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.     

Sequencing,Annotation, and Characterization of the Influenza Ferret Infectome

Ferrets have become an indispensable tool in the understanding of influenza virus virulence and pathogenesis. Furthermore, ferrets are the preferred preclinical model for influenza vaccine and therapeutic testing. Here we characterized the influenza infectome during the different stages of the infectious process in ferrets with and without prior specific immunity to influenza. RNA from lung tissue and lymph nodes from infected and naïve animals was subjected to next-generation sequencing, followed by de novo data assembly and annotation of the resulting sequences; this process generated a library comprising 13,202 ferret mRNAs. Gene expression profiles during pandemic H1N1 (pdmH1N1) influenza virus infection were analyzed by digital gene expression and solid support microarrays. As expected during primary infection, innate immune responses were triggered in the lung tissue; meanwhile, in the lymphoid tissue, genes encoding antigen presentation and maturation of effector cells of adaptive immunity increased dramatically. After 5 days postinfection, the innate immune gene expression was replaced by the adaptive immune response, which correlates with viral clearance. Reinfection with homologous pandemic influenza virus resulted in a diminished innate immune response, early adaptive immune gene regulation, and a reduction in clinical severity. The fully annotated ferret infectome will be a critical aid to the understanding of the molecular events that regulate disease severity and host-influenza virus interactions among seasonal, pandemic, and highly pathogenic avian influenzas.     

Genetic relatedness does not retain spatial pattern across multiple spatial scales: dispersal and colonization in the coral,Pocillopora damicornis

Patterns of isolation by distance are uncommon in coral populations. Here, we depart from historical trends of large‐scale, geographical genetic analyses by scaling down to a single patch reef in Kāne‘ohe Bay, Hawai‘i, USA, and map and genotype all colonies of the coral, Pocillopora damicornis. Six polymorphic microsatellite loci were used to assess population genetic and clonal structure and to calculate individual colony pairwise relatedness values. Our results point to an inbred, highly clonal reef (between 53 and 116 clonal lineages of 2352 genotyped colonies) with a much skewed genet frequency distribution (over 70% of the reef was composed of just seven genotypes). Spatial autocorrelation analyses revealed that corals found close together on the reef were more genetically related than corals further apart. Spatial genetic structure disappears, however, as spatial scale increases and then becomes negative at the largest distances. Stratified, random sampling of three neighbouring reefs confirms that reefs are demographically open and inter‐reef genetic structuring was not detected. Attributing process to pattern in corals is complicated by their mixed reproductive strategies. Separate autocorrelation analyses, however, show that the spatial distribution of both clones and nonclones contributes to spatial genetic structure. Overall, we demonstrate genetic structure on an intrareef scale and genetic panmixia on an inter‐reef scale indicating that, for P. damicornis, the effect of small‐ and large‐scale dispersal processes on genetic diversity are not the same. By starting from an interindividual, intrareef level before scaling up to an inter‐reef level, this study demonstrates that isolation‐by‐distance patterns for the coral P. damicornis are limited to small scales and highlights the importance of investigating genetic patterns and ecological processes at multiple scales.     

The anabolic action of intermittent parathyroid hormone on cortical bone depends partly on its ability to induce nitric oxide‐mediated vasorelaxation in BALB/c mice

There is strong evidence that vasodilatory nitric oxide (NO) donors have anabolic effects on bone in humans. Parathyroid hormone (PTH), the only osteoanabolic drug currently approved, is also a vasodilator. We investigated whether the NO synthase inhibitor L‐NAME might alter the effect of PTH on bone by blocking its vasodilatory effect. BALB/c mice received 28 daily injections of PTH[1–34] (80 µg/kg/day) or L‐NAME (30 mg/kg/day), alone or in combination. Hindlimb blood perfusion was measured by laser Doppler imaging. Bone architecture, turnover and mechanical properties in the femur were analysed respectively by micro‐CT, histomorphometry and three‐point bending. PTH increased hindlimb blood flow by >30% within 10 min of injection (P < 0.001). Co‐treatment with L‐NAME blocked the action of PTH on blood flow, whereas L‐NAME alone had no effect. PTH treatment increased femoral cortical bone volume and formation rate by 20% and 110%, respectively (P < 0.001). PTH had no effect on trabecular bone volume in the femoral metaphysis although trabecular thickness and number were increased and decreased by 25%, respectively. Co‐treatment with L‐NAME restricted the PTH‐stimulated increase in cortical bone formation but had no clear‐cut effects in trabecular bone. Co‐treatment with L‐NAME did not affect the mechanical strength in femurs induced by iPTH. These results suggest that NO‐mediated vasorelaxation plays partly a role in the anabolic action of PTH on cortical bone. © 2016 The Authors. Cell Biochemistry and Function published by John Wiley & Sons, Ltd.     

Repeated immunostaining of the same tissue section using alkaline phosphatase as a reporter

One can determine the best dilution of a primary antibody for immunohistochemistry that uses horseradish peroxidase conjugated to a secondary antibody by testing increasing concentrations sequentially on the same tissue section. When the same tissue section is incubated repeatedly with increasing concentrations of primary antibodies to epithelial membrane antigen, smooth muscle α-actin, or vimentin using alkaline phosphatase conjugated to a secondary antibody as the reporter, the best staining was obtained with a less concentrated primary antibody than was optimal for a single staining test. The best concentration of primary antibody for single run staining using an alkaline phosphatase reporting system is usually four times the best concentration for staining with multiple runs. The optimal concentration can be determined by denaturing the residual alkaline phosphatase and extracting residual stain by incubating the section in 4:1 diglyme:phosphate buffered saline for 20 min at 80^o C between tests of primary antibody concentrations. I tested the method for four chromogens from one supplier and one chromogen from a different supplier.     

Differential responses by two closely related native fishes to restoration actions

Globally, river degradation has decimated freshwater fish populations. To help reverse this trend in a southeastern Australia river, we used multiple restoration actions, including reintroduction of instream woody habitat, riparian revegetation, removal of a weir hindering fish movement, fencing out livestock, and controlling riparian weeds. We monitored the responses of native fish at the segment scale (20 km) and reach scale (0.3 km) over 7 years to assess the effectiveness of the different restoration strategies. Two closely related species, Murray cod Maccullochella peeli and trout cod Maccullochella macquariensis, increased at the restored segment compared with the control segment. However, inherent differences between river segments and low sample size hampered assessment of the mechanisms responsible for segment‐scale changes in fish abundance. In contrast, at the reach scale, only M. peeli abundance significantly increased in reaches supplemented with wood. These differential responses by 2 closely related fish species likely reflect species‐specific responses to increased habitat availability and enhanced longitudinal connectivity when the weir improved passage around a fishway. Changes in M. peeli abundance in segments supplemented with and without wood suggest an increase in carrying capacity and not simply a redistribution of individuals within the segment, facilitated the observed expansion. Our findings confirm the need to consider individual fish species' habitat preferences carefully when designing restoration interventions. Further, species‐specific responses to restoration actions provide waterway managers with precise strategies to target fish species for recovery and the potential to predict fish outcomes based on ecological preferences.     

Modulating biochemical perturbations during 72-hour machine perfusion of kidneys: role of preservation solution

This study documents renal biochemistry during hypothermic machine perfusion of kidneys. It is intended to demonstrate that a comprehensive evaluation of organ viability during ex-vivo preservation is needed to increase the number of organs available for transplantation and to reduce the current renal discard rate. Porcine kidneys were hypothermically machine perfused for 72 h with either Unisol-UHK or Belzer-Machine Perfusion Solution, (Belzer-MPS). Renal perfusate samples were periodically collected and biochemically analyzed. Significant differences were measured in the renal metabolic activity between the two experimental groups while similar values for traditional parameters such as renal flow rate and vascular resistance values were recorded. The effluent of UHK perfused kidneys showed strong metabolites and NH(4)(+) dynamics (P<0.05 vs. baseline), while the Belzer-MPS kidneys metabolic activity led to little or no change of the effluent biochemistry relative to baseline.     

O-fucosylation is required for ADAMTS13 secretion

ADAMTS13 is a plasma metalloproteinase that cleaves von Willebrand factor to smaller, less thrombogenic forms. Deficiency of ADAMTS13 activity in plasma leads to thrombotic thrombocytopenic purpura. ADAMTS13 contains eight thrombospondin type 1 repeats (TSR), seven of which contain a consensus sequence for the direct addition of fucose to the hydroxyl group of serine or threonine. Mass spectral analysis of tryptic peptides derived from human ADAMTS13 indicate that at least six of the TSRs are modified with an O-fucose disaccharide. Analysis of [(3)H]fucose metabolically incorporated into ADAMTS13 demonstrated that the disaccharide has the structure glucose-beta1,3-fucose. Mutation of the modified serine to alanine in TSR2, TSR5, TSR7, and TSR8 reduced the secretion of ADAMTS13. Mutation of more than one site dramatically reduced secretion regardless of the sites mutated. When the expression of protein O-fucosyltransferase 2 (POFUT2), the enzyme that transfers fucose to serines in TSRs, was reduced using siRNA, the secretion of ADAMTS13 decreased. A similar outcome was observed when ADAMTS13 was expressed in a cell line unable to synthesize the donor for fucose addition, GDP-fucose. Although overexpression of POFUT2 did not affect the secretion of wild-type ADAMTS13, it did increase the secretion of the ADAMTS13 TSR1,2 double mutant but not that of ADAMTS13 TSR1-8 mutant. Together these findings indicate that O-fucosylation is functionally significant for secretion of ADAMTS13.