Nonparametric supervised learning by linear interpolation with maximum entropy

Nonparametric neighborhood methods for learning entail estimation of class conditional probabilities based on relative frequencies of samples that are "near-neighbors" of a test point. We propose and explore the behavior of a learning algorithm that uses linear interpolation and the principle of maximum entropy (LIME). We consider some theoretical properties of the LIME algorithm: LIME weights have exponential form; the estimates are consistent; and the estimates are robust to additive noise. In relation to bias reduction, we show that near-neighbors contain a test point in their convex hull asymptotically. The common linear interpolation solution used for regression on grids or look-up-tables is shown to solve a related maximum entropy problem. LIME simulation results support use of the method, and performance on a pipeline integrity classification problem demonstrates that the proposed algorithm has practical value.     

Novel ligands that target the mitochondrial membrane protein mitoNEET

Ligands of the thiazolidinedione (TZD) class of compounds, pioglitazone (Actos?) and rosiglitazone (Avandia?) are currently approved for treatment of type 2 diabetes and are known to bind to the PPAR-γ nuclear receptor subtype. Recent evidence suggesting PPAR-γ independent action of the TZDs led to the discovery of a novel integral outer mitochondrial membrane protein, mitoNEET. In spite of the several reported X-ray crystal structures of the unbound form of mitoNEET, the location and nature of the mitoNEET ligand binding sites (LBS) remain unknown. In this study, a molecular blind docking (BD) method was used to discover potential mitoNEET LBS and novel ligands, utilizing the program AutoDock Vina (v 1.0.2). Validation of BD was performed on the PPAR-γ receptor (PDB ID: 1ZGY) with the test compound rosiglitazone, demonstrating that the binding conformation of rosiglitazone determined by AutoDock Vina matches well with that of the cocrystallized ligand (root mean square deviation of the heavy atoms 1.45?). The locations and a general ligand binding interaction model for the LBS were determined, leading to the discovery of novel mitoNEET ligands. An in vitro fluorescence binding assay utilizing purified recombinant mitoNEET protein was used to determine the binding affinity of a predicted mitoNEET ligand, and the data obtained is in good agreement with AutoDock Vina results. The discovery of potential mitoNEET ligand binding sites and novel ligands, opens up the possibility for detailed structural studies of mitoNEET-ligand complexes, as well as rational design of novel ligands specifically targeted for mitoNEET.     

A Bayesian framework for sensory adaptation

Adaptation allows biological sensory systems to adjust to variations in the environment and thus to deal better with them. In this article, we propose a general framework of sensory adaptation. The underlying principle of this framework is the setting of internal parameters of the system such that certain prespecified tasks can be performed optimally. Because sensorial inputs vary probabilistically with time and biological mechanisms have noise, the tasks could be performed incorrectly. We postulate that the goal of adaptation is to minimize the number of task errors. This minimization requires prior knowledge of the environment and of the limitations of the mechanisms processing the information. Because these processes are probabilistic, we formulate the minimization with a Bayesian approach. Application of this Bayesian framework to the retina is successful in accounting for a host of experimental findings.     

Helmet design to facilitate thermoneutrality during forest harvesting

The purpose of this investigation was to evaluate the ventilatory characteristics of vented and non-vented helmets for use in forestry harvesting operations. A ventilation index developed by Birnbaum and Crockford (1978) was used to determine the ventilation capacity of twelve helmets varying in design and presence, location, and dimension of vents. Helmets with top vents had higher ventilation indices than non-vented, side and side/top-vented helmets. Ten physically fit men participated in a maximal oxygen consumption test and four trials wearing a non-vented, round-side-vented, round-top-vented or rectangulartop-vented helmet. Trials simulated typical summer environmental conditions (28 degrees C, 80% rh) and physiological work loads (40% VO(2)max) experienced during forest harvesting in New Zealand. The temperature and humidity under the helmet was typically lowest when the helmet with the largest vented area (288 mm(2)) in the crown was worn, although physiological responses (temperatures at the tympanum, ear and scalp; humidity at the centre of the helmet) of the participants to wearing the helmets differed significantly during some rest periods only. Top-vented helmets (in this experiment measured 288 mm(2)) should be worn to minimize temperature and humidity of the head during forest harvesting operations.     

Atmospheric and children's blood lead as indicators of vehicular traffic and other emission sources in Mumbai, India

Average concentration of Pb in atmospheric air particulates in different suburbs of Mumbai was studied for almost a decade and its spatial and temporal profiles are discussed in relation to emission sources. In general the concentration of Pb in all the residential suburban atmosphere is well below the Central Pollution Control Board (CPCB, 1994) prescribed limit of 1.5 microg m(-3) barring a few exceptions for some residential/industrial sites, such as those of Thane and Kurla scrap yards. The correlation between blood lead of children and air lead reveals that the blood Pb level in children could increase by 3.6 microg dl(-1) for an incremental rise of 1.0 microg Pb m(-3) of air. The temporal profile of air Pb values indicates a decreasing trend in residential suburbs (Khar: 1984, 0.39 microg m(-3); 1996, 0.17 microg m(-3)) as well as in suburban residential areas with low traffic (Goregaon: 1984, 0.53 microg m(-3); 1996, 0.30 microg m(-3)).     

Scanning electron microscopy studies of protein-functionalized atomic force microscopy cantilever tips

Protein-functionalized atomic force microscopy (AFM) tips have been used to investigate the interaction of individual ligand-receptor complexes. Herein we present results from scanning electron microscopy (SEM) studies of protein-functionalized AFM cantilever tips. The goals of this study were (1) to examine the surface morphology of protein-coated AFM tips and (2) to determine the stability of the coated tips. Based on SEM images, we found that bovine serum albumin (BSA) in solution spontaneously adsorbed onto the surface of silicon nitride cantilevers, forming a uniform protein layer over the surface. Additional protein layers deposited over the initial BSA-coated surface did not significantly alter the surface morphology. However, we found that avidin-functionalized tips were contaminated with debris after a series of force measurements with biotinylated agarose beads. The bound debris presumably originated from the transfer of material from the agarose bead. This observation is consistent with the observed deterioration of functional activity as measured in ligand-receptor binding force experiments.     

Reconstruction from Sparsely Sampled Data by ART with Interpolated Rays

After a brief discussion of the algebraic reconstruction techniques (ART), we introduce the attenuation problem in positron emission tomography (PET). We anticipate that a generalization of ART, the so-called cyclic subgradient projection (CSP) method, may be useful for solving this problem. This, however, has not been successfully realized, due to the fact that data collected by our proposed stationary PET detector ring are too sparsely sampled. That this is, in fact, a major problem is demonstrated by showing that ordinary ART produces reconstructions with unacceptably strong artifacts even on perfect (no attenuation) data collected according to the PET geometry. We demonstrate that the source of this artifact is the sparse sampling, and we propose the use of interpolated rays to overcome the problem. This approach is successful, as is illustrated by showing reconstructions from sparsely sampled data by ART with interpolated rays.