Adaptive Top-K Algorithm for Medical Conversational Diagnostic Model.

With advancements in computing technology and the rapid progress of data science, machine learning has been widely applied in various fields, showing great potential, especially in digital healthcare. In recent years, conversational diagnostic systems have been used to predict diseases through symptom checking. Early systems predicted the likelihood of a single disease by minimizing the number of questions asked. However, doctors typically perform differential diagnoses in real medical practice, considering multiple possible diseases to address diagnostic uncertainty. This requires systems to ask more critical questions to improve diagnostic accuracy. Nevertheless, such systems in acute medical situations need to process information quickly and accurately, but the complexity of differential diagnosis increases the system's computational cost. To improve the efficiency and accuracy of telemedicine diagnostic systems, this study developed an optimized algorithm for the Top-K algorithm. This algorithm dynamically adjusts the number of the most likely diseases and symptoms by real-time monitoring of case progress, optimizing the diagnostic process, enhancing accuracy (99.81%), and increasing the exclusion rate of severe pathologies. Additionally, the Top-K algorithm optimizes the diagnostic model through a policy network loss function, effectively reducing the number of symptoms and diseases processed and improving the system's response speed by 1.3-1.9 times compared to the state-of-the-art differential diagnosis systems.

Adversarial Defense Method Based on Latent Representation Guidance for Remote Sensing Image Scene Classification.

Deep neural networks have made great achievements in remote sensing image analyses; however, previous studies have shown that deep neural networks exhibit incredible vulnerability to adversarial examples, which raises concerns about regional safety and production safety. In this paper, we propose an adversarial denoising method based on latent representation guidance for remote sensing image scene classification. In the training phase, we train a variational autoencoder to reconstruct the data using only the clean dataset. At test time, we first calculate the normalized mutual information between the reconstructed image using the variational autoencoder and the reference image as denoised by a discrete cosine transform. The reconstructed image is selectively utilized according to the result of the image quality assessment. Then, the latent representation of the current image is iteratively updated according to the reconstruction loss so as to gradually eliminate the influence of adversarial noise. Because the training of the denoiser only involves clean data, the proposed method is more robust against unknown adversarial noise. Experimental results on the scene classification dataset show the effectiveness of the proposed method. Furthermore, the method achieves better robust accuracy compared with state-of-the-art adversarial defense methods in image classification tasks.

Prognostic value of systemic inflammatory factors NLR, LMR, PLR and LDH in penile cancer.

BACKGROUND: Penile squamous cell carcinoma (PSCC) represents an important public health problem for developing countries. The major prognostic factors in PSCC are pathological subtype, perineural invasion, lymphovascular invasion, depth of invasion and grade, which are hard to obtain precisely before the operation. Besides, micro-metastases will be detected in about 30% of intermediate-risk patients with clinically non-palpable inguinal lymph nodes after inguinal lymph node dissection (ILND). It means approximately 70% of patients are unable to benefit from ILND who might suffered from the complications of surgery. We hope some biomarkers could be found which are able to predict the outcome before surgery and reflect the inguinal lymph nodes metastasis. METHODS: A total of 349 consecutive patients of penile cancer in Yunnan cancer hospital in China between October 2002 and December2017. Two hundred twenty-five was succeed to follow-up. The association between NLR, LMR, PLR, LDH and Overall survival (OS), progression free survival (PFS), inguinal lymph node (N stage) was analyzed with K-M analysis, univariable, multivariable logistic regression and Kendall's tau-b correlation coefficient. RESULTS: Multivariable analysis reveal that only PLR was significant independent factor which is associated with inferior OS and PFS; Age and LDH was associated with inferior OS; Lymph node and metastatic status remained significant for OS and PFS as NCCN and EAU Guidelines indicated; the tumor type, initial treatment and NLR LMR were not significant in predicting both OS and PFS. NLR, LMR and PLR were corresponded to N stage, while LDH was not associated with the N stage based on logistic regression model analysis. NLR, LMR and PLR were found weakly related to N stage through an application of Kendall's tau-b correlation coefficient. CONCLUSIONS: PLR was significant independent factors for OS and PFS, Age and LDH was significant independent factors for OS. NLR, LMR, PLR was corresponded to N stage.

Molecular characterization and expression of interleukin-10 and interleukin-22 in golden pompano (Trachinotus ovatus) in response to Streptococcus agalactiae stimulus.

In the present study, members of the interleukin (IL)-10 family of cytokines, including IL-10 (TOIL-10) and IL-22 (TOIL-22) of golden pompano (Trachinotus ovatus), were cloned for the first time, and their expression patterns and 3D structures analyzed. The full-length cDNA sequences of TOIL-10 and TOIL-22 contained open reading frames of 564 and 567 bp, respectively. TOIL-10 and TOIL-22 shared higher homology (78%-89%) with the corresponding genes from various fish relative to other species (25%-34%) and contained the IL-10 family signature and four cysteine residues that are well conserved in other vertebrate IL-10 members. Phylogenetic tree analysis of our sequences alongside other IL-10 family proteins revealed that TOIL-10 and TOIL-22 cluster together with other teleost IL-10 and IL-22 molecules. Expression of TOIL-10 and TOIL-22 genes was ubiquitous in all tissues examined. The TOIL-10 gene was also highly expressed in skin, heart, gill, spleen, kidney, brain and liver, and lower levels were detected in intestine and muscle. High expression of the TOIL-22 gene was observed in gill, intestine, kidney, spleen, with the lowest levels in liver. TOIL-10 and TOIL-22 were rapidly activated after SAΔphoB immunization and significantly increased to peak levels at 12 h and 4 d in golden pompano kidney and spleen respectively following challenge. Expression in the brain reached peak levels at 4 d and 3 d respectively after post-immunization. Our results collectively indicate that TOIL-10 and TOIL-22 participate in the host immune response to bacterial infection. Moreover, TOIL-22 plays a potentially important role in mucosal immunity.

Current insights into computer-aided immunotherapeutic design strategies.

Drug designing costs as well as design of immunotherapeutic agents could be nearly halved through the involvement of computer-aided drug designing methods in discovery and research. The inter-disciplinary, time-, and money-consuming process of drug discovery is amended by the development of drug designing, the technique of creating or finding a molecule that can render stimulatory or inhibitory activity upon various biological organisms. Meanwhile, the advancements made within this scientific domain in the last couple of decades have significantly modified and affected the way new bioactive molecules have been produced by the pharmaceutical industry. In this regard, improvements made in hardware solutions and computational techniques along with their efficient integration with biological processes have revolutionized the in silico methods in speeding up the lead identification and optimization processes. In this review, we will discuss various methods of recent computer-aided drug designing techniques that forms the basis of modern day drug discovery projects.

Macrovoid-Inhibited PVDF Hollow Fiber Membranes via Spinning Process Delay for Direct Contact Membrane Distillation.

A polyvinylidene fluoride (PVDF) hollow fiber membrane was fabricated through water-induced dope crystallization by allowing a facile spinning process delay (SPD) in the nonsolvent-induced phase separation (NIPS) process for direct contact membrane distillation (DCMD). The SPD was achieved by the addition of a small amount of water to the PVDF dope solution that was held in a closed container for a particular time. The crystalline property of the PVDF dope solution was investigated by differential scanning calorimetry. The obtained PVDF hollow fiber membranes were characterized with different techniques, including field emission scanning electron microscopy, X-ray diffraction, and the mechanical strength. Both the formation mechanism and properties were studied for the membranes with different SPD times. The results showed that macrovoid-inhibited PVDF membranes were obtained from 12 days of the SPD via the crystallization-dominated membrane formation process. The obtained membrane 4D-12 exhibited desirable membrane structure and properties for DCMD, which includes an improved liquid entry pressure of 2.25 bar, a surface water contact angle of 129°, a maximum pore size of 0.40 µm, and a mean pore size of 0.34 µm. The membrane 4D-12 possessed a twofold increase in both energy efficiency and permeate water flux in DCMD and stable permeate water flux and salt rejection through 224 h of continuous desalination operation. Compared to the commonly used approach by adding chemicals to the external coagulant, the SPD method provided a low-cost and environmentally friendly alternative to pursuing the macrovoid-free PVDF membranes for DCMD.

De Novo High-Titer Production of Delta-Tocotrienol in Recombinant Saccharomyces cerevisiae.

Delta-tocotrienol as a vitamin E isomer has received much attention because of its diverse biomedical applications. Microbial biosynthesis of delta-tocotrienol is a promising strategy for its economic and environmental advantages. Here, we accomplished complete biosynthesis of delta-tocotrienol in Saccharomyces cerevisiae from glucose. We first constructed and incorporated a heterologous pathway into the genome of S. cerevisiae by incorporating the genes hpd (from Pseudomonas putida KT2440), hpt (from Synechocystis sp. PCC 6803), and vte1 (from Arabidopsis thaliana) for the biosynthesis of delta-tocotrienol. We further enhanced the biosynthesis of the precursor geranylgeranyl diphosphate by overexpressing the thmg1 and ggppssa (from Sulfolobus acidocaldarius) genes, leading to a production titer of delta-tocotrienol of 1.39 ± 0.01 mg/L. Finally, we optimized the fermentation medium using the response surface methodology, enabling a high-titer production of delta-tocotrienol (3.56 ± 0.25 mg/L), ∼2.6-fold of that of the initial culture medium. Fed-batch fermentation in a 2 L fermenter was further used to enhance the production titer of delta-tocotrienol (4.10 ± 0.10 mg/L). To the best of our knowledge, this is the first report on the de novo biosynthesis of delta-tocotrienol in S. cerevisiae, and the highest titer obtained for microbial production of delta-tocotrienol.

Deep high-temperature hydrothermal circulation in a detachment faulting system on the ultra-slow spreading ridge.

Coupled magmatic and tectonic activity plays an important role in high-temperature hydrothermal circulation at mid-ocean ridges. The circulation patterns for such systems have been elucidated by microearthquakes and geochemical data over a broad spectrum of spreading rates, but such data have not been generally available for ultra-slow spreading ridges. Here we report new geophysical and fluid geochemical data for high-temperature active hydrothermal venting at Dragon Horn area (49.7°E) on the Southwest Indian Ridge. Twin detachment faults penetrating to the depth of 13 ± 2 km below the seafloor were identified based on the microearthquakes. The geochemical composition of the hydrothermal fluids suggests a long reaction path involving both mafic and ultramafic lithologies. Combined with numerical simulations, our results demonstrate that these hydrothermal fluids could circulate ~ 6 km deeper than the Moho boundary and to much greater depths than those at Trans-Atlantic Geotraverse and Logachev-1 hydrothermal fields on the Mid-Atlantic Ridge.

A fine granularity based user collaboration algorithm for location privacy protection.

As the location trajectory contains more spatial-temporal information about the user, it will be even dangerous for jeopardizing the privacy of the user. In order to cope with the correlation, an algorithm that utilizes the query division had been proposed. In this algorithm, random blocks of query context was used, so as the adversary was obfuscated and difficult to correlate the real result. However, this algorithm fails to dispose the size of each query block, as once same size blocks were obtained by the adversary continuously, so the adversary can regard them as blocks from the same query context, and then obtains the query context to correlate the discrete locations. In view of above conditions, in this paper we propose a fine granularity block division algorithm based on the conception of granularity measurement as well as granularity layer division, so with the help of collaborative users the location privacy of the user will be protected. In this algorithm, the query context will be divided into fine granularity size of information blocks that difficult to be distinguished with others, and then these blocks will be exchanged with other collaborative users to eliminate the difference in block size. In addition, as each block is divided into fine granularity size, the adversary will be difficult to correlate the discrete locations into location trajectory, so the location privacy will be protected. At last, through security analysis and experimental verification, this granularity indistinguishable algorithm is analyzed and verified at both theoretical and practical levels, which further demonstrate the superiority of the proposed algorithm compared with other similar algorithms.

Modular Pathway Engineering of Bacillus subtilis To Promote De Novo Biosynthesis of Menaquinone-7.

Menaquinone-7 (MK-7), a valuable vitamin K2, plays an important role in the prevention of osteoporosis and cardiovascular calcification. We chose B. subtilis 168 as the chassis for the modular metabolic engineering design to promote the biosynthesis of MK-7. The biosynthetic pathway of MK-7 was categorized into four modules, namely, the MK-7 pathway (Module I), the shikimate (SA) pathway (Module II), the methylerythritol phosphate (MEP) pathway (Module III), and the glycerol metabolism pathway (Module IV). Overexpression of menA (Module I) resulted in 6.6 ± 0.1 mg/L of MK-7 after 120 h fermentation, which was 2.1-fold that of the starting strain BS168NU (3.1 ± 0.2 mg/L). Overexpression of aroA, aroD, and aroE (Module II) had a negative effect on the synthesis of MK-7. Simultaneous overexpression of dxs, dxr, yacM, and yacN (Module III) enabled the yield of MK-7 to 12.0 ± 0.1 mg/L. Moreover, overexpression of glpD (Module IV) resulted in an increase of the yield of MK-7 to 13.7 ± 0.2 mg/L. Furthermore, deletion of dhbB reduced the consumption of the intermediate metabolite isochorismate, thus promoting the yield of MK-7 to 15.4 ± 0.6 mg/L. Taken together, the final resulting strain MK3-MEP123-Gly2-Δ dhbB with simultaneous overexpression of menA, dxs, dxr, yacM-yacN, glpD and deletion of dhbB enabled the yield of MK-7 to 69.5 ± 2.8 mg/L upon 144 h fermentation in a 2 L baffled flask.

Structural inference embedded adversarial networks for scene parsing.

Explicit structural inference is one key point to improve the accuracy of scene parsing. Meanwhile, adversarial training method is able to reinforce spatial contiguity in output segmentations. To take both advantages of the structural learning and adversarial training simultaneously, we propose a novel deep learning network architecture called Structural Inference Embedded Adversarial Networks (SIEANs) for pixel-wise scene labeling. The generator of our SIEANs, a novel designed scene parsing network, makes full use of convolutional neural networks and long short-term memory networks to learn the global contextual information of objects in four different directions from RGB-(D) images, which is able to describe the (three-dimensional) spatial distributions of objects in a more comprehensive and accurate way. To further improve the performance, we explore the adversarial training method to optimize the generator along with a discriminator, which can not only detect and correct higher-order inconsistencies between the predicted segmentations and corresponding ground truths, but also exploit full advantages of the generator by fine-tuning its parameters so as to obtain higher consistencies. The experimental results demonstrate that our proposed SIEANs is able to achieve a better performance on PASCAL VOC 2012, SIFT FLOW, PASCAL Person-Part, Cityscapes, Stanford Background, NYUDv2, and SUN-RGBD datasets compared to the most of state-of-the-art methods.

Distribution of tetraether lipids in sulfide chimneys at the Deyin hydrothermal field, southern Mid-Atlantic Ridge: Implication to chimney growing stage.

This study presents analysis of four chimney samples in terms of glycerol dialkyl glycerol tetraether lipids (GDGTs), representing different growing stages of sulfide chimneys at the Deyin hydrothermal field, the southern mid-Atlantic ridge. The modified Bligh-Dyer method was used for lipid extraction and purification. GDGTs were analyzed with an Agilent 1200 series liquid chromatograph and 6460A triple quadrupole mass spectrometer. Our results showed that the intact polar GDGTs were more abundant than the core GDGTs in the 4 samples. The intact polar isoprenoidal GDGT-0 was the dominant composition (>70% of isoprenoidal GDGTs), indicating input of thermophilic Euryarchaeota. Most branched GDGTs were likely originated from the in situ thermophilic bacteria. However, the intact polar GDGTs in the sample at the late growing stage was similar to that in normal marine sediments, suggesting that the archaea mainly came from the planktonic Thaumarchaeota input. Our results suggested that the ratio of H-GDGTs to iGDGTs could be considered as a proxy to differentiated growing stages of a chimney. This study shed light on how to assess hydrothermal venting and sulfide chimneys in deep marine environments with a biomarker method in terms of different groups of GDGTs.

Minimal Increase Network Coding for Dynamic Networks.

Because of the mobility, computing power and changeable topology of dynamic networks, it is difficult for random linear network coding (RLNC) in static networks to satisfy the requirements of dynamic networks. To alleviate this problem, a minimal increase network coding (MINC) algorithm is proposed. By identifying the nonzero elements of an encoding vector, it selects blocks to be encoded on the basis of relationship between the nonzero elements that the controls changes in the degrees of the blocks; then, the encoding time is shortened in a dynamic network. The results of simulations show that, compared with existing encoding algorithms, the MINC algorithm provides reduced computational complexity of encoding and an increased probability of delivery.

Promising Anti-influenza Properties of Active Constituent of Withania somnifera Ayurvedic Herb in Targeting Neuraminidase of H1N1 Influenza: Computational Study.

Neuraminidase (NA) is a membrane surface antigen which helps in the release of influenza viruses from the host cells after replication. Anti-influenza drugs such as zanamivir bind with eight highly conserved functional residues (R118, D151, R152, R224, E276, R292, R371, and Y406) in the active site of NA, thus restricting the viral release the from host cells. Binding of the drug in active site inhibits the ability of enzyme to cleave sialic acid residues on the cell membrane. Reports on the emergence of zanamivir-resistant strains of H1N1 Influenza virus necessitated a search for alternative drug candidates, preferably from plant source due to their known benefits such as less or no side effects, availability, and low cost. Withaferin A (WA), an active constituent of Withania somnifera ayurvedic herb, has been shown to have a broad range of medicinal properties including its anti-viral activity. The present study demonstrated that WA has the potential to attenuate the neuraminidase of H1N1 influenza. Our docking and simulation results predicted high binding affinity of the WA toward NA and revealed several interesting molecular interactions with the residues which are catalytically important during molecular dynamic simulations. The results presented in the article could be of high importance for further designing of target-specific anti-influenza drug candidates.