Using Manifold Learning for Atlas Selection in Multi-Atlas Segmentation

Multi-atlas segmentation has been widely used to segment various anatomical structures. The success of this technique partly relies on the selection of atlases that are best mapped to a new target image after registration. Recently, manifold learning has been proposed as a method for atlas selection. Each manifold learning technique seeks to optimize a unique objective function. Therefore, different techniques produce different embeddings even when applied to the same data set. Previous studies used a single technique in their method and gave no reason for the choice of the manifold learning technique employed nor the theoretical grounds for the choice of the manifold parameters. In this study, we compare side-by-side the results given by 3 manifold learning techniques (Isomap, Laplacian Eigenmaps and Locally Linear Embedding) on the same data set. We assess the ability of those 3 different techniques to select the best atlases to combine in the framework of multi-atlas segmentation. First, a leave-one-out experiment is used to optimize our method on a set of 110 manually segmented atlases of hippocampi and find the manifold learning technique and associated manifold parameters that give the best segmentation accuracy. Then, the optimal parameters are used to automatically segment 30 subjects from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). For our dataset, the selection of atlases with Locally Linear Embedding gives the best results. Our findings show that selection of atlases with manifold learning leads to segmentation accuracy close to or significantly higher than the state-of-the-art method and that accuracy can be increased by fine tuning the manifold learning process.     

Generation of a transmembrane electric potential during respiration by Azotobacter vinelandii membrand vesicles.

Membrane vesicles isolated from Azotobacter vinelandii strain O by lysis of spheroplasts in potassium of sodium phosphate buffer develop a transmembrane electric potential during respiration. The magnitude of this potential was determined by three independent methods: (i) fluorescence of 3,3'-dipropylthiodicarbocyanine and 3,3'-dihexyloxacarbocyanine; (ii) uptake of 86Rb+ in the presence of valinomycin; and (iii) uptake of [3H]triphenylmethyl phosphonium. In method (i), the relative fluorescence of these cyanine dyes in the presence of intact cells or derived vesicles is quenched during oxication of electron donors. A linear relationship between this quenching and a potassium diffusion potential was employed to calibrate the probe response. In method (ii), the steady-state concentration ratio of rubidium across the vesicle membrane during oxidation of L-malate was converted to potential by the Nernst equation. In method (iii), the steady-state concentration ratio of this lipophilic cation was likewise converted to a potential. With the exception of 3,3'-dihexyloxacarbocyanine fluorescence, these methods gave good agreement for the potential developed during L-malate oxidation by membrane vesicles. A value of 75 to 80 mV (inside negative) was obtained for vesicles prepared in potassium phosphate, and 104 mV (inside negative) was obtained for vesicles prepared in sodium phosphate. Electrogenic expulsion of hydrogen ion was observed during L-malate oxidation, and the amount of proton exodus was greater in potassium rather than the sodium-containing vesicles. This indicates the presence of a sodium-proton antiport mechanism. In addition, D-glucose uptake was observed during development of a potassium diffusion potential that was artificially imposed across the vesicle membrane. These observations suggest the presence of a glucose-proton symport mechanism in accordance with the principles of Mitchell.     

Ecological species groups of landform-level ecosystems dominated by jack pine in northern Lower Michigan,USA

A combination of field and tabular methods and multivariate analyses were used to develop groups of ground flora species (i.e., ecological species groups) that characterize and distinguish highly disturbed, landform-level ecosystems dominated by jack pine in northern Lower Michigan. The endangered Kirtland's warbler formerly or currently occupied the large glacial landforms for which species groups were developed. Eight such ecological species groups were created using 31 woody and herbaceous species sampled in 144 plots within a 20,000 km² geographic region of uniform climate and regional physiography. The groups were initially constructed using subjective, observation-based groupings of species with similar presence and abundance along soil moisture and fertility gradients. Species groups were corroborated using TWINSPAN and detrended correspondence analysis, and the environmental conditions indicated by each were described and contrasted based upon field observations and canonical correspondence analysis. Two of the eight species groups indicated very dry, infertile sites, and one was indicative of a very broad range of sites dominated by jack pine. The remaining five groups reflected a relatively gentle environmental gradient within the set of ecosystems we sampled, indicating small differences along a soil moisture gradient and less so along a soil fertility and light availability gradient. The groups were applied successfully for the majority of an area that had been repeatedly logged and/or burned for at least 120 years. In addition, the groups were successful when tested on sites with highly disturbed upper soil strata furrowed for the establishment of plantations between the 1960s and 1980s. Within the boundaries of the regional ecosystems for which they were developed, ecological species groups reflect the integrated effects of multiple site factors that control the height growth rates of jack pine trees that, in turn, determine the duration of Kirtland's warbler occupancy. While usually applied at the scale of ecosystem types, our results demonstrate that ecological species groups may also provide the ecological basis for distinguishing ecosystems at broader scales. When examined simultaneously in the field with physiography, microclimate, and soil factors, the groups are therefore useful in identifying and classifying ecosystem units at the scale of landforms, the appropriate scale of management for the Kirtland's warbler.     

Arg16/Gly beta2-adrenergic receptor polymorphism alters the cardiac output response to isometric exercise.

Normotensive adults homozygous for glycine (Gly) of the Arg16/Gly beta2-adrenergic-receptor polymorphism have 1) greater forearm beta2-receptor mediated vasodilation and 2) a higher heart rate (HR) response to isometric handgrip than arginine (Arg) homozygotes. To test the hypothesis that the higher HR response in Gly16 subjects serves to maintain the pressor response [increased cardiac output (CO)] in the setting of augmented peripheral vasodilation to endogenous catecholamines, we measured continuous HR (ECG), arterial pressure (Finapres), and CO (transthoracic echocardiography) during isometric, 40% submaximal handgrip to fatigue in healthy subjects homozygous for Gly (n = 30; mean age +/- SE: 30 +/- 1.2, 13 women) and Arg (n = 17, age 30 +/- 1.6, 11 women). Resting data were similar between groups. Handgrip produced similar increases in arterial pressure and venous norepinephrine and epinephrine concentrations; however, HR increased more in the Gly group (60.1 +/- 4.3% increase from baseline vs. 45.5 +/- 3.9%, P = 0.03), and this caused CO to be higher (Gly: 7.6 +/- 0.3 l/m vs. Arg: 6.5 +/- 0.3 l/m, P = 0.03), whereas the decrease in systemic vascular resistance in the Gly group did not reach significance (P = 0.09). We conclude that Gly16 homozygotes generate a higher CO to maintain the pressor response to handgrip. The influence of polymorphic variants in the beta2-adrenergic receptor gene on the cardiovascular response to sympathoexcitation may have important implications in the development of hypertension and heart failure.     

Genes,agents and the institution of responsible action

The concern of this paper is with how the accounts of human beings and their behaviour now emerging from genetics, genomics and the new human biotechnology should be related to traditional accounts in which we identify ourselves as responsible agents, capable of choice, who normally act freely and voluntarily. The paper addresses these apparently competing accounts in terms of their functions and modes of use, and thereby arrives at a general solution to this current version of the ancient problem of free will and determinism. The causal scientific discourse of genetics and the everyday discourse of responsibility and choice do different things for us, it suggests, and should not be regarded as articulating conflicting theories. Whilst the former is oriented to the task of naturalistic explanation, the latter is predominantly, if not entirely, a medium of communication through which we affect each other and thereby mutually regulate our conduct. If this is indeed the case, then interesting implications follow concerning the proper relationship of the two kinds of account, which need no longer be regarded as incompatible with each other. And insights emerge into both the limitations and the profound importance of the contribution that genetics and genomics seem destined to make to the understanding of human behaviour.     

Plant competition for light analyzed with a multispecies canopy model

Summary Competition for light among species in a mixed canopy can be assessed quantitatively by a simulation model which evaluates the importance of different morphological and photosynthetic characteristics of each species. A model was developed that simulates how the foliage of all species attenuate radiation in the canopy and how much radiation is received by foliage of each species. The model can account for different kinds of foliage (leaf blades, stems, etc.) for each species. The photosynthesis and transpiration for sunlit and shaded foliage of each species is also computed for different layers in the canopy. The model is an extension of previously described single-species canopy photosynthesis simulation models. Model predictions of the fraction of foliage sunlit and interception of light by sunlit and shaded foliage for monoculture and mixed canopies of wheat (Triticum aestivum) and wild oat (Avena fatua) in the field compared very well with measured values. The model was used to calculate light interception and canopy photosynthesis for both species of wheat/wild oat mixtures grown under normal solar and enhanced ultraviolet-B (290–320 nm) radiation (UV-B) in a glasshouse experiment with no root competition. In these experiments, measurements showed that the mixtures receiving enhanced UV-B radiation had a greater proportion of the total foliage area composed of wheat compared to mixtures in the control treatments. The difference in species foliage area and its position in the canopy resulted in a calculated increase in the portion of total canopy radiation interception and photosynthesis by wheat. This, in turn, is consistent with greater canopy biomass of wheat reported in canopies irradiated with supplemental UV-B.     

Conservation in a cup of water: estimating biodiversity and population abundance from environmental DNA

Three mantras often guide species and ecosystem management: (i) for preventing invasions by harmful species, ‘early detection and rapid response’; (ii) for conserving imperilled native species, ‘protection of biodiversity hotspots’; and (iii) for assessing biosecurity risk, ‘an ounce of prevention equals a pound of cure.’ However, these and other management goals are elusive when traditional sampling tools (e.g. netting, traps, electrofishing, visual surveys) have poor detection limits, are too slow or are not feasible. One visionary solution is to use an organism’s DNA in the environment (eDNA), rather than the organism itself, as the target of detection. In this issue of Molecular Ecology, Thomsen et al. (2012) provide new evidence demonstrating the feasibility of this approach, showing that eDNA is an accurate indicator of the presence of an impressively diverse set of six aquatic or amphibious taxa including invertebrates, amphibians, a fish and a mammal in a wide range of freshwater habitats. They are also the first to demonstrate that the abundance of eDNA, as measured by qPCR, correlates positively with population abundance estimated with traditional tools. Finally, Thomsen et al. (2012) demonstrate that next‐generation sequencing of eDNA can quantify species richness. Overall, Thomsen et al. (2012) provide a revolutionary roadmap for using eDNA for detection of species, estimates of relative abundance and quantification of biodiversity.     

The MIF Antagonist ISO-1 Attenuates Corticosteroid-Insensitive Inflammation and Airways Hyperresponsiveness in an Ozone-Induced Model of COPD

Introduction

Macrophage migration inhibitory factor (MIF) is an inflammatory cytokine associated with acute and chronic inflammatory disorders and corticosteroid insensitivity. Its expression in the airways of patients with chronic obstructive pulmonary disease (COPD), a relatively steroid insensitive inflammatory disease is unclear, however.

Methods

Sputum, bronchoalveolar lavage (BAL) macrophages and serum were obtained from non-smokers, smokers and COPD patients. To mimic oxidative stress-induced COPD, mice were exposed to ozone for six-weeks and treated with ISO-1, a MIF inhibitor, and/or dexamethasone before each exposure. BAL fluid and lung tissue were collected after the final exposure. Airway hyperresponsiveness (AHR) and lung function were measured using whole body plethysmography. HIF-1α binding to the Mif promoter was determined by Chromatin Immunoprecipitation assays.

Results

MIF levels in sputum and BAL macrophages from COPD patients were higher than those from non-smokers, with healthy smokers having intermediate levels. MIF expression correlated with that of HIF-1α in all patients groups and in ozone-exposed mice. BAL cell counts, cytokine mRNA and protein expression in lungs and BAL, including MIF, were elevated in ozone-exposed mice and had increased AHR. Dexamethasone had no effect on these parameters in the mouse but ISO-1 attenuated cell recruitment, cytokine release and AHR.

Conclusion

MIF and HIF-1α levels are elevated in COPD BAL macrophages and inhibition of MIF function blocks corticosteroid-insensitive lung inflammation and AHR. Inhibition of MIF may provide a novel anti-inflammatory approach in COPD.     

Reproductive and metabolic hormones during estrus after fasting in Holstein heifers

The estrous cycles of 23 Holstein heifers were synchronized with three prostaglandin F₂α (PG) injections at 0600 h 11 d apart, designated as Days ?11, 0 and 11. Twelve of the animals were randomly assigned to receive no solid food (Group F) from Day 6 to 14, while the other animals remained on full feed to serve as controls (Group C). Jugular blood samples were collected at 6-h intervals beginning with PG injection at 0600 h on Day 0 until 1800 h on Day 4 and at 0600, 1200 and 1700 h on Day 8 through 10. Samples were collected again at 6-h intervals from PG Day 11 (0600 h) until 1800 h on Day 15. Period 1 was defined as those samples collected from Day 0 through 4.5, Period 2 from Day 7 through 10, Period 3 from Day 11 through 14.25, and Period 4 from Day 14.5 through 15. Plasma growth hormone concentrations were increased (P<0.01) in F as compared with C animals during Periods 2, 3 and 4. Plasma concentrations of prolactin (P<0.01) were decreased in F as compared with C animals during Periods 2 and 3. Plasma urea concentrations were increased (P<0.01) in F as compared with C animals during the first 3 d of the fast (Period 2) but were decreased (P<0.01) during the remainder of the experiment (Periods 3 and 4). Thus, fasting was effective in altering several metabolic parameters. Although plasma progesterone and luteinizing hormone (LH) concentrations remained similar (P>0.05) between F and C animals, plasma estradiol-17β concentrations decreased in F as compared with C animals during Periods 2, 3 and 4. No differences (P>0.05) between F and C animals were found in duration to LH peak after PG injection, estrous behavior, or pregnancy rates. Results from this study indicate that fasting reduced plasma estradiol-17β concentrations during estrus but did not alter occurrence of estrus or pregnancy rate.     

Plant genome archaeology: evidence for conserved ancestral chromosome segments in dicotyledonous plant species

We have developed genetic maps, based on expressed sequence tags (ESTs) that are homologous to Arabidopsis genes, in four dicotyledonous crop plant species from different families. A comparison of these maps with the physical map of Arabidopsis reveals common genome segments that appear to have been conserved throughout the evolution of the dicots. In the four crop species analysed these segments comprise between 16 and 33% of the Arabidopsis genome. Our findings extend the synteny patterns previously observed only within plant families, and indicate that structural and functional information from the model species will be, at least in part, applicable in crop plants with large genomes.