Alpha&ESMhFolds: A Web Server for Comparing AlphaFold2 and ESMFold Models of the Human Reference Proteome

We develop a novel database Alpha&ESMhFolds which allows the direct comparison of AlphaFold2 and ESMFold predicted models for 42,942 proteins of the Reference Human Proteome, and when available, their comparison with 2,900 directly associated PDB structures with at least a structure to sequence coverage of 70%. Statistics indicate that good quality models tend to overlap with a TM-score >0.6 as long as some PDB structural information is available. As expected, a direct model superimposition to the PDB structure highlights that AlphaFold2 models are slightly superior to ESMFold ones. However, some 55% of the database is endowed with models overlapping with TM-score <0.6. This highlights the different outputs of the two methods. The database is freely available for usage at https://alpha-esmhfolds.biocomp.unibo.it/.

Deep learning methods in metagenomics: a review

The ever-decreasing cost of sequencing and the growing potential applications of metagenomics have led to an unprecedented surge in data generation. One of the most prevalent applications of metagenomics is the study of microbial environments, such as the human gut. The gut microbiome plays a crucial role in human health, providing vital information for patient diagnosis and prognosis. However, analysing metagenomic data remains challenging due to several factors, including reference catalogues, sparsity and compositionality. Deep learning (DL) enables novel and promising approaches that complement state-of-the-art microbiome pipelines. DL-based methods can address almost all aspects of microbiome analysis, including novel pathogen detection, sequence classification, patient stratification and disease prediction. Beyond generating predictive models, a key aspect of these methods is also their interpretability. This article reviews DL approaches in metagenomics, including convolutional networks, autoencoders and attention-based models. These methods aggregate contextualized data and pave the way for improved patient care and a better understanding of the microbiome's key role in our health.

Multi-modal features-based human-herpesvirus protein-protein interaction prediction by using LightGBM

The identification of human-herpesvirus protein-protein interactions (PPIs) is an essential and important entry point to understand the mechanisms of viral infection, especially in malignant tumor patients with common herpesvirus infection. While natural language processing (NLP)-based embedding techniques have emerged as powerful approaches, the application of multi-modal embedding feature fusion to predict human-herpesvirus PPIs is still limited. Here, we established a multi-modal embedding feature fusion-based LightGBM method to predict human-herpesvirus PPIs. In particular, we applied document and graph embedding approaches to represent sequence, network and function modal features of human and herpesviral proteins. Training our LightGBM models through our compiled non-rigorous and rigorous benchmarking datasets, we obtained significantly better performance compared to individual-modal features. Furthermore, our model outperformed traditional feature encodings-based machine learning methods and state-of-the-art deep learning-based methods using various benchmarking datasets. In a transfer learning step, we show that our model that was trained on human-herpesvirus PPI dataset without cytomegalovirus data can reliably predict human-cytomegalovirus PPIs, indicating that our method can comprehensively capture multi-modal fusion features of protein interactions across various herpesvirus subtypes. The implementation of our method is available at https://github.com/XiaodiYangpku/MultimodalPPI/.

Single-residue linear and conformational B cell epitopes prediction using random and ESM-2 based projections

B cell epitope prediction methods are separated into linear sequence-based predictors and conformational epitope predictions that typically use the measured or predicted protein structure. Most linear predictions rely on the translation of the sequence to biologically based representations and the applications of machine learning on these representations. We here present CALIBER 'Conformational And LInear B cell Epitopes pRediction', and show that a bidirectional long short-term memory with random projection produces a more accurate prediction (test set AUC=0.789) than all current linear methods. The same predictor when combined with an Evolutionary Scale Modeling-2 projection also improves on the state of the art in conformational epitopes (AUC = 0.776). The inclusion of the graph of the 3D distances between residues did not increase the prediction accuracy. However, the long-range sequence information was essential for high accuracy. While the same model structure was applicable for linear and conformational epitopes, separate training was required for each. Combining the two slightly increased the linear accuracy (AUC 0.775 versus 0.768) and reduced the conformational accuracy (AUC = 0.769).

Modeling Similarity and Psychological Space

Similarity and categorization are fundamental processes in human cognition that help complex organisms make sense of the cacophony of information in their environment. These processes are critical for tasks such as recognizing objects, making decisions, and forming memories. In this review, we provide an overview of the current state of knowledge on similarity and psychological spaces, discussing the theories, methods, and empirical findings that have been generated over the years. Although the concept of similarity has important limitations, it plays a key role in cognitive modeling. The review surfaces three key themes. First, similarity and mental representations are merely two sides of the same coin, existing as a similarity-representation duality that defines a psychological space. Second, both the brain's mental representations and the study of mental representations are made possible by exploiting second-order isomorphism. Third, similarity analysis has near-universal applicability across all levels of cognition, providing a common research language.

Use of a Backflush Needle with a Silicone Tip Cannula to Embed Lamellar Hole-Associated Epiretinal Proliferation

PURPOSE

For the treatment of lamellar macular hole, the recent development of a lamellar hole-associated epiretinal proliferation (LHEP) embedding technique is likely to improve functional and anatomical results. However, the peeling of LHEP is often technically challenging. We have developed a new technique using a backflush needle with a silicone tip cannula that seems safer and more effective for use in LHEP embedding.

METHODS

A 25-gauge vitrectomy system with an enhancing visual acuity system (D.O.R.C., Zuidland, Netherlands) was used in all cases. After core vitrectomy, triamcinolone acetonide (Wakamoto Pharmaceutical Co., Ltd., Tokyo, Japan) was used to visualize the membrane. A 25-gauge backflush needle with a silicone tip cannula was used to remove the thin preretinal membrane centripetally, leaving an LHEP on the edge of the hole. Brilliant Blue G (internal limiting membrane Blue; D.O.R.C.) was then used to stain the internal limiting membrane.

RESULTS

This technique was used in six eyes with lamellar macular holes. In all cases, peeling and embedding of the LHEP was effectively performed without damaging the internal limiting membrane or causing retinal hemorrhage. No other intraoperative or postoperative complications were experienced.

CONCLUSION

Using a silicone-tipped backflush needle with passive aspiration was a simple and effective technique for peeling and embedding of LHEPs in this small series.

Histology of the whole body of honey bees: tissue fixation and processing

Beekeeping plays a crucial role in biodiversity, pollination, commercial farming, and the worldwide agricultural economy. Histopathology, which is an important tool for the investigation of diseases in vertebrates, is not commonly used in honey bees (Apis mellifera). However, histopathology could potentially help the diagnostic investigation of high mortality in bees. We developed a tissue fixation and processing method enabling systematic production of histologic slides adequate for diagnostic and research purposes. Our method uses inexpensive, accessible products and can be realized with conventional pathology laboratory equipment. The quality of histologic slides obtained is similar to those of vertebrate animals processed routinely in pathology laboratories. Histopathology as a diagnostic and research tool will improve the services currently offered to apiarists and could help decrease the mean mortality rate, increase apiarists' profits, and ensure long-term pollination services.

Transglutaminase, glucono-δ-lactone, and citric acid-induced whey protein isolation-milk fat emulsion gel embedding lutein and its application in processed cheese

In this study, transglutaminase (TG), glucono-δ-lactone (GDL), and citric acid (CA) were used to induce the formation of whey protein isolate (WPI)-milk fat emulsion gels to embed lutein, and the emulsion gels induced in different ways were used for the preparation of processed cheese. The protective effect of emulsion gels induced in different ways on lutein was investigated, and the stability of lutein in emulsion gels and processed cheese was analyzed. The results showed that the acidification rate of CA was higher than that of GDL, which was the key step in acid-induced gels, and that the difference in acidification rate led to differences in gel structure. Compared with the 2 acid inducers (GDL and CA), TG exhibited greater potential for forming gel structures with high strength. The TG-induced emulsion gels showed the best physical stability and the highest embedding efficiency for lutein. After heat treatment (85°C), the GDL-induced emulsion gels had higher retention rate of lutein and showed good thermal stability compared with the CA-induced emulsion gels. The processed cheese added with the TG-induced emulsion gel had higher hardness and springiness compared with the processed cheese added with the other 2 kinds of emulsion gels, whereas the processed cheese added with the CA-induced emulsion gel had a lower density of network structure, showing porosity and a larger aggregated structure, but the highest bioavailability of lutein. These results provide valuable information for the formation of cold-set emulsion gel and provide the possibility for the application of emulsion gel embedding active substances in processed cheese.

A Review on Acoustic Emission Testing for Structural Health Monitoring of Polymer-Based Composites

Acoustic emission (AE) has received increased interest as a structural health monitoring (SHM) technique for various materials, including laminated polymer composites. Piezoelectric sensors, including PZT (piezoelectric ceramic) and PVDF (piezoelectric polymer), can monitor AE in materials. The thickness of the piezoelectric sensors (as low as 28 µm-PVDF) allows embedding the sensors within the laminated composite, creating a smart material. Incorporating piezoelectric sensors within composites has several benefits but presents numerous difficulties and challenges. This paper provides an overview of acoustic emission testing, concluding with a discussion on embedding piezoelectric AE sensors within fibre-polymer composites. Various aspects are covered, including the underlying AE principles in fibre-based composites, factors that influence the reliability and accuracy of AE measurements, methods to artificially induce acoustic emission, and the correlation between AE events and damage in polymer composites.

LMPhosSite: A Deep Learning-Based Approach for General Protein Phosphorylation Site Prediction Using Embeddings from the Local Window Sequence and Pretrained Protein Language Model

Phosphorylation is one of the most important post-translational modifications and plays a pivotal role in various cellular processes. Although there exist several computational tools to predict phosphorylation sites, existing tools have not yet harnessed the knowledge distilled by pretrained protein language models. Herein, we present a novel deep learning-based approach called LMPhosSite for the general phosphorylation site prediction that integrates embeddings from the local window sequence and the contextualized embedding obtained using global (overall) protein sequence from a pretrained protein language model to improve the prediction performance. Thus, the LMPhosSite consists of two base-models: one for capturing effective local representation and the other for capturing global per-residue contextualized embedding from a pretrained protein language model. The output of these base-models is integrated using a score-level fusion approach. LMPhosSite achieves a precision, recall, Matthew's correlation coefficient, and F1-score of 38.78%, 67.12%, 0.390, and 49.15%, for the combined serine and threonine independent test data set and 34.90%, 62.03%, 0.298, and 44.67%, respectively, for the tyrosine independent test data set, which is better than the compared approaches. These results demonstrate that LMPhosSite is a robust computational tool for the prediction of the general phosphorylation sites in proteins.

Fast-Embeddable Grooved Microneedles by Shear Actuation for Accurate Transdermal Drug Delivery

Percutaneous drug delivery using microneedles (MNs) has been extensively exploited to increase the transdermal permeability of therapeutic drugs. However, it is difficult to control the precise dosage with existing MNs and they need to be attached for a long time, so a more simple and scalable method is required for accurate transdermal drug delivery. In this study, we developed grooved MNs that can be embedded into the skin by mechanical fracture following simple shear actuation. Grooved MNs are prepared from hyaluronic acid (HA), which is a highly biocompatible and biodegradable biopolymer. By adjusting the aspect ratio (length:diameter) of the MN and the position of the groove, the MN tip inserted into the skin can be easily broken by shear force. In addition, it was demonstrated that it is possible to deliver the desired amount of triamcinolone acetonide (TCA) for alopecia areata by controlling the position of the groove structure and the concentration of TCA loaded in the MN. It was also confirmed that the tip of the TCA MN can be accurately delivered into the skin with a high probability (98% or more) by fabricating an easy-to-operate applicator to provide adequate shear force. The grooved MN platform has proven to be able to load the desired amount of a drug and deliver it at the correct dose.

Metric Learning in Histopathological Image Classification: Opening the Black Box

The application of machine learning techniques to histopathology images enables advances in the field, providing valuable tools that can speed up and facilitate the diagnosis process. The classification of these images is a relevant aid for physicians who have to process a large number of images in long and repetitive tasks. This work proposes the adoption of metric learning that, beyond the task of classifying images, can provide additional information able to support the decision of the classification system. In particular, triplet networks have been employed to create a representation in the embedding space that gathers together images of the same class while tending to separate images with different labels. The obtained representation shows an evident separation of the classes with the possibility of evaluating the similarity and the dissimilarity among input images according to distance criteria. The model has been tested on the BreakHis dataset, a reference and largely used dataset that collects breast cancer images with eight pathology labels and four magnification levels. Our proposed classification model achieves relevant performance on the patient level, with the advantage of providing interpretable information for the obtained results, which represent a specific feature missed by the all the recent methodologies proposed for the same purpose.

Using Hybrid HMM/DNN Embedding Extractor Models in Computational Paralinguistic Tasks

The field of computational paralinguistics emerged from automatic speech processing, and it covers a wide range of tasks involving different phenomena present in human speech. It focuses on the non-verbal content of human speech, including tasks such as spoken emotion recognition, conflict intensity estimation and sleepiness detection from speech, showing straightforward application possibilities for remote monitoring with acoustic sensors. The two main technical issues present in computational paralinguistics are (1) handling varying-length utterances with traditional classifiers and (2) training models on relatively small corpora. In this study, we present a method that combines automatic speech recognition and paralinguistic approaches, which is able to handle both of these technical issues. That is, we trained a HMM/DNN hybrid acoustic model on a general ASR corpus, which was then used as a source of embeddings employed as features for several paralinguistic tasks. To convert the local embeddings into utterance-level features, we experimented with five different aggregation methods, namely mean, standard deviation, skewness, kurtosis and the ratio of non-zero activations. Our results show that the proposed feature extraction technique consistently outperforms the widely used x-vector method used as the baseline, independently of the actual paralinguistic task investigated. Furthermore, the aggregation techniques could be combined effectively as well, leading to further improvements depending on the task and the layer of the neural network serving as the source of the local embeddings. Overall, based on our experimental results, the proposed method can be considered as a competitive and resource-efficient approach for a wide range of computational paralinguistic tasks.

3D Neuron Morphology Analysis

We consider the problem of finding an accurate representation of neuron shapes, extracting sub-cellular features, and classifying neurons based on neuron shapes. In neuroscience research, the skeleton representation is often used as a compact and abstract representation of neuron shapes. However, existing methods are limited to getting and analyzing"curve"skeletons which can only be applied for tubular shapes. This paper presents a 3D neuron morphology analysis method for more general and complex neuron shapes. First, we introduce the concept of skeleton mesh to represent general neuron shapes and propose a novel method for computing mesh representations from 3D surface point clouds. A skeleton graph is then obtained from skeleton mesh and is used to extract sub-cellular features. Finally, an unsupervised learning method is used to embed the skeleton graph for neuron classification. Extensive experiment results are provided and demonstrate the robustness of our method to analyze neuron morphology.

Fabrication and Analysis of Microcapsule Electrets with a Tunable Flexoelectric-like Response

The electret has drawn considerable attention as an emerging flexible energy collector. In this work, a charged microcapsule is designed which can provide a stable storage space for electric charge in the electret. The flexoelectric-like response is achieved by embedding a layer of charged microcapsules in the middle plane of the flexible polymer to form an electret. The results of Fourier transform infrared spectroscopy and energy-dispersive X-ray spectroscopy verified the successful preparation of microcapsules. Zeta potential analysis showed the negative electrical properties of the microcapsules. The prepared microcapsule electret has a significant flexoelectric effect under loading conditions. This work presents a preliminary theoretical study of the microcapsule electret to optimize the output characteristics of the electret by varying the parameters, including the number of microcapsules, the size of the electret, and the external load. Good agreement was achieved with the experimental results, which verified the validity of the theoretical study. This work provides a new method for preparing electret and further promotes its application in electromechanical transducers.

SMURF: embedding single-cell RNA-seq data with matrix factorization preserving self-consistency

The advance in single-cell RNA-sequencing (scRNA-seq) sheds light on cell-specific transcriptomic studies of cell developments, complex diseases and cancers. Nevertheless, scRNA-seq techniques suffer from 'dropout' events, and imputation tools are proposed to address the sparsity. Here, rather than imputation, we propose a tool, SMURF, to extract the low-dimensional embeddings from cells and genes utilizing matrix factorization with a mixture of Poisson-Gamma divergent as objective while preserving self-consistency. SMURF exhibits feasible cell subpopulation discovery efficacy with obtained cell embeddings on replicated in silico and eight web lab scRNA datasets with ground truth cell types. Furthermore, SMURF can reduce the cell embedding to a 1D-oval space to recover the time course of cell cycle. SMURF can also serve as an imputation tool; the in silico data assessment shows that SMURF parades the most robust gene expression recovery power with low root mean square error and high Pearson correlation. Moreover, SMURF recovers the gene distribution for the WM989 Drop-seq data. SMURF is available at https://github.com/deepomicslab/SMURF.

NeuroPpred-SVM: A New Model for Predicting Neuropeptides Based on Embeddings of BERT

Neuropeptides play pivotal roles in different physiological processes and are related to different kinds of diseases. Identification of neuropeptides is of great benefit for studying the mechanism of these physiological processes and the treatment of neurological disorders. Several state-of-the-art neuropeptide predictors have been developed by using a two-layer stacking ensemble algorithm. Although the two-layer stacking ensemble algorithm can improve the feature representability, these models are complex, which are not as efficient as the models based on one classifier. In this study, we proposed a new model, NeuroPpred-SVM, to predict neuropeptides based on the embeddings of Bidirectional Encoder Representations from Transformers and other sequential features by using a support vector machine (SVM). The experimental results indicate that our model achieved a cross-validation area under the receiver operating characteristic (AUROC) curve of 0.969 on the training data set and an AUROC of 0.966 on the independent test set. By comparing our model with the other four state-of-the-art models including NeuroPIpred, PredNeuroP, NeuroPpred-Fuse, and NeuroPpred-FRL on the independent test set, our model achieved the highest AUROC, Matthews correlation coefficient, accuracy, and specificity, which indicate that our model outperforms the existing models. We believed that NeuroPpred-SVM could be a useful tool for identifying neuropeptides with high accuracy and low cost. The data sets and Python code are available at https://github.com/liuyf-a/NeuroPpred-SVM.

A Graph-Attention-Based Method for Single-Resident Daily Activity Recognition in Smart Homes

In ambient-assisted living facilitated by smart home systems, the recognition of daily human activities is of great importance. It aims to infer the household's daily activities from the triggered sensor observation sequences with varying time intervals among successive readouts. This paper introduces a novel deep learning framework based on embedding technology and graph attention networks, namely the time-oriented and location-oriented graph attention (TLGAT) networks. The embedding technology converts sensor observations into corresponding feature vectors. Afterward, TLGAT provides a sensor observation sequence as a fully connected graph to the model's temporal correlation as well as the sensor's location correlation among sensor observations and facilitates the feature representation of each sensor observation through receiving other sensor observations and weighting operations. The experiments were conducted on two public datasets, based on the diverse setups of sensor event sequence length. The experimental results revealed that the proposed method achieved favorable performance under diverse setups.

Quantum computing in pharma: A multilayer embedding approach for near future applications

Quantum computers are special purpose machines that are expected to be particularly useful in simulating strongly correlated chemical systems. The quantum computer excels at treating a moderate number of orbitals within an active space in a fully quantum mechanical manner. We present a quantum phase estimation calculation on F₂ in a (2,2) active space on Rigetti's Aspen-11 QPU. While this is a promising start, it also underlines the need for carefully selecting the orbital spaces treated by the quantum computer. In this work, a scheme for selecting such an active space automatically is described and simulated results obtained using both the quantum phase estimation (QPE) and variational quantum eigensolver (VQE) algorithms are presented and combined with a subtractive method to enable accurate description of the environment. The active occupied space is selected from orbitals localized on the chemically relevant fragment of the molecule, while the corresponding virtual space is chosen based on the magnitude of interactions with the occupied space calculated from perturbation theory. This protocol is then applied to two chemical systems of pharmaceutical relevance: the enzyme [Fe] hydrogenase and the photosenzitizer temoporfin. While the sizes of the active spaces currently amenable to a quantum computational treatment are not enough to demonstrate quantum advantage, the procedure outlined here is applicable to any active space size, including those that are outside the reach of classical computation.

Theoretical Study on Non-Linear Optics Properties of Polycyclic Aromatic Hydrocarbons and the Effect of Their Intercalation with Carbon Nanotubes

Results of a theoretical study devoted to comparing NLO (non-linear optics) responses of derivatives of tetracene, isochrysene, and pyrene are reported. The static hyperpolarizability β, the dipole moment μ, the HOMO and LUMO orbitals, and their energy gap were calculated using the CAM-B3LYP density functional combined with the cc-pVDZ basis set. The para-disubstituted NO₂-tetracene-N(CH₃)₂ has the highest NLO response, which is related to a large intramolecular charge transfer. Adding vinyl groups to the para-disubstituted NO₂-tetracene-N(CH₃)₂ results in an increase in the NLO responses. We further investigated the effect of the intercalation of various push-pull molecules inside an armchair single-walled carbon nanotube. The intercalation leads to increased NLO responses, something that depends critically on the position of the guest molecule and/or on functionalization of the nanotube by donor and attractor groups.

Early maternal sensitivity and markers of physical health: Enduring or transient associations from childhood to adulthood?

Individual differences in the quality of early experiences with primary caregivers have been reliably implicated in the development of socioemotional adjustment and, more recently, physical health. However, few studies have examined the development of such associations with physical health into the adult years. To that end, the current study used prospective, longitudinal data from the National Institute of Child Health and Human Development Study of Early Child Care and Youth Development (N = 1,306, 52% male, 77% White/non-Hispanic) to investigate whether associations between direct observations of maternal sensitivity in the first 3 years of life and repeated assessments of two commonly used, objective indicators of physical health (i.e., body mass and mean arterial blood pressure) remained stable or diminished in magnitude over time. Associations between early maternal sensitivity and lower body mass remained relatively stable from age 54 months to 26 years and were robust to the modeling of autoregressive and second-order stability processes as well as the inclusion of potential demographic confounders. In contrast, although associations between early caregiving and lower mean arterial pressure remained relatively stable from Grade 4 to age 15 years (the oldest age for which mean arterial pressure was assessed thus far), these associations were not robust to the inclusion of covariates and the modeling of second-order stability processes. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

binny: an automated binning algorithm to recover high-quality genomes from complex metagenomic datasets

The reconstruction of genomes is a critical step in genome-resolved metagenomics and for multi-omic data integration from microbial communities. Here, we present binny, a binning tool that produces high-quality metagenome-assembled genomes (MAG) from both contiguous and highly fragmented genomes. Based on established metrics, binny outperforms or is highly competitive with commonly used and state-of-the-art binning methods and finds unique genomes that could not be detected by other methods. binny uses k-mer-composition and coverage by metagenomic reads for iterative, nonlinear dimension reduction of genomic signatures as well as subsequent automated contig clustering with cluster assessment using lineage-specific marker gene sets. When compared with seven widely used binning algorithms, binny provides substantial amounts of uniquely identified MAGs and almost always recovers the most near-complete ($\gt 95\%$ pure, $\gt 90\%$ complete) and high-quality ($\gt 90\%$ pure, $\gt 70\%$ complete) genomes from simulated datasets from the Critical Assessment of Metagenome Interpretation initiative, as well as substantially more high-quality draft genomes, as defined by the Minimum Information about a Metagenome-Assembled Genome standard, from a real-world benchmark comprised of metagenomes from various environments than any other tested method.

Improved handling and embedding schemes for cultured murine neuroretinal explants

Traumatic, inherited, and age-related degenerative diseases of the retina, such as retinal detachment, retinitis pigmentosa, and age-related macular degeneration, are characterized by the irreversible loss of retinal neurons. While current treatments aim to prevent neuronal degeneration, there are no available treatments to restore neurons after loss. Cultured murine neuroretinal tissue explants model retinal injury and offer a high throughput approach to identify experimental interventions capable of regenerating neurons. Formalin-fixed paraffin-embedded (FFPE) preparations of murine neuroretinal explants can be used to identify cells throughout the retinal layers to provide information on proliferation and activity following exposure to therapeutics. However, retinal explants are friable, particularly after ex vivo culture, sample handling and FFPE processing steps can result in tissue loss and damage. Friability also prohibits bisecting samples post-culture to display more than one region of interest for analysis. We developed a sample handling and embedding technique for cultured murine neuroretinal explants using HistogelTM in combination with a post-processing trimming step that eliminates tissue loss, increases cross-sectional retinal representation, and captures proximal and central retina on one slide to facilitate analysis of explants subjected to neurotrophic compounds.

MedCV: An Interactive Visualization System for Patient Cohort Identification from Medical Claim Data

Healthcare providers generate a medical claim after every patient visit. A medical claim consists of a list of medical codes describing the diagnosis and any treatment provided during the visit. Medical claims have been popular in medical research as a data source for retrospective cohort studies. This paper introduces a medical claim visualization system (MedCV) that supports cohort selection from medical claim data. MedCV was developed as part of a design study in collaboration with clinical researchers and statisticians. It helps a researcher to define inclusion rules for cohort selection by revealing relationships between medical codes and visualizing medical claims and patient timelines. Evaluation of our system through a user study indicates that MedCV enables domain experts to define high-quality inclusion rules in a time-efficient manner.

Transfer learning in proteins: evaluating novel protein learned representations for bioinformatics tasks

A representation method is an algorithm that calculates numerical feature vectors for samples in a dataset. Such vectors, also known as embeddings, define a relatively low-dimensional space able to efficiently encode high-dimensional data. Very recently, many types of learned data representations based on machine learning have appeared and are being applied to several tasks in bioinformatics. In particular, protein representation learning methods integrate different types of protein information (sequence, domains, etc.), in supervised or unsupervised learning approaches, and provide embeddings of protein sequences that can be used for downstream tasks. One task that is of special interest is the automatic function prediction of the huge number of novel proteins that are being discovered nowadays and are still totally uncharacterized. However, despite its importance, up to date there is not a fair benchmark study of the predictive performance of existing proposals on the same large set of proteins and for very concrete and common bioinformatics tasks. Therefore, this lack of benchmark studies prevent the community from using adequate predictive methods for accelerating the functional characterization of proteins. In this study, we performed a detailed comparison of protein sequence representation learning methods, explaining each approach and comparing them with an experimental benchmark on several bioinformatics tasks: (i) determining protein sequence similarity in the embedding space; (ii) inferring protein domains and (iii) predicting ontology-based protein functions. We examine the advantages and disadvantages of each representation approach over the benchmark results. We hope the results and the discussion of this study can help the community to select the most adequate machine learning-based technique for protein representation according to the bioinformatics task at hand.

scHiCEmbed: Bin-Specific Embeddings of Single-Cell Hi-C Data Using Graph Auto-Encoders

Most publicly accessible single-cell Hi-C data are sparse and cannot reach a higher resolution. Therefore, learning latent representations (bin-specific embeddings) of sparse single-cell Hi-C matrices would provide us with a novel way of mining valuable information hidden in the limited number of single-cell Hi-C contacts. We present scHiCEmbed, an unsupervised computational method for learning bin-specific embeddings of single-cell Hi-C data, and the computational system is applied to the tasks of 3D structure reconstruction of whole genomes and detection of topologically associating domains (TAD). The only input of scHiCEmbed is a raw or scHiCluster-imputed single-cell Hi-C matrix. The main process of scHiCEmbed is to embed each node/bin in a higher dimensional space using graph auto-encoders. The learned n-by-3 bin-specific embedding/latent matrix is considered the final reconstructed 3D genome structure. For TAD detection, we use constrained hierarchical clustering on the latent matrix to classify bins: S_Dbw is used to determine the optimal number of clusters, and each cluster is considered as one potential TAD. Our reconstructed 3D structures for individual chromatins at different cell stages reveal the expanding process of chromatins during the cell cycle. We observe that the TADs called from single-cell Hi-C data are not shared across individual cells and that the TAD boundaries called from raw or imputed single-cell Hi-C are significantly different from those called from bulk Hi-C, confirming the cell-to-cell variability in terms of TAD definitions. The source code for scHiCEmbed is publicly available, and the URL can be found in the conclusion section.

Comparison of approximate intermolecular potentials for ab initio fragment calculations on medium sized N-heterocycles

The ground state intermolecular potential of bimolecular complexes of N‐heterocycles is analyzed for the impact of individual terms in the interaction energy as provided by various, conceptually different theories. Novel combinations with several formulations of the electrostatic, Pauli repulsion, and dispersion contributions are tested at both short‐ and long‐distance sides of the potential energy surface, for various alignments of the pyrrole dimer as well as the cytosine–uracil complex. The integration of a DFT/CCSD density embedding scheme, with dispersion terms from the effective fragment potential (EFP) method is found to provide good agreement with a reference CCSD(T) potential overall; simultaneously, a quantum mechanics/molecular mechanics approach using CHELPG atomic point charges for the electrostatic interaction, augmented by EFP dispersion and Pauli repulsion, comes also close to the reference result. Both schemes have the advantage of not relying on predefined force fields; rather, the interaction parameters can be determined for the system under study, thus being excellent candidates for ab initio modeling.

Embedding decomposition for artifacts removal in EEG signals

Electroencephalogram (EEG) recordings are often contaminated with artifacts. Various methods have been developed to eliminate or weaken the influence of artifacts. However, most of them rely on prior experience for analysis. Here, we propose an deep learning framework to separate neural signal and artifacts in the embedding space and reconstruct the denoised signal, which is called DeepSeparator. DeepSeparator employs an encoder to extract and amplify the features in the raw EEG, a module called decomposer to extract the trend, detect and suppress artifact and a decoder to reconstruct the denoised signal. Besides, DeepSeparator can extract the artifact, which largely increases the model interpretability. The proposed method is tested with a semi-synthetic EEG dataset and a real task-related EEG dataset, suggesting that DeepSeparator outperforms the conventional models in both EOG and EMG artifact removal. DeepSeparator can be extended to multi-channel EEG and data with any arbitrary length. It may motivate future developments and application of deep learning-based EEG denoising. The code for DeepSeparator is available at https://github.com/ncclabsustech/DeepSeparator.

A Crypto-Steganography Approach for Hiding Ransomware within HEVC Streams in Android IoT Devices

Steganography is a vital security approach that hides any secret content within ordinary data, such as multimedia. This hiding aims to achieve the confidentiality of the IoT secret data; whether it is benign or malicious (e.g., ransomware) and for defensive or offensive purposes. This paper introduces a hybrid crypto-steganography approach for ransomware hiding within high-resolution video frames. This proposed approach is based on hybridizing an AES (advanced encryption standard) algorithm and LSB (least significant bit) steganography process. Initially, AES encrypts the secret Android ransomware data, and then LSB embeds it based on random selection criteria for the cover video pixels. This research examined broad objective and subjective quality assessment metrics to evaluate the performance of the proposed hybrid approach. We used different sizes of ransomware samples and different resolutions of HEVC (high-efficiency video coding) frames to conduct simulation experiments and comparison studies. The assessment results prove the superior efficiency of the introduced hybrid crypto-steganography approach compared to other existing steganography approaches in terms of (a) achieving the integrity of the secret ransomware data, (b) ensuring higher imperceptibility of stego video frames, (3) introducing a multi-level security approach using the AES encryption in addition to the LSB steganography, (4) performing randomness embedding based on RPS (random pixel selection) for concealing secret ransomware bits, (5) succeeding in fully extracting the ransomware data at the receiver side, (6) obtaining strong subjective and objective qualities for all tested evaluation metrics, (7) embedding different sizes of secret data at the same time within the video frame, and finally (8) passing the security scanning tests of 70 antivirus engines without detecting the existence of the embedded ransomware.

SAAED: Embedding and Deep Learning Enhance Accurate Prediction of Association Between circRNA and Disease

Emerging evidence indicates that circRNA can regulate various diseases. However, the mechanisms of circRNA in these diseases have not been fully understood. Therefore, detecting potential circRNA–disease associations has far-reaching significance for pathological development and treatment of these diseases. In recent years, deep learning models are used in association analysis of circRNA–disease, but a lack of circRNA–disease association data limits further improvement. Therefore, there is an urgent need to mine more semantic information from data. In this paper, we propose a novel method called Semantic Association Analysis by Embedding and Deep learning (SAAED), which consists of two parts, a neural network embedding model called Entity Relation Network (ERN) and a Pseudo-Siamese network (PSN) for analysis. ERN can fuse multiple sources of data and express the information with low-dimensional embedding vectors. PSN can extract the feature between circRNA and disease for the association analysis. CircRNA–disease, circRNA–miRNA, disease–gene, disease–miRNA, disease–lncRNA, and disease–drug association information are used in this paper. More association data can be introduced for analysis without restriction. Based on the CircR2Disease benchmark dataset for evaluation, a fivefold cross-validation experiment showed an AUC of 98.92%, an accuracy of 95.39%, and a sensitivity of 93.06%. Compared with other state-of-the-art models, SAAED achieves the best overall performance. SAAED can expand the expression of the biological related information and is an efficient method for predicting potential circRNA–disease association.

EMBER-Embedding Multiple Molecular Fingerprints for Virtual Screening

In recent years, the debate in the field of applications of Deep Learning to Virtual Screening has focused on the use of neural embeddings with respect to classical descriptors in order to encode both structural and physical properties of ligands and/or targets. The attention on embeddings with the increasing use of Graph Neural Networks aimed at overcoming molecular fingerprints that are short range embeddings for atomic neighborhoods. Here, we present EMBER, a novel molecular embedding made by seven molecular fingerprints arranged as different “spectra” to describe the same molecule, and we prove its effectiveness by using deep convolutional architecture that assesses ligands’ bioactivity on a data set containing twenty protein kinases with similar binding sites to CDK1. The data set itself is presented, and the architecture is explained in detail along with its training procedure. We report experimental results and an explainability analysis to assess the contribution of each fingerprint to different targets.

Solid-Liquid Host-Guest Composites: The Marriage of Porous Solids and Functional Liquids

Composite materials can provide remarkable improvements over the individual constituents. Especially, with a liquid component introduced into a solid porous host, solid-liquid host-guest composites have recently come to the forefront with exceptional functions that promise them for a wealth of applications. Combining the unprecedented dynamic, transparent, omniphobic, self-healing, diffusive and adaptive nature of functional liquid with inherent solid host's property, solid-liquid host-guest composites can realize the ease of fabrication, long-term stability, and a broad spectrum of enhanced properties, which cannot be fully met by conventional solid-solid composites or liquid-liquid composites. This review presents the state-of-the-art progress in solid-liquid host-guest composites. Initially, the concept, classification, design strategy, as well as fabrication methods as a path forward to develop the composites are unraveled, and further it is elaborated on how the functionality of porous solid and functional liquid can be harnessed to create composites with a broad range of unique properties, especially, the optical, thermal, electric, mechanical, sorption, and separation properties. With these fascinating properties, a myriad of emerging applications such as optical devices, thermal management, electromagnetic-interference shielding, soft electronics, gas capture and release, and multiphase separations are touched upon, inspiring more frontier researches in materials science, interfacial chemistry, membrane science, engineering, and multidisciplinary. Finally, this review provides the perspective on the future directions of solid-liquid host-guest composites and assesses the challenges and opportunities ahead.

Developing an Embedding, Koopman and Autoencoder Technologies-Based Multi-Omics Time Series Predictive Model (EKATP) for Systems Biology research

It is very important for systems biologists to predict the state of the multi-omics time series for disease occurrence and health detection. However, it is difficult to make the prediction due to the high-dimensional, nonlinear and noisy characteristics of the multi-omics time series data. For this reason, this study innovatively proposes an Embedding, Koopman and Autoencoder technologies-based multi-omics time series predictive model (EKATP) to predict the future state of a high-dimensional nonlinear multi-omics time series. We evaluate this EKATP by using a genomics time series with chaotic behavior, a proteomics time series with oscillating behavior and a metabolomics time series with flow behavior. The computational experiments demonstrate that our proposed EKATP can substantially improve the accuracy, robustness and generalizability to predict the future state of a time series for multi-omics data.

D-SCRIPT translates genome to phenome with sequence-based, structure-aware, genome-scale predictions of protein-protein interactions

We combine advances in neural language modeling and structurally motivated design to develop D-SCRIPT, an interpretable and generalizable deep-learning model, which predicts interaction between two proteins using only their sequence and maintains high accuracy with limited training data and across species. We show that a D-SCRIPT model trained on 38,345 human PPIs enables significantly improved functional characterization of fly proteins compared with the state-of-the-art approach. Evaluating the same D-SCRIPT model on protein complexes with known 3D structure, we find that the inter-protein contact map output by D-SCRIPT has significant overlap with the ground truth. We apply D-SCRIPT to screen for PPIs in cow (Bos taurus) at a genome-wide scale and focusing on rumen physiology, identify functional gene modules related to metabolism and immune response. The predicted interactions can then be leveraged for function prediction at scale, addressing the genome-to-phenome challenge, especially in species where little data are available.

Differentiation Protocol for 3D Retinal Organoids, Immunostaining and Signal Quantitation

Structures resembling whole organs, called organoids, are generated using pluripotent stem cells and 3D culturing methods. This relies on the ability of cells to self-reorganize after dissociation. In combination with certain supplemented factors, differentiation can be directed toward the formation of several organ-like structures. Here, a protocol for the generation of retinal organoids containing all seven retinal cell types is described. This protocol does not depend on Matrigel, and by keeping the organoids single and independent at all times, fusion is prevented and monitoring of differentiation is improved. Comprehensive phenotypic characterization of the in vitro-generated retinal organoids is achieved by the protocol for immunostaining outlined here. By comparing different stages of retinal organoids, the decrease and increase of certain cell populations can be determined. In order to be able to detect even small differences, it is necessary to quantify the immunofluorescent signals, for which we have provided a detailed protocol describing signal quantitation using the image-processing program Fiji. © 2020 The Authors. Basic Protocol 1: Differentiation protocol for 3D retinal organoids Basic Protocol 2: Immunostaining protocol for cryosections of retinal organoids Support Protocol: Embedding and sectioning protocol for 3D retinal organoids Basic Protocol 3: Quantitation protocol using Fiji.

Comparative effectiveness of medical concept embedding for feature engineering in phenotyping

Objective

Feature engineering is a major bottleneck in phenotyping. Properly learned medical concept embeddings (MCEs) capture the semantics of medical concepts, thus are useful for retrieving relevant medical features in phenotyping tasks. We compared the effectiveness of MCEs learned from knowledge graphs and electronic healthcare records (EHR) data in retrieving relevant medical features for phenotyping tasks.

Materials and Methods

We implemented 5 embedding methods including node2vec, singular value decomposition (SVD), LINE, skip-gram, and GloVe with 2 data sources: (1) knowledge graphs obtained from the observational medical outcomes partnership (OMOP) common data model; and (2) patient-level data obtained from the OMOP compatible electronic health records (EHR) from Columbia University Irving Medical Center (CUIMC). We used phenotypes with their relevant concepts developed and validated by the electronic medical records and genomics (eMERGE) network to evaluate the performance of learned MCEs in retrieving phenotype-relevant concepts. Hits@k% in retrieving phenotype-relevant concepts based on a single and multiple seed concept(s) was used to evaluate MCEs.

Results

Among all MCEs, MCEs learned by using node2vec with knowledge graphs showed the best performance. Of MCEs based on knowledge graphs and EHR data, MCEs learned by using node2vec with knowledge graphs and MCEs learned by using GloVe with EHR data outperforms other MCEs, respectively.

Conclusion

MCE enables scalable feature engineering tasks, thereby facilitating phenotyping. Based on current phenotyping practices, MCEs learned by using knowledge graphs constructed by hierarchical relationships among medical concepts outperformed MCEs learned by using EHR data.

Application of dyes to cytology cell blocks and biopsy tissues before processing enhances specimen visualization during embedding and microtomy

Cytology specimens and biopsy tissues are frequently small and pale, making them difficult to visualize grossly in paraffin. Ten dyes were assayed on small tissues to determine if specimen discernibility could be increased during the embedding and microtomy steps in the histological process. The ideal dye should not remain visible in a tissue section microscopically after subsequent staining and must not interfere with immunohistochemistry (IHC) assays. This study found that Harris hematoxylin and 1% aq. toluidine blue solution were the best labelers for gross tissue visualization and did not adversely affect post-processing staining and IHC assays.

Sentiment Analysis Techniques Applied to Raw-Text Data from a Csq-8 Questionnaire about Mindfulness in Times of COVID-19 to Improve Strategy Generation

The use of artificial intelligence in health care has grown quickly. In this sense, we present our work related to the application of Natural Language Processing techniques, as a tool to analyze the sentiment perception of users who answered two questions from the CSQ-8 questionnaires with raw Spanish free-text. Their responses are related to mindfulness, which is a novel technique used to control stress and anxiety caused by different factors in daily life. As such, we proposed an online course where this method was applied in order to improve the quality of life of health care professionals in COVID 19 pandemic times. We also carried out an evaluation of the satisfaction level of the participants involved, with a view to establishing strategies to improve future experiences. To automatically perform this task, we used Natural Language Processing (NLP) models such as swivel embedding, neural networks, and transfer learning, so as to classify the inputs into the following three categories: negative, neutral, and positive. Due to the limited amount of data available-86 registers for the first and 68 for the second-transfer learning techniques were required. The length of the text had no limit from the user's standpoint, and our approach attained a maximum accuracy of 93.02% and 90.53%, respectively, based on ground truth labeled by three experts. Finally, we proposed a complementary analysis, using computer graphic text representation based on word frequency, to help researchers identify relevant information about the opinions with an objective approach to sentiment. The main conclusion drawn from this work is that the application of NLP techniques in small amounts of data using transfer learning is able to obtain enough accuracy in sentiment analysis and text classification stages.

Deep-learning-based automated terminology mapping in OMOP-CDM

OBJECTIVE

Accessing medical data from multiple institutions is difficult owing to the interinstitutional diversity of vocabularies. Standardization schemes, such as the common data model, have been proposed as solutions to this problem, but such schemes require expensive human supervision. This study aims to construct a trainable system that can automate the process of semantic interinstitutional code mapping.

MATERIALS AND METHODS

To automate mapping between source and target codes, we compute the embedding-based semantic similarity between corresponding descriptive sentences. We also implement a systematic approach for preparing training data for similarity computation. Experimental results are compared to traditional word-based mappings.

RESULTS

The proposed model is compared against the state-of-the-art automated matching system, which is called Usagi, of the Observational Medical Outcomes Partnership common data model. By incorporating multiple negative training samples per positive sample, our semantic matching method significantly outperforms Usagi. Its matching accuracy is at least 10% greater than that of Usagi, and this trend is consistent across various top-k measurements.

DISCUSSION

The proposed deep learning-based mapping approach outperforms previous simple word-level matching algorithms because it can account for contextual and semantic information. Additionally, we demonstrate that the manner in which negative training samples are selected significantly affects the overall performance of the system.

CONCLUSION

Incorporating the semantics of code descriptions more significantly increases matching accuracy compared to traditional text co-occurrence-based approaches. The negative training sample collection methodology is also an important component of the proposed trainable system that can be adopted in both present and future related systems.

When Gesture "Takes Over": Speech-Embedded Nonverbal Depictions in Multimodal Interaction

The framework of depicting put forward by Clark (2016) offers a schematic vantage point from which to examine iconic language use. Confronting the framework with empirical data, we consider some of its key theoretical notions. Crucially, by reconceptualizing the typology of depictions, we identify an overlooked domain in the literature: "speech-embedded nonverbal depictions," namely cases where meaning is communicated iconically, nonverbally, and without simultaneously co-occurring speech. In addition to contextualizing the phenomenon in relation to existing research, we demonstrate, with examples from American TV talk shows, how such depictions function in real-life language use, offering a brief sketch of their complexities and arguing also for their theoretical significance.

Embedding Silver Nanowires into a Hydroxypropyl Methyl Cellulose Film for Flexible Electrochromic Devices with High Electromechanical Stability

Transparent conductive films (TCFs) based on silver nanowires (AgNWs) are becoming one of the best candidates in realizing flexible optoelectronic devices. The AgNW-based TCF is usually prepared by coating AgNWs on a transparent polymer film; however, the coated AgNWs easily detach from the polymer underneath because of the weak adhesion between them. Herein, a network of AgNWs is embedded in the transparent hydroxypropyl methyl cellulose film, which has a strong adhesion with the AgNWs. The obtained TCF shows high optical transmittance (>85%), low roughness (rms = 4.8 ± 0.5 nm), and low haze (<0.2%). More importantly, owing to the embedding structure and strong adhesion, this TCF also shows excellent electromechanical stability, which is superior to the reported ones. Employing this TCF in a flexible electrochromic device, the obtained device exhibits excellent cyclic electromechanical stability and high coloring efficiency. Our work demonstrates a promising TCF with superior electromechanical stability for future applications in flexible optoelectronics.

Secure Patient Data Transfer Using Information Embedding and Hyperchaos

Health 4.0 is an extension of the Industry standard 4.0 which is aimed at the virtualization of health-care services. It employs core technologies and services for integrated management of electronic health records (EHRs), captured through various sensors. The EHR is processed and transmitted to distant experts for better diagnosis and improved healthcare delivery. However, for the successful implementation of Heath 4.0 many challenges do exist. One of the critical issues that needs attention is the security of EHRs in smart health systems. In this work, we have developed a new interpolation scheme capable of providing better quality cover media and supporting reversible EHR embedding. The scheme provides a double layer of security to the EHR by firstly using hyperchaos to encrypt the EHR. The encrypted EHR is reversibly embedded in the cover images produced by the proposed interpolation scheme. The proposed interpolation module has been found to provide better quality interpolated images. The proposed system provides an average peak signal to noise ratio (PSNR) of 52.38 dB for a high payload of 0.75 bits per pixel. In addition to embedding EHR, a fragile watermark (WM) is also encrypted using the hyperchaos embedded into the cover image for tamper detection and authentication of the received EHR. Experimental investigations reveal that our scheme provides improved performance for high contrast medical images (MI) when compared to various techniques for evaluation parameters like imperceptibility, reversibility, payload, and computational complexity. Given the attributes of the scheme, it can be used for enhancing the security of EHR in health 4.0.

Exploring the Limits of Combined Image/'omics Analysis for Non-cancer Histological Phenotypes

The last several years have witnessed an explosion of methods and applications for combining image data with 'omics data, and for prediction of clinical phenotypes. Much of this research has focused on cancer histology, for which genetic perturbations are large, and the signal to noise ratio is high. Related research on chronic, complex diseases is limited by tissue sample availability, lower genomic signal strength, and the less extreme and tissue-specific nature of intermediate histological phenotypes. Data from the GTEx Consortium provides a unique opportunity to investigate the connections among phenotypic histological variation, imaging data, and 'omics profiling, from multiple tissue-specific phenotypes at the sub-clinical level. Investigating histological designations in multiple tissues, we survey the evidence for genomic association and prediction of histology, and use the results to test the limits of prediction accuracy using machine learning methods applied to the imaging data, genomics data, and their combination. We find that expression data has similar or superior accuracy for pathology prediction as our use of imaging data, despite the fact that pathological determination is made from the images themselves. A variety of machine learning methods have similar performance, while network embedding methods offer at best limited improvements. These observations hold across a range of tissues and predictor types. The results are supportive of the use of genomic measurements for prediction, and in using the same target tissue in which pathological phenotyping has been performed. Although this last finding is sensible, to our knowledge our study is the first to demonstrate this fact empirically. Even while prediction accuracy remains a challenge, the results show clear evidence of pathway and tissue-specific biology.

Local Bacteriophage Delivery for Treatment and Prevention of Bacterial Infections

As viruses with high specificity for their bacterial hosts, bacteriophages (phages) are an attractive means to eradicate bacteria, and their potential has been recognized by a broad range of industries. Against a background of increasing rates of antibiotic resistance in pathogenic bacteria, bacteriophages have received much attention as a possible "last-resort" strategy to treat infections. The use of bacteriophages in human patients is limited by their sensitivity to acidic pH, enzymatic attack and short serum half-life. Loading phage within a biomaterial can shield the incorporated phage against many of these harmful environmental factors, and in addition, provide controlled release for prolonged therapeutic activity. In this review, we assess the different classes of biomaterials (i.e., biopolymers, synthetic polymers, and ceramics) that have been used for phage delivery and describe the processing methodologies that are compatible with phage embedding or encapsulation. We also elaborate on the clinical or pre-clinical data generated using these materials. While a primary focus is placed on the application of phage-loaded materials for treatment of infection, we also include studies from other translatable fields such as food preservation and animal husbandry. Finally, we summarize trends in the literature and identify current barriers that currently prevent clinical application of phage-loaded biomaterials.

Syntactic Comprehension of Relative Clauses and Center Embedding Using Pseudowords

Relative clause (RC) formation and center embedding (CE) are two primary syntactic operations fundamental for creating and understanding complex sentences. Ample evidence from previous cross-linguistic studies has revealed several similarities and differences between RC and CE. However, it is not easy to investigate the effect of pure syntactic constraints for RC and CE without the interference of semantic and pragmatic interactions. Here, we show how readers process CE and RC using a self-paced reading task in Korean. More interestingly, we adopted a novel self-paced pseudoword reading task to exploit syntactic operations of the RC and CE, eliminating the semantic and pragmatic interference in sentence comprehension. Our results showed that the main effects of RC and CE conform to previous studies. Furthermore, we found a facilitation effect of sentence comprehension when we combined an RC and CE in a complex sentence. Our study provides a valuable insight into how the purely syntactic processing of RC and CE assists comprehension of complex sentences.

Superpixel Embedding Network

Superpixel segmentation is a fundamental computer vision technique that finds application in a multitude of high level computer vision tasks. Most state-of-the-art superpixel segmentation methods are unsupervised in nature and thus cannot fully utilize frequently occurring texture patterns or incorporate multiscale context. In this paper, we show that superpixel segmentation can be improved by leveraging the superior modeling power of deep convolutional autoencoders in a fully unsupervised manner. We pose the superpixel segmentation problem as one of manifold learning where pixels that belong to similar texture patterns are assigned near identical embedding vectors. The proposed deep network is able to learn image-wide and dataset-wide feature patterns and the relationships between them. This knowledge is used to segment and group pixels in a way that is consistent with a more global definition of pattern coherence. Experiments demonstrate that the superpixels obtained from the embeddings learned by the proposed method outperform the state-of-theart superpixel segmentation methods for boundary precision and recall values. Additionally, we find that semantic edges obtained from the superpixel embeddings to be significantly better than the contemporary unsupervised approaches.

Using Low-Dimensional Manifolds to Map Relationships Between Dynamic Brain Networks

As the field of dynamic brain networks continues to expand, new methods are needed to allow for optimal handling and understanding of this explosion in data. We propose here a novel approach that embeds dynamic brain networks onto a two-dimensional (2D) manifold based on similarities and differences in network organization. Each brain network is represented as a single point on the low dimensional manifold with networks of similar topology being located in close proximity. The rich spatio-temporal information has great potential for visualization, analysis, and interpretation of dynamic brain networks. The fact that each network is represented by a single point makes it possible to switch between the low-dimensional space and the full connectivity of any given brain network. Thus, networks in a specific region of the low-dimensional space can be examined to identify network features, such as the location of brain network hubs or the interconnectivity between brain circuits. In this proof-of-concept manuscript, we show that these low dimensional manifolds contain meaningful information, as they were able to successfully discriminate between cognitive tasks and study populations. This work provides evidence that embedding dynamic brain networks onto low dimensional manifolds has the potential to help us better visualize and understand dynamic brain networks with the hope of gaining a deeper understanding of normal and abnormal brain dynamics.

Facile Fabrication of Highly Conductive, Ultrasmooth, and Flexible Silver Nanowire Electrode for Organic Optoelectronic Devices

So far, one of the fundamental limitations of silver nanowires (Ag NWs) is the high contact resistance among their junctions. Moreover, a rough surface due to its random arrangement is inevitable to electrical short when the nanowire-based electronics is driving. To improve the contact resistance, we suggest that the particle shape nanocrystals are intentionally reduced at the junctions by a localized joule-heat reduction approach from the silver ions. Via localized reductions, the reduced nanoparticles effectively weld the junction's areas, resulting in a 19% decrease in sheet resistance to 9.9 Ω sq^-1. Besides, the nanowires are embedded into a polyamide film with gentle hot pressing. Consequently, the roughness was considerably dropped so that it was successful to demonstrate organic light-emitting diodes (OLEDs) with nanowires, which was beneficial to be laminated with OLEDs under the low temperature. The experimental results show that the Ag NW-embedded films reach 10.9 Ω sq^-1 of the sheet resistance at 92% transmittance and the roughness was only 1.92 nm.