A*STAR Graduate Academy

The compact genome of Fugu rubripes has been sequenced to over 95% coverage, and more than 80% of the assembly is in multigene-sized scaffolds. In this 365-megabase vertebrate genome, repetitive DNA accounts for less than one-sixth of the sequence, and gene loci occupy about one-third of the genome. As with the human genome, gene loci are not evenly distributed, but are clustered into sparse and dense regions. Some "giant" genes were observed that had average coding sequence sizes but were spread over genomic lengths significantly larger than those of their human orthologs. Although three-quarters of predicted human proteins have a strong match to Fugu, approximately a quarter of the human proteins had highly diverged from or had no pufferfish homologs, highlighting the extent of protein evolution in the 450 million years since teleosts and mammals diverged. Conserved linkages between Fugu and human genes indicate the preservation of chromosomal segments from the common vertebrate ancestor, but with considerable scrambling of gene order.

The recent developments of lignin were reviewed in terms of different approaches to synthesize lignin-based copolymers, the resulting features and the potential applications of such copolymers.

Polymer encapsulated organic nanoparticles have recently attracted increasing attention in the biomedical field because of their unique optical properties, easy fabrication and outstanding performance as imaging and therapeutic agents. Of particular importance is the polymer encapsulated nanoparticles containing conjugated polymers (CP) or fluorogens with aggregation induced emission (AIE) characteristics as the core, which have shown significant advantages in terms of tunable brightness, superb photo- and physical stability, good biocompatibility, potential biodegradability and facile surface functionalization. In this review, we summarize the latest advances in the development of polymer encapsulated CP and AIE fluorogen nanoparticles, including preparation methods, material design and matrix selection, nanoparticle fabrication and surface functionalization for fluorescence and photoacoustic imaging. We also discuss their specific applications in cell labeling, targeted in vitro and in vivo imaging, blood vessel imaging, cell tracing, inflammation monitoring and molecular imaging. We specially focus on strategies to fine-tune the nanoparticle property (e.g. size and fluorescence quantum yield) through precise engineering of the organic cores and careful selection of polymer matrices. The review also highlights the merits and limitations of these nanoparticles as well as strategies used to overcome the limitations. The challenges and perspectives for the future development of polymer encapsulated organic nanoparticles are also discussed.

The recent progress, challenges and promising future on design, synthesis and fabrication of MnO₂for supercapacitors are reviewed and discussed.

Machine learning algorithms are often vulnerable to adversarial examples that have imperceptible alterations from the original counterparts but can fool the state-of-the-art models. It is helpful to evaluate or even improve the robustness of these models by exposing the maliciously crafted adversarial examples. In this paper, we present TextFooler, a simple but strong baseline to generate adversarial text. By applying it to two fundamental natural language tasks, text classification and textual entailment, we successfully attacked three target models, including the powerful pre-trained BERT, and the widely used convolutional and recurrent neural networks. We demonstrate three advantages of this framework: (1) effective—it outperforms previous attacks by success rate and perturbation rate, (2) utility-preserving—it preserves semantic content, grammaticality, and correct types classified by humans, and (3) efficient—it generates adversarial text with computational complexity linear to the text length.1

Aromaticity is an important concept to understand the stability and physical properties of π-conjugated molecules. Recent studies on pro-aromatic and anti-aromatic molecules revealed their irresistible tendency to become diradicals in the ground state. Diradical character thus becomes another very important concept and it is fundamentally correlated to the physical (optical, electronic and magnetic) properties and chemical reactivity of most of the organic optoelectronic materials. Molecules with distinctive diradical character show unique properties which are very different from those of traditional closed-shell π-conjugated systems, and thus they have many potential applications in organic electronics, spintronics, non-linear optics and energy storage. This critical review first introduces the fundamental electronic structure of Kekulé diradicals within the concepts of anti-aromaticity and pro-aromaticity in the context of Hückel aromaticity and diradical character. Then recent research studies on various stable/persistent diradicaloids based on pro-aromatic and anti-aromatic compounds are summarized and discussed with regard to their synthetic chemistry, physical properties, structure-property relationships and potential material applications. A summary and personal perspective is given at the end.

The development of the next generation of advanced lithium-ion batteries (LIBs) requires new & advanced materials and novel fabrication techniques in order to push the boundaries of performance and open up new and exciting markets. Structured carbon materials, with controlled pore features on the micron and nanometer scales, are explored as advanced alternatives to conventional graphite as the active material of the LIB anode. Mesoporous carbon materials, carbon nanotube-based materials, and graphene-based materials have been extensively investigated and reviewed. Morphology control (e.g., colloids, thin films, nanofibrous mats, monoliths) and hierarchical pores (particularly the presence of large pores) exhibit an increasing influence on LIB performance. This tutorial review focuses on the synthetic techniques for preparation of porous carbon spheres and carbon monoliths, including hydrothermal carbonization, emulsion templating, ice templating and new developments in making porous carbons from sustainable biomass and metal-organic framework templating. We begin with a brief introduction to LIBs, defining key parameters and terminology used to assess the performance of anode materials, and then address synthetic techniques for the fabrication of carbon spheres & monoliths and the relevant composites, followed, respectively, by a review of their performance as LIB anode materials. The review is completed with a prospective view on the possible direction of future research in this field.

Item-to-item collaborative filtering (aka.item-based CF) has been long used for building recommender systems in industrial settings, owing to its interpretability and efficiency in real-time personalization. It builds a user's profile as her historically interacted items, recommending new items that are similar to the user's profile. As such, the key to an item-based CF method is in the estimation of item similarities. Early approaches use statistical measures such as cosine similarity and Pearson coefficient to estimate item similarities, which are less accurate since they lack tailored optimization for the recommendation task. In recent years, several works attempt to learn item similarities from data, by expressing the similarity as an underlying model and estimating model parameters by optimizing a recommendation-aware objective function. While extensive efforts have been made to use shallow linear models for learning item similarities, there has been relatively less work exploring nonlinear neural network models for item-based CF. In this work, we propose a neural network model named Neural Attentive Item Similarity model (NAIS) for item-based CF. The key to our design of NAIS is an attention network, which is capable of distinguishing which historical items in a user profile are more important for a prediction. Compared to the state-of-the-art item-based CF method Factored Item Similarity Model (FISM) [1] , our NAIS has stronger representation power with only a few additional parameters brought by the attention network. Extensive experiments on two public benchmarks demonstrate the effectiveness of NAIS. This work is the first attempt that designs neural network models for item-based CF, opening up new research possibilities for future developments of neural recommender systems.

Abstract Electrocatalysts based on hierarchically structured and heteroatom‐doped non‐noble metal oxide materials are of great importance for efficient and low‐cost electrochemical water splitting systems. Herein, the synthesis of a series of hierarchical hollow nanoplates (NPs) composed of ultrathin Co 3 O 4 nanosheets doped with 13 different metal atoms is reported. The synthesis involves a cooperative etching−coordination−reorganization approach starting from zeolitic imidazolate framework‐67 (ZIF‐67) NPs. First, metal atom decorated ZIF‐67 NPs with unique cross‐channels are formed through a Lewis acid etching and metal species coordination process. Afterward, the composite NPs are converted to hollow Co 3 O 4 hierarchical NPs composed of ultrathin nanosheets through a solvothermal reaction, during which the guest metal species is doped into the octahedral sites of Co 3 O 4 . Density functional theory calculations suggest that doping of small amount of Fe atoms near the surface of Co 3 O 4 can greatly enhance the electrocatalytic activity toward the oxygen evolution reaction (OER). Benefiting from the structural and compositional advantages, the obtained Fe‐doped Co 3 O 4 hierarchical NPs manifest superior electrocatalytic performance for OER with an overpotential of 262 mV at 10 mA cm −2 , a Tafel slope of 43 mV dec −1 , and excellent stability even at a high current density of 100 mA cm −2 for 50 h.

The large availability of depth sensors provides valuable complementary information for salient object detection (SOD) in RGBD images. However, due to the inherent difference between RGB and depth information, extracting features from the depth channel using ImageNet pre-trained backbone models and fusing them with RGB features directly are sub-optimal. In this paper, we utilize contrast prior, which used to be a dominant cue in none deep learning based SOD approaches, into CNNs-based architecture to enhance the depth information. The enhanced depth cues are further integrated with RGB features for SOD, using a novel fluid pyramid integration, which can make better use of multi-scale cross-modal features. Comprehensive experiments on 5 challenging benchmark datasets demonstrate the superiority of the architecture CPFP over 9 state-of-the-art alternative methods.

A mono- to multilayer thick MoS₂ film has been grown by using the atomic layer deposition (ALD) technique at 300 °C on a sapphire wafer. ALD provides precise control of the MoS₂ film thickness due to pulsed introduction of the reactants and self-limiting reactions of MoCl₅ and H₂S. A post-deposition annealing of the ALD-deposited monolayer film improves the crystallinity of the film, which is evident from the presence of triangle-shaped crystals that exhibit strong photoluminescence in the visible range.

Mutations in the TP53 (p53) gene are present in a large fraction of human tumours, which frequently express mutant p53 proteins at high but heterogeneous levels. The clinical significance of this protein accumulation remains clouded. Mouse models bearing knock-in mutations of p53 have established that the mutant p53 proteins can drive tumour formation, invasion and metastasis through dominant negative inhibition of wild-type p53 as well as through gain of function or 'neomorphic' activities that can inhibit or activate the function of other proteins. These models have also shown that mutation alone does not confer stability, so the variable staining of mutant proteins seen in human cancers reflects tumour-specific activation of p53-stabilizing pathways. Blocking the accumulation and activity of mutant p53 proteins may thus provide novel cancer therapeutic and diagnostic targets, but their induction by chemotherapy may paradoxically limit the effectiveness of these treatments.

The delivery of genetic materials into cells to elicit cellular responses has been extensively studied by biomaterials scientists globally. Many materials such as lipids, peptides, viruses, synthetically modified cationic polymers and certain inorganic nanomaterials could be used to complex the negatively charged plasmids and deliver the formed package into cells. The recent literature on the delivery of genetic materials utilising inorganic nanoparticles is carefully examined in this review. We have picked out the most relevant references and concisely summarised the findings with illustrated examples. We further propose alternative approaches and suggest future pathways towards the practical use of multifunctional nanocarriers.

Abstract The causes of impaired skeletal muscle mass and strength during aging are well-studied in healthy populations. Less is known on pathological age-related muscle wasting and weakness termed sarcopenia, which directly impacts physical autonomy and survival. Here, we compare genome-wide transcriptional changes of sarcopenia versus age-matched controls in muscle biopsies from 119 older men from Singapore, Hertfordshire UK and Jamaica. Individuals with sarcopenia reproducibly demonstrate a prominent transcriptional signature of mitochondrial bioenergetic dysfunction in skeletal muscle, with low PGC-1α/ERRα signalling, and downregulation of oxidative phosphorylation and mitochondrial proteostasis genes. These changes translate functionally into fewer mitochondria, reduced mitochondrial respiratory complex expression and activity, and low NAD + levels through perturbed NAD + biosynthesis and salvage in sarcopenic muscle. We provide an integrated molecular profile of human sarcopenia across ethnicities, demonstrating a fundamental role of altered mitochondrial metabolism in the pathological loss of skeletal muscle mass and function in older people.

2D transition metal dichalcogenides (2D TMDs) (MoS₂, WS₂,<italic>etc.</italic>) have attracted considerable attention recently due to their unique structures, strong chemical stability and attractive semiconducting characteristics.

The strong π-π interactions in the stacking layers of two-dimensional covalent organic frameworks (2D-COFs), together with rotationally labile imine linkages, make most of the solid state imine-linked COFs non-fluorescent due to fluorescence quenching processes. Here, we report the successful synthesis of highly photoluminescent imine-based 2D-COFs by integrating a non-planar building unit with a pyrene-based unit and transforming the COF into spherical, sub-micron particles. High photoluminescence quantum yields (PLQY) were achieved with COF sub-micron particles dispersed in organic solvents. The as-prepared COF sub-micron particles can be used as a chemical sensor for the detection of explosive chemicals, with high sensitivity and selectivity (up to ppm level).

A low-cost, intense, broadband, noise resistive, magnetic field controllable, flexible, and low power driven THz emitter based on thin nonmagnetic/ferromagnetic metallic heterostructures is demonstrated. The THz emission origins from the inverse spin Hall Effect. The proposed devices are not only promising for a wide range of THz equipment, but also offer an alternative approach to characterize the spin-orbit interaction in nonmagnetic/ferromagnetic bilayers. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Radio-frequency (RF) enabled wireless energy transfer (WET), as a promising solution to provide cost-effective and reliable power supplies for energy-constrained wireless networks, has drawn growing interests recently. To overcome the significant propagation loss over distance, employing multi-antennas at the energy transmitter (ET) to more efficiently direct wireless energy to desired energy receivers (ERs), termed energy beamforming, is an essential technique for enabling WET. However, the achievable gain of energy beamforming crucially depends on the available channel state information (CSI) at the ET, which needs to be acquired practically. In this paper, we study the design of an efficient channel acquisition method for a point-to-point multiple-input multiple-output (MIMO) WET system by exploiting the channel reciprocity, i.e., the ET estimates the CSI via dedicated reverse-link training from the ER. Considering the limited energy availability at the ER, the training strategy should be carefully designed so that the channel can be estimated with sufficient accuracy, and yet without consuming excessive energy at the ER. To this end, we propose to maximize the net harvested energy at the ER, which is the average harvested energy offset by that used for channel training. An optimization problem is formulated for the training design over MIMO Rician fading channels, including the subset of ER antennas to be trained, as well as the training time and power allocated. Closed-form solutions are obtained for some special scenarios, based on which useful insights are drawn on when training should be employed to improve the net transferred energy in MIMO WET systems.

An electrochemically favorable Ni(OH)2 with porously hierarchical structure and ultrathin nanosheets in a core–shell structure H-TiO2@Ni(OH)2 is achieved through modulating the surface chemical activity of TiO2 by hydrogenation, which creates a defect-rich surface of TiO2, thereby facilitating the subsequent nucleation and growth of Ni(OH)2. These configuration-tailored H-TiO2@Ni(OH)2 core–shell nanowires exhibit a superior electrochemical performance and good flexibility. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Over the past few years, state-of-the-art image segmentation algorithms are based on deep convolutional neural networks. To render a deep network with the ability to understand a concept, humans need to collect a large amount of pixel-level annotated data to train the models, which is time-consuming and tedious. Recently, few-shot segmentation is proposed to solve this problem. Few-shot segmentation aims to learn a segmentation model that can be generalized to novel classes with only a few training images. In this paper, we propose a cross-reference network (CRNet) for few-shot segmentation. Unlike previous works which only predict the mask in the query image, our proposed model concurrently makes predictions for both the support image and the query image. With a cross-reference mechanism, our network can better find the co-occurrent objects in two images, thus helping the few-shot segmentation task. We also develop a mask refinement module to recurrently refine the prediction of the foreground regions. For the k-shot learning, we propose to finetune parts of networks to take advantage of multiple labeled support images. Experiments on the PASCAL VOC 2012 dataset show that our network achieves state-of-the-art performance.