MLMSeg: A multi-view learning model for ultrasound thyroid nodule segmentation.

Accurate segmentation of the thyroid gland in ultrasound images is an essential initial step in distinguishing between benign and malignant nodules, thus facilitating early diagnosis. Most existing deep learning-based methods to segment thyroid nodules are learned from only a single view or two views, which limits the performance of segmenting nodules at different scales in complex ultrasound scanning environments. To address this limitation, this study proposes a multi-view learning model, abbreviated as MLMSeg. First, a deep convolutional neural network is introduced to encode the features of the local view. Second, a multi-channel transformer module is designed to capture long-range dependency correlations of global view between different nodules. Third, there are semantic relationships of structural view between features of different layers. For example, low-level features and high-level features are endowed with hidden relationships in the feature space. To this end, a cross-layer graph convolutional module is proposed to adaptively learn the correlations of high-level and low-level features by constructing graphs across different layers. In addition, in the view fusion, a channel-aware graph attention block is devised to fuse the features from the aforementioned views for accurate segmentation of thyroid nodules. To demonstrate the effectiveness of the proposed method, extensive comparative experiments were conducted with 14 baseline methods. MLMSeg achieved higher Dice coefficients (92.10% and 83.84%) and Intersection over Union scores (86.60% and 73.52%) on two different thyroid datasets. The exceptional segmentation capability of MLMSeg for thyroid nodules can greatly assist in localizing thyroid nodules and facilitating more precise measurements of their transverse and longitudinal diameters, which is of significant clinical relevance for the diagnosis of thyroid nodules.

Improvement of the oral bioavailability of coenzyme Q10 with lecithin nanocapsules.

Coenzyme Q10-loaded lecithin nanocapsules (CoQ10-LNCs), composed of a CoQ10/lecithin/ GTCC/glycerol aqueous solution, were prepared by high-pressure homogenization. The zeta potential of the CoQ10-LNCs above -60 mV was determined on a Malvern Zetasize 2000 (Malvern Instruments, UK). The spherical shape of the CoQ10-LNCs was observed by using freeze-fracture transmission electron microscopy (FF-TEM), and the particle size was found to be below 100 nm. The supercooled state of the CoQ10-LNCs was observed by differential scanning calorimetry (DSC). In an oral bioavailability study, the CoQ10 plasma level after administering CoQ10-LNCs was higher than that after administering a CoQ10 tablet over 24 hours, and the relative bioavailability of CoQ10 was improved to 176.6% in mice. Based on the above results, the LNC delivery system might be a potential vehicle for improving the oral bioavailability of CoQ10.

Dual Attention Adversarial Attacks With Limited Perturbations.

The construction of undetectable adversarial examples with few perturbances remains a difficult problem in adversarial attacks. At present, most solutions use the standard gradient optimization algorithm to build adversarial examples by applying global perturbations to benign samples and then launch attacks on the targets (e.g., face recognition systems). However, when the perturbance size is limited, the performance of these approaches suffers substantially. The content of crucial places in an image, on the other hand, will impact the final prediction; if these areas can be investigated and limited perturbances introduced, an acceptable adversarial example will be constructed. Based on the foregoing research, this article offers a dual attention adversarial network (DAAN) to produce adversarial examples with limited perturbations. DAAN initially searches for effective areas in an input image using the spatial attention network and channel attention network, and then creates space and channel weights. Following that, these weights direct an encoder and a decoder to generate effective perturbation, which is then combined with the input to produce an adversarial example. Finally, the discriminator determines if the created adversarial examples are true or false, and the attacked model is utilized to determine whether the generated samples fit the attack targets. Extensive studies on various datasets show that DAAN not only delivers the best attack performance across all comparison algorithms with few perturbations, but it can also significantly improve the defensiveness of the attacked models.

The End-to-end Fetal Head Circumference Detection and Estimation in Ultrasound Images.

In prenatal examinations, the fetal head circumference (HC) measurement is essential for assessing fetal weight and health conditions. The sonographers obtain the fetal HC manually by fitting peripheral skull ellipse in clinical practice, which is highly subjective, time-consuming, and experience-dependent. Recently, many fetal HC automatic measurement algorithms have been proposed to improve workflow efficiency in prenatal examination. But most automatic measurement algorithms focus on using fetal head segmentation as an intermediate processing step, and HC estimation relies heavily on segmentation results, which causes the accumulation of errors in the above two stages. Independent of the segmentation method, we design a regression network to generate the oriented bounding box to detect the head contour, and directly obtain the fetal head parameters with a pixel-based ellipse regression (PER) loss. Moreover, an effective 3D attention mechanism is integrated into the network to estimate HC more precisely without adding parameters in complex ultrasound images. The extensive experimental results on the public HC18 and our clinical dataset show that the proposed network provides a feasible scheme for end-to-end estimating fetal HC, and avoids the mistake brought by the intermediary processes.

Characterization of a novel annexin gene from cotton (Gossypium hirsutum cv CRI 35) and antioxidative role of its recombinant protein.

Plant annexins represent a multigene family involved in cellular elongation and development. A cDNA encoding a novel annexin was isolated from a cotton (Gossypium hirsutum) fiber cDNA library and designated GhAnx1. This gene encodes a 316 amino acid protein with a theoretical molecular mass of 36.06 kDa and a theoretical pI of 6.19. At the amino acid level, it shares high sequence similarity and has evolutionary relationships with annexins from higher plants. The purified recombinant protein expressed in Escherichia coli was used to investigate its physicochemical properties. Circular dichroism spectrum analyses showed a positive peak rising to the maximum at 196 nm and a broad negative band rounding 215 nm, suggesting that the GhAnx1 protein was prominently α-helical. The fluorescence measurements indicated that it could bind to Ca(2+) in vitro. These results demonstrated that GhAnx1 was a typical annexin protein in cotton. A bioassay experiment was conducted to analyze its potential function and showed that E. coli cells expressing GhAnx1 were protected from tert-butyl hydroperoxide (tBH) stress, suggesting that it had a potential antioxidative role. Northern blot analyses revealed that GhAnx1 was highly expressed in fibers, especially during the elongation stage, suggesting that it might be important for fiber elongation.

New evidence suggests Southern China as a common source of multiple clusters of highly pathogenic H5N1 avian influenza virus.

Highly pathogenic H5N1 avian influenza is considered an avian disease, although there is some evidence of limited human-to-human transmission of the virus. A global effort is underway to control or eradicate the highly pathogenic H5N1 avian influenza virus in poultry and prevent human exposure, both of which may also reduce the risk of pandemic emergence. Hemagglutinin gene sequences from 215 human H5N1 influenza viruses were used to trace the source and dispersal pattern of human H5N1 influenza viruses on a global scale. A mutation network and phylogenetic analyses of the hemagglutinin gene show that human H5N1 influenza viruses can be clearly divided among 4 clusters across geographic space. On the basis of analysis of the N-glycosylation sites at positions 100 and 170 in the hemagglutinin protein, human H5N1 influenza viruses were also divided into 3 types. When we combined these analyses with geographic information system data analyses, we found that Southern China is often a common source of multiple clusters of H5N1 influenza viruses and that each cluster has different dispersal patterns and individual evolutionary features. In summary, the genetic evidence presented here provides clear evidence for multiple clusters of human H5N1 influenza viruses that initially originated in Southern China.

In situ hydrodeoxygenation of vanillin over Ni-Co-P/HAP with formic acid as a hydrogen source.

A new noble metal-free Ni-Co-P/HAP (hydroxyapatite) amorphous alloy catalyst was synthesized by an impregnation-chemical reduction method; the structure and properties of the catalysts were characterized by XRD, SEM, BET, XPS and DSC. Based on the model of the hydrodeoxygenation (HDO) of vanillin to 2-methoxy-4-methylphenol (MMP) with formic acid as a hydrogen source, the catalytic performance of the catalyst was studied. The results found that the Ni-Co-P/HAP catalyst exhibited excellent catalytic activity for the in situ HDO reaction of vanillin compared with Ni-P and Ni-Co-P. The conversion of vanillin could be high to 97.86% with MMP selectivity of 93.97% under optimized reaction conditions. In addition, mechanism studies have shown that the side reaction of carbocation and vanillyl alcohol (HMP) condensation can be effectively reduced with increasing the hydrogenation rate, thereby the selectivity of MMP was effectively increased.

Selenium deficiency impairs host innate immune response and induces susceptibility to Listeria monocytogenes infection.

BACKGROUND: Susceptibility or resistance to infection with Listeria monocytogenes correlates with Selenium (Se) deficiency in response to infection. RESULTS: Se-deficient mouse models of listeriosis were used to study the innate immune response during the course of L. monocytogenes infection. Blood samples from mouse models were used for Se status. The concentration of MDA, SOD, GPx and CAT in blood has revealed that lower Se level exist in Se-deficient mice. Intestine, mesenteric lymph node, liver, spleen and brain from each mouse were to study the bacterial burden in organs. The analysis of cell types of spleen from Se-deficient mice revealed that the ability of the host to elicit a rapid recruitment and activation of systemic innate immune response to infection was to a certain extent compromised under conditions of Se deficiency. The cytokine levels in the serum and cytokine expression levels in the livers from Se-deficient mice revealed that the innate immune response of Se-deficient mice was impaired throughout the course of infection. These results suggest that innate immune response is altered by Se deficiency after infection with L. monocytogenes. CONCLUSION: In conclusion, induced susceptibility of host resistance is associated with an impaired innate immune response following infection with L. monocytogenes in C57BL/6 Se-deficient mice.

Construction of effective receptor for recognition of avian influenza H5N1 protein HA1 by assembly of monohead glycolipids on polydiacetylene vesicle surface.

Fast and sensitive detection of epidemic virus is of the utmost importance for human being in nowadays. Various biosensors have been designed for this goal based on conjugation event between host cell glycolipids and invading virus. However, multihead glycolipids analogous to native receptors on cell surface are known to be very difficult to mimic because of the complexity of chemical synthesis. Here, we developed a new approach where two types of monohead glycolipids, active sialic acid-beta-glucoside (G1) and inactive lactose-beta-glucoside (G2), are embedded onto the surface of a polydiacetylene (PDA) vesicle to mimic native glycolipids on the cell surface. Vesicles prepared in this manner show good selectivity with a 10 ng/mL detection limit and 5 min response time to Hemagglutinin (HA1), which is more sensitive than any HA1 biosensors ever known. Moreover, in the formation of color-changeable vesicles, a very strong synergistic effect between G1 and G2 has been found, offering a novel strategy to construct effective biosensor receptors, as well as a new way to study the surface combination effect that is potentially important to the immunology study of epidemic disease.

Development and evaluation of a recombinant CP23 antigen-based ELISA for serodiagnosis of Cryptosporidium parvum.

The CP23 gene of Cryptosporidium parvum strain isolated from Changchun in China was expressed in Escherichia coli BL21 strain and was purified as a recombinant protein. An indirect ELISA assay (CP23-ELISA) for antibody detection was established using the purified recombinant CP23 protein. Antigen coating conditions and serum dilution for the CP23-ELISA were optimized. The S/P ratio of the absorbency value was calculated in the CP23-ELISA to evaluate the serum antibody level of the field cow samples. It indicated that the CP23-ELISA assay, which was more convenient and easier to prepare than traditional methods, was a good candidate for evaluation of C. parvum exposure to domestic animal in field.

Computing Time-Varying Quadratic Optimization With Finite-Time Convergence and Noise Tolerance: A Unified Framework for Zeroing Neural Network.

Zeroing neural network (ZNN), as a powerful calculating tool, is extensively applied in various computation and optimization fields. Convergence and noise-tolerance performance are always pursued and investigated in the ZNN field. Up to now, there are no unified ZNN models that simultaneously achieve the finite-time convergence and inherent noise tolerance for computing time-varying quadratic optimization problems, although this superior property is highly demanded in practical applications. In this paper, for computing time-varying quadratic optimization within finite-time convergence in the presence of various additive noises, a new framework for ZNN is designed to fill this gap in a unified manner. Specifically, different from the previous design formulas either possessing finite-time convergence or possessing noise-tolerance performance, a new design formula with finite-time convergence and noise tolerance is proposed in a unified framework (and thus called unified design formula). Then, on the basis of the unified design formula, a unified ZNN (UZNN) is, thus, proposed and investigated in the unified framework of ZNN for computing time-varying quadratic optimization problems in the presence of various additive noises. In addition, theoretical analyses of the unified design formula and the UZNN model are given to guarantee the finite-time convergence and inherent noise tolerance. Computer simulation results verify the superior property of the UZNN model for computing time-varying quadratic optimization problems, as compared with the previously proposed ZNN models.

RNA interference of avian influenza virus H5N1 by inhibiting viral mRNA with siRNA expression plasmids.

Avian influenza virus H5N1 causes widespread infection in the birds and human respiratory tract, but existing vaccines and drug therapy are of limited value. Here we show that small interfering RNAs (siRNAs) specific for conserved regions of the viral genome can potently inhibit influenza virus production in cell lines, embryonated chicken eggs and BALB/c mice. siRNA expression plasmid pBabe-Super was chosen in the study, which directed the synthesis of small interfering RNAs in cells. The inhibition depended on the presence of a functional antisense strand in the small interfering RNA duplex, suggesting that viral mRNA is the target of RNA interference (RNAi). Among the three small interfering RNA expression plasmids we designed, we found that small interfering RNA for nucleocapsid protein (NP) had a specific effect in inhibiting the accumulation of RNAs in infected cells because of a critical requirement for newly synthesized nucleocapsid proteins in avian influenza viral RNA transcription and replication. The findings reveal that newly synthesized nucleocapsid, polymerase A (PA) and polymerase B1 (PB1) proteins are required for avian influenza virus transcription and replication and provide a basis for the development of small interfering RNAs as prophylaxis and therapy for avian influenza infection in birds and humans.

A Parallel Multiclassification Algorithm for Big Data Using an Extreme Learning Machine.

As data sets become larger and more complicated, an extreme learning machine (ELM) that runs in a traditional serial environment cannot realize its ability to be fast and effective. Although a parallel ELM (PELM) based on MapReduce to process large-scale data shows more efficient learning speed than identical ELM algorithms in a serial environment, some operations, such as intermediate results stored on disks and multiple copies for each task, are indispensable, and these operations create a large amount of extra overhead and degrade the learning speed and efficiency of the PELMs. In this paper, an efficient ELM based on the Spark framework (SELM), which includes three parallel subalgorithms, is proposed for big data classification. By partitioning the corresponding data sets reasonably, the hidden layer output matrix calculation algorithm, matrix decomposition algorithm, and matrix decomposition algorithm perform most of the computations locally. At the same time, they retain the intermediate results in distributed memory and cache the diagonal matrix as broadcast variables instead of several copies for each task to reduce a large amount of the costs, and these actions strengthen the learning ability of the SELM. Finally, we implement our SELM algorithm to classify large data sets. Extensive experiments have been conducted to validate the effectiveness of the proposed algorithms. As shown, our SELM achieves an speedup on a cluster with ten nodes, and reaches a speedup with 15 nodes, an speedup with 20 nodes, a speedup with 25 nodes, a speedup with 30 nodes, and a speedup with 35 nodes.

Construction of a eukaryotic expression plasmid encoding partial S gene fragments of the SARS-CoV and its potential utility as a DNA vaccine.

The spike (S) protein, a main surface antigen of the SARS coronavirus (SARS-CoV), is considered to be one of the most important protective antigen candidates for targets for vaccine design against the virus. In this study, a secreted recombinant expression plasmid, pVAX-S1, encoding the partial S protein with a signal peptide, was constructed and used to immunize BALB/c mice through electroporation. It was demonstrated that the eukaryotic expression vector pVAX-S1 was successfully constructed by restriction enzyme and sequence analysis. The expressed protein could be detected specifically by Western blot analysis. The serum IgG level of the vaccine group mice was significantly higher than that of the corresponding control group at day 14 after vaccination (P < 0.05). The vaccine group demonstrated significantly higher S1 protein lymphocyte proliferation index (LPI) than the control groups (P < 0.05). Furthermore, in the experimental group, a decrease in the ratio of CD4(+) to CD8(+) T-lymphocytes and an increase level of IFN-gamma in serum were observed. However, interleukin-4 (IL-4) was not detectable in two groups. These results strongly demonstrated that the pVAX-S1 plasmid could induce humoral and cellular immune responses in mice, and may be a potential candidate for a DNA vaccine against the SARS coronavirus.

The humoral and cellular immune responses in mice induced by DNA vaccine expressing the sporozoite surface protein of Cryptosporidium parvum.

Cryptosporidiosis, a protozoan disease, is caused by Cryptosporidium parvum in animals and humans. To study the humoral and cellular immune responses induced by DNA vaccine expressing the sporozoite surface protein, CP15/60, of Cryptosporidium parvum, the recombinant plasmid containing the CP15/60 gene was injected into tibialis a interior muscle of BALB/c mice. The mice were subsequently given booster doses twice at 3-week intervals. The humoral and cellular immune responses were detected at different times after immunization. The mice were then challenged by inoculation of 1 x 10(6) oocysts of C. parvum. The experimental results have shown that the recombinant plasmid can induce corresponding specific immune responses and thus protect the mice from challenge of the oocysts, suggesting that the recombinant plasmid could be a potential candidate of DNA vaccine.

Rapid quantitative determination and assessment of insulin in oil formulation by micellar electrokinetic capillary chromatography.

A rapid, simple and precise micellar electrokinetic capillary chromatrography (MEKC) with diode array detector had been developed to the determination of insulin in oil formulation. As the micelle-forming ampholytic surfactant, Labrasol was chosen for the separation and analysis of the hydrophobic insulin in oil formulation in this experiment. A buffer containing 10 mmol/ml Tris-HCl plus 10% acetonitrile (ACN) at pH 8.2 was used for running buffer. Samples and tested samples were successfully separated during 8 min with a good linear range(r>0.99) and a concentration detection limit of 0.01 mg/ml. Furthermore, quantitative determinations of insulin in oil formulation had been described and assessed after handled with different pre-treatment conditions to reveal the effect of pH, temperature and tryptic digestion.

Detection of drug-resistant Klebsiella pneumoniae in Chinese hares (Lepus sinensis).

We investigated an outbreak of acute pneumonia among adult Chinese hares (Lepus sinensis) and diarrhea among juvenile hares in Hebei Province, China, in 2012. Diagnosis was based on necropsy examination, microbial characteristics, biochemical identification, and nucleotide sequence analysis. The isolated bacteria from tissue samples of dead hares were identified as Klebsiella pneumoniae ssp. pneumoniae (K. pneumoniae). This K. pneumoniae was resistant to the antimicrobials imipenem, meropenem, penicillin, and vancomycin, but was highly sensitive to cefepime, cotrimoxazole, and enrofloxacin. Klebsiella pneumoniae is an important opportunistic pathogen, which often causes nosocomial infections in immunocompromised patients. However, the emergence of drug-resistant K. pneumoniae in hares indicates the existence of increasing risk of pathogen transmission between humans and wildlife. Given the close association between wildlife, livestock, and humans, it is important to identify epidemiologic factors associated with infection in these hares to minimize the risk of K. pneumoniae transmission.

Novel lipid-free nanoformulation for improving oral bioavailability of coenzyme Q10.

To improve the bioavailability of orally administered lipophilic coenzyme Q10 (CoQ10), we formulated a novel lipid-free nano-CoQ10 system stabilized by various surfactants. Nano-CoQ10s, composed of 2.5% (w/w) CoQ10, 1.67% (w/w) surfactant, and 41.67% (w/w) glycerol, were prepared by hot high-pressure homogenization. The resulting formulations were characterized by particle size, zeta potential, differential scanning calorimetry, and cryogenic transmission electron microscopy. We found that the mean particle size of all nano-CoQ10s ranged from 66.3 ± 1.5 nm to 92.7 ± 1.5 nm and the zeta potential ranged from -12.8 ± 1.4 mV to -41.6 ± 1.4 mV. The CoQ10 in nano-CoQ10s likely existed in a supercooled state, and nano-CoQ10s stored in a brown sealed bottle were stable for 180 days at 25 °C. The bioavailability of CoQ10 was evaluated following oral administration of CoQ10 formulations in Sprague-Dawley rats. Compared to the values observed following administration of CoQ10-Suspension, nano-CoQ10 modified with various surfactants significantly increased the maximum plasma concentration and the area under the plasma concentration-time curve. Thus, the lipid-free system of a nano-CoQ10 stabilized with a surfactant may be an effective vehicle for improving oral bioavailability of CoQ10.

Evolutionary characterization of the pandemic H1N1/2009 influenza virus in humans based on non-structural genes.

The 2009 influenza pandemic had a tremendous social and economic impact. To study the genetic diversity and evolution of the 2009 H1N1 virus, a mutation network for the non-structural (NS) gene of the virus was constructed. Strains of the 2009 H1N1 pandemic influenza A virus could be divided into two categories based on the V123I mutation in the NS1 gene: G1 (characterized as 123 Val) and G2 (characterized as 123 Ile). Sequence homology analysis indicated that one type of NS sequence, primarily isolated from Mexico, was likely the original type in this pandemic. The two genotypes of the virus presented distinctive clustering features in their geographic distributions. These results provide additional insight into the genetics and evolution of human pandemic influenza H1N1.

Preliminary evaluation of a novel oral delivery system for rhPTH1-34: in vitro and in vivo.

rhPTH1-34 is clinically used for osteoporosis treatment. However, this peptide drug has no oral bioavailability because of proteolysis and low membrane permeability in gastrointestinal gut. This study explored the possibility of absorption enhancement for rhPTH1-34 through the oral delivery of the microemulsion. The microemulsion (85:15, oil/water) consisting of Labrasol, Crodamol GTCC, Solutol HS 15, d-α-tocopheryl acetate (6:2:1:1, w/w) and saline water was developed and characterized, including particle size, morphology, drug loading efficiency and permeability, stability and pharmacokinetics. The microemulsion showed high drug loading efficiency (83%) and permeability, and significantly higher resistance to proteolysis in vitro study. The relative oral bioavailability was 5.4% and 12.0% when delivered to gastric and ileum. Besides, osteoporosis rats were induced and treated with oral rhPTH1-34 microemulsion (0.05 mg/kg), injection (0.01 mg/kg) and vehicle, respectively, for 8 weeks. The proximal tibia bone mineral content and density in oral rats (0.188 ± 0.008 g, 0.283 ± 0.014 g/cm(2)) was significantly increased compared to the control rats (0.169 ± 0.006 g, 0.266 ± 0.011 g/cm(2)), reaching to the sham rats. And the proximal tibia microstructure of oral rats was improved greatly, approaching sham level too. These findings revealed that oral microemulsion may represent an effective oral delivery system for rhPTH1-34.