using subsets of occurrence and environmental data. Catchment risk was determined through incorporating habitat suitability, the current abundance and distribution of H. amplexicaulis, and catchment connectivity. Our results indicate geographic differences between predictions of different approaches. Despite these differences a number of catchments in northern, central, and southern Australia were identified as high risk of invasion or further spread by all models suggesting they should be given priority for the management of H. amplexicaulis. The study also highlighted the utility of ensemble approaches in indentifying areas of uncertainty and commonality regarding the species invasive potential.

Microstructure and mechanical properties of unidirectionally solidified Al-Si-Ni ternary eutectic

Al-10.98 pct Si-4.9 pct Ni ternary eutectic alloy was unidirectionally solidified at growth rates from 1.39μm/sec to 6.95μm/sec. Binary Al-Ni and Al-Si eutectics prepared from the same purity metals were also solidified under similar conditions to characterize the growth conditions under the conditions of present study. NiAl3 phase appeared as fibers in the binary Al-Ni eutectic and silicon appeared as irregular plates in the binary Al-Si eutectic. However, in the ternary Al-Si-Ni eutectic alloy both NiAl3 and silicon phases appeared as irregular plates dispersed in α-Al phase, without any regular repctitive arrangement. The size and spacing of NiAl3 and Si platelets in cone shaped colonies decreased with an increase in the growth rate of the ternary eutectic. Examination of specimen quenched during unidirectional solidification indicated that the ternary eutectic grows with a non-planar interface with both Si and NiAl3 phases protruding into the liquid. It is concluded that it will be difficult to grow regular ternary eutectic structures even if only one phase has a high entropy of melting. The tensile strength and modulus of unidirectionally solidified Al-Si-Ni eutectic was lower than the chill cast alloys of the same composition, and decreased with a decrease in growth rate. Tensile modulus and strength of ternary Al-Si-Ni eutectic alloys was greater than binary Al-Si eutectic alloy under similar growth conditions, both in the chill cast and in unidirectionally solidified conditions.

On walls of marginal stability in N=2 string theories

We study the properties of walls of marginal stability for BPS decays in a class of N = 2 theories. These theories arise in N = 2 string compactifications obtained as freely acting orbifolds of N = 4 theories, such theories include the STU model and the FHSV model. The cross sections of these walls for a generic decay in the axion-dilaton plane reduce to lines or circles. From the continuity properties of walls of marginal stability we show that central charges of BPS states do not vanish in the interior of the moduli space. Given a charge vector of a BPS state corresponding to a large black hole in these theories, we show that all walls of marginal stability intersect at the same point in the lower half of the axion-dilaton plane. We isolate a class of decays whose walls of marginal stability always lie in a region bounded by walls formed by decays to small black holes. This enables us to isolate a region in moduli space for which no decays occur within this class. We then study entropy enigma decays for such models and show that for generic values of the moduli, that is when moduli are of order one compared to the charges, entropy enigma decays do not occur in these models.

A Generalized Approach to Multiresolution Complex SAR Signal Processing

Multiresolution synthetic aperture radar (SAR) image formation has been proven to be beneficial in a variety of applications such as improved imaging and target detection as well as speckle reduction. SAR signal processing traditionally carried out in the Fourier domain has inherent limitations in the context of image formation at hierarchical scales. We present a generalized approach to the formation of multiresolution SAR images using biorthogonal shift-invariant discrete wavelet transform (SIDWT) in both range and azimuth directions. Particularly in azimuth, the inherent subband decomposition property of wavelet packet transform is introduced to produce multiscale complex matched filtering without involving any approximations. This generalized approach also includes the formulation of multilook processing within the discrete wavelet transform (DWT) paradigm. The efficiency of the algorithm in parallel form of execution to generate hierarchical scale SAR images is shown. Analytical results and sample imagery of diffuse backscatter are presented to validate the method.

Free energy calculations of glycosaminoglycan - Protein interactions

Glycosaminoglycans (GAGs) are complex highly charged linear polysaccharides that have a variety of roles in biological processes. We report the first use of molecular dynamics (MD) free energy calculations using the MM/PBSA method to investigate the binding of GAGs to protein molecules, namely the platelet endothelial cell adhesion molecule 1 (PECAM-1) and annexin A2. Calculations of the free energy of the binding of heparin fragments of different sizes reveal the existence of a region of low GAG-binding affinity in domains 5-6 of PECAM-1 and a region of high affinity in domains 2-3, consistent with experimental data and ligand-protein docking studies. A conformational hinge movement between domains 2 and 3 was observed, which allows the binding of heparin fragments of increasing size (pentasaccharides to octasaccharides) with an increasingly higher binding affinity. Similar simulations of the binding of a heparin fragment to annexin A2 reveal the optimization of electrostatic and hydrogen bonding interactions with the protein and protein-bound calcium ions. In general, these free energy calculations reveal that the binding of heparin to protein surfaces is dominated by strong electrostatic interactions for longer fragments, with equally important contributions from van der Waals interactions and vibrational entropy changes, against a large unfavorable desolvation penalty due to the high charge density of these molecules.

Urban Growth Analysis Using Spatial and Temporal Data

Urban growth identification, quantification, knowledge of rate and the trends of growth would help in regional planning for better infrastructure provision in environmentally sound way. This requires analysis of spatial and temporal data, which help in quantifying the trends of growth on spatial scale. Emerging technologies such as Remote Sensing, Geographic Information System (GIS) along with Global Positioning System (GPS) help in this regard. Remote sensing aids in the collection of temporal data and GIS helps in spatial analysis. This paper focuses on the analysis of urban growth pattern in the form of either radial or linear sprawl along the Bangalore - Mysore highway. Various GIS base layers such as builtup areas along the highway, road network, village boundary etc. were generated using collateral data such as the Survey of India toposheet, etc. Further, this analysis was complemented with the computation of Shannon's entropy, which helped in identifying prevalent sprawl zone, rate of growth and in delineating potential sprawl locations. The computation Shannon's entropy helped in delineating regions with dispersed and compact growth. This study reveals that the Bangalore North and South taluks contributed mainly to the sprawl with 559% increase in built-up area over a period of 28 years and high degree of dispersion. The Mysore and Srirangapatna region showed 128% change in built-up area and a high potential for sprawl with slightly high dispersion. The degree of sprawl was found to be directly proportional to the distances from the cities.

Thermodynamics of Monosaccharide and Disaccharide Binding to Erythrina corallodendron Lectin

Isothermal titration calorimetry measurements of the binding of 2′-fucosyllactose, lactose, N-acetyllactosamine, galactopyranose, 2-acetamido-2-deoxygalactopyranoside, methyl α-N-dansylgalactosaminide (Me-α-DNS-GalN), methyl α-D-galactopyranoside, methyl β-D-galactopyranoside, and fucose to Erythrina corallodendron lectin (ECorL), a dimer with one binding site per subunit, were performed at 283-286 and 297-299 K. The site binding enthalpies, ΔHb, with the exception of Me-α-DNS-GalN, are the same at both temperatures and range from −47.1 ± 1.0 kJ mol−1 for N-acetyllactosamine to −4.4 ± 0.3 kJ mol−1 for fucose, and the site binding constants range from 3.82 ± 0.9 × 105 M−1 for Me-α-DNS-GalN at 283.2 K to 0.46 ± 0.05 × 103 M−1 for fucose at 297.2 K. The binding reactions are mainly enthalpically driven except for fucose and exhibit enthalpy-entropy compensation. The binding enthalpies of the disaccharides are about twice the binding enthalpies of the monosaccharides in contrast to concanavalin A where the binding enthalpies do not double for the disaccharides. Differential scanning calorimetry measurements show that denaturation of the ECorL dimer results in dissociation into its monomer subunits. The binding constants from the increase in denaturation temperature of ECorL in the presence of saccharides are in agreement with values from isothermal titration calorimetry results. The thermal denaturation of ECorL occurs around 333 K, well below the 344-360 K denaturation temperature of other legume lectins of similar size and tertiary structure, undoubtedly due to the difference in its quaternary structure relative to other legume lectins. This is also apparent from the independent unfolding of its two domains.

On the relationship of thermodynamic parameters with the buried surface area in protein-ligand complex formation

Prediction of thermodynamic parameters of protein-protein and antigen-antibody complex formation from high resolution structural parameters has recently received much attention, since an understanding of the contributions of different fundamental processes like hydrophobic interactions, hydrogen bonding, salt bridge formation, solvent reorganization etc. to the overall thermodynamic parameters and their relations with the structural parameters would lead to rational drug design. Using the results of the dissolution of hydrocarbons and other model compounds the changes in heat capacity (DeltaCp), enthalpy (DeltaH) and entropy (DeltaS) have been empirically correlated with the polar and apolar surface areas buried during the process of protein folding/unfolding and protein-ligand complex formation. In this regard, the polar and apolar surfaces removed from the solvent in a protein-ligand complex have been calculated from the experimentally observed values of changes in heat capacity (DeltaCp) and enthalpy (DeltaH) for protein-ligand complexes for which accurate thermodynamic and high resolution structural data are available, and the results have been compared with the x-ray crystallographic observations. Analyses of the available results show poor correlation between the thermodynamic and structural parameters. Probable reasons for this discrepancy are mostly related with the reorganization of water accompanying the reaction which is indeed proven by the analyses of the energetics of the binding of the wheat germ agglutinin to oligosaccharides.

Multicarrier analysis of magnetotransport data at low and high electric fields

We present some results on multicarrier analysis of magnetotransport data, Both synthetic as well as data from narrow gap Hg0.8Cd0.2Te samples are used to demonstrate applicability of various algorithms vs. nonlinear least square fitting, Quantitative Mobility Spectrum Analysis (QMSA) and Maximum Entropy Mobility Spectrum Analysis (MEMSA). Comments are made from our experience oil these algorithms, and, on the inversion procedure from experimental R/sigma-B to S-mu specifically with least square fitting as an example. Amongst the conclusions drawn are: (i) Experimentally measured resistivity (R-xx, R-xy) should also be used instead of just the inverted conductivity (sigma(xx), sigma(xy)) to fit data to semiclassical expressions for better fits especially at higher B. (ii) High magnetic field is necessary to extract low mobility carrier parameters. (iii) Provided the error in data is not large, better estimates to carrier parameters of remaining carrier species can be obtained at any stage by subtracting highest mobility carrier contribution to sigma from the experimental data and fitting with the remaining carriers. (iv)Even in presence of high electric field, an approximate multicarrier expression can be used to guess the carrier mobilities and their variations before solving the full Boltzmann equation.

Sequence Specific Interaction between Promoter DNA and Escherichia coli RNA Polymerase: Comparative Thermodynamic Analysis with One Immobilized Partner

Sequence specific interaction between DNA and protein molecules has been a subject of active investigation for decades now. Here, we have chosen single promoter containing bacteriophage Delta D-III T7 DNA and Escherichia coli RNA polymerase and followed their recognition at the air-water interface by using the surface plasmon resonance (SPR) technique, where the movement of one of the reacting species is restricted by way of arraying them on an immobilized support. For the Langmuir monolayer studies, we used a RNA polymerase with a histidine tag attached to one of its subunits, thus making it an xcellent substrate for Ni(II) ions, while the SPR Studies were done using biotin-labeled DNA immobilized on a streptavidin-coated chip. Detailed analysis of the thermodynamic parameters as a function of concentration and temperature revealed that the interaction of RNA polymerase with T7 DNA is largely entropy driven (83 (+/- 12) kcal mol(-1)) with a positive enthalpy of 13.6 (+/- 3.6) kcal mol(-1), The free energy of reaction determined by SPR and Langmuir-Blodgett technique was -11 (+/- 2) and -15.6 kcal mol(-1), respectively. The ability of these methods to retain the specificity of the recognition process was also established.

Titration calorimetric studies to elucidate the specificity of the interactions of polymyxin B with lipopolysaccharides and lipid A

Lipopolysaccharide (LPS), the major cell wall constituent of Gram-negative bacteria, evokes a multitude of biological effects in mammals including pyrogenicity and toxic shock syndrome. Polymyxin B (PmB), a polycationic cyclic peptide, is known to neutralize most of its activities. The nature of the interaction of PmB with LPS and lipid A was investigated by isothermal titration calorimetry. PmB binds to LPS as well as lipid A stoichiometrically and non-co-operatively with micromolar affinity. These interactions are driven primarily by a favourable change in entropy (delta S) and are endothermic in nature. These positive changes in enthalpies decrease with increasing temperature, yielding a heat capacity change, delta Cp, of -2385 J.mol-1.degree-1 for PmB-LPS interactions while the binding of PmB to lipid A displays a delta Cp of -2259 J.mol-1.degree-1. The negative heat capacity changes provide strong evidence for the role of hydrophobic interactions as the driving force for the association of PmB with LPS and lipid A. A correlation of the energetics of these interactions with analyses of the molecular models of PmB suggests that a cluster of solvent-exposed non-polar amino acid side-chains that line one surface of the molecule, together with a ring of positively charged residues on its other surface, are responsible for its strong and stoichiometric binding to LPS.