TransPTM: a transformer-based model for non-histone acetylation site prediction.

Protein acetylation is one of the extensively studied post-translational modifications (PTMs) due to its significant roles across a myriad of biological processes. Although many computational tools for acetylation site identification have been developed, there is a lack of benchmark dataset and bespoke predictors for non-histone acetylation site prediction. To address these problems, we have contributed to both dataset creation and predictor benchmark in this study. First, we construct a non-histone acetylation site benchmark dataset, namely NHAC, which includes 11 subsets according to the sequence length ranging from 11 to 61 amino acids. There are totally 886 positive samples and 4707 negative samples for each sequence length. Secondly, we propose TransPTM, a transformer-based neural network model for non-histone acetylation site predication. During the data representation phase, per-residue contextualized embeddings are extracted using ProtT5 (an existing pre-trained protein language model). This is followed by the implementation of a graph neural network framework, which consists of three TransformerConv layers for feature extraction and a multilayer perceptron module for classification. The benchmark results reflect that TransPTM has the competitive performance for non-histone acetylation site prediction over three state-of-the-art tools. It improves our comprehension on the PTM mechanism and provides a theoretical basis for developing drug targets for diseases. Moreover, the created PTM datasets fills the gap in non-histone acetylation site datasets and is beneficial to the related communities. The related source code and data utilized by TransPTM are accessible at https://www.github.com/TransPTM/TransPTM.

Discovering DNA shape motifs with multiple DNA shape features: generalization, methods, and validation.

DNA motifs are crucial patterns in gene regulation. DNA-binding proteins (DBPs), including transcription factors, can bind to specific DNA motifs to regulate gene expression and other cellular activities. Past studies suggest that DNA shape features could be subtly involved in DNA-DBP interactions. Therefore, the shape motif annotations based on intrinsic DNA topology can deepen the understanding of DNA-DBP binding. Nevertheless, high-throughput tools for DNA shape motif discovery that incorporate multiple features altogether remain insufficient. To address it, we propose a series of methods to discover non-redundant DNA shape motifs with the generalization to multiple motifs in multiple shape features. Specifically, an existing Gibbs sampling method is generalized to multiple DNA motif discovery with multiple shape features. Meanwhile, an expectation-maximization (EM) method and a hybrid method coupling EM with Gibbs sampling are proposed and developed with promising performance, convergence capability, and efficiency. The discovered DNA shape motif instances reveal insights into low-signal ChIP-seq peak summits, complementing the existing sequence motif discovery works. Additionally, our modelling captures the potential interplays across multiple DNA shape features. We provide a valuable platform of tools for DNA shape motif discovery. An R package is built for open accessibility and long-lasting impact: https://zenodo.org/doi/10.5281/zenodo.10558980.

CoaDTI: multi-modal co-attention based framework for drug-target interaction annotation.

MOTIVATION: The identification of drug-target interactions (DTIs) plays a vital role for in silico drug discovery, in which the drug is the chemical molecule, and the target is the protein residues in the binding pocket. Manual DTI annotation approaches remain reliable; however, it is notoriously laborious and time-consuming to test each drug-target pair exhaustively. Recently, the rapid growth of labelled DTI data has catalysed interests in high-throughput DTI prediction. Unfortunately, those methods highly rely on the manual features denoted by human, leading to errors. RESULTS: Here, we developed an end-to-end deep learning framework called CoaDTI to significantly improve the efficiency and interpretability of drug target annotation. CoaDTI incorporates the Co-attention mechanism to model the interaction information from the drug modality and protein modality. In particular, CoaDTI incorporates transformer to learn the protein representations from raw amino acid sequences, and GraphSage to extract the molecule graph features from SMILES. Furthermore, we proposed to employ the transfer learning strategy to encode protein features by pre-trained transformer to address the issue of scarce labelled data. The experimental results demonstrate that CoaDTI achieves competitive performance on three public datasets compared with state-of-the-art models. In addition, the transfer learning strategy further boosts the performance to an unprecedented level. The extended study reveals that CoaDTI can identify novel DTIs such as reactions between candidate drugs and severe acute respiratory syndrome coronavirus 2-associated proteins. The visualization of co-attention scores can illustrate the interpretability of our model for mechanistic insights. AVAILABILITY: Source code are publicly available at https://github.com/Layne-Huang/CoaDTI.

Developing Isoxazole as a Native Photo-Cross-Linker for Photoaffinity Labeling and Chemoproteomics.

Photoaffinity labeling is a powerful technique to interrogate drug-protein interactions in native cellular environments. Photo-cross-linkers are instrumental for this technique. However, the introduction of unnatural photo-cross-linkers may significantly reduce the bioactivity of the drug, thus impairing the chemoproteomic outcomes. Herein, we developed a common pharmacophore, isoxazole, into a natively embedded photo-cross-linker for chemoproteomics, which minimally perturbs the drug structure. The photo-cross-linking reactions of the isoxazole were thoroughly investigated for the first time. Functionalized isoxazoles were then designed and applied to protein labeling, demonstrating the superior photo-cross-linking efficiency. Subsequently, two isoxazole-based drugs, Danazol and Luminespib, were employed in chemoproteomic studies, revealing their potential cellular targets. These results provide valuable strategies for future chemoproteomic study and drug development.

Design, synthesis and biological evaluation of amino acids-oleanolic acid conjugates as influenza virus inhibitors.

Viral entry inhibitors are of great importance in current efforts to develop a new generation of anti-influenza drugs. Inspired by the discovery of a series of pentacyclic triterpene derivatives as entry inhibitors targeting the HA protein of influenza virus, we designed and synthesized 32 oleanolic acid (OA) analogues in this study by conjugating different amino acids to the 28-COOH of OA. The antiviral activity of these compounds was evaluated in vitro. Some of these compounds revealed impressive anti-influenza potencies against influenza A/WSN/33 (H1N1) virus. Among them, compound 15a exhibited robust potency and broad antiviral spectrum with IC50 values at the low-micromolar level against four different influenza strains. Hemagglutination inhibition (HI) assay and docking experiment indicated that these OA analogues may act in the same way as their parent compound by interrupting the interaction between HA protein of influenza virus and the host cell sialic acid receptor via binding to HA, thus blocking viral entry.

Synthesis, Structure Activity Relationship and Anti-influenza A Virus Evaluation of Oleanolic Acid-Linear Amino Derivatives.

Oleanolic acid (OA) was discovered as a mild influenza hemagglutinin (HA) inhibitor in our earlier studies. In the present work, 20 compounds were prepared by structural modifications of OA, and their antiviral activities against influenza A/WSN/33 (H1N1) virus in Madin-Darby canine kidney (MDCK) cells were evaluated. Based on the biological result, structure-activity relationship (SAR) was discussed. Compound 10 with six-carbon chain and a terminal hydroxyl group showed the strongest anti-influenza activity with an IC50 of 2.98 µM, which is an order of magnitude more potent than OA. Hemagglutination inhibition and Surface plasmon resonance (SPR) assay indicated that compound 10 might interfere with influenza invasion by interacting with HA protein.

A straightforward and CMOS-compatible nanofabrication technique of periodic SiO2 nanohole arrays.

In this work, we have demonstrated a straightforward and CMOS-compatible nanofabrication technique that can produce well-ordered periodic SiO2 nanohole arrays in wafer-scale using a single amorphous silicon (α-Si) layer. It is the first time that α-Si material has been used as an etch mask to fabricate SiO2 nanostructures. Our results have shown that the diameter and shape of SiO2 nanohole arrays, with vertical and smooth sidewalls, can be precisely controlled by an optimized two-step etch process. The diameter and pitch of nanoholes as small as 45 nm and 140 nm, respectively, have been successfully achieved. Moreover, the technique is independent of any specific lithographic approaches and, therefore, is capable of fabricating SiO2 nanohole arrays with smaller diameters and higher densities. Furthermore, since our approach is completely metal-free, it can be incorporated and integrated very easily into the standard semiconductor industry. It has a potential for wide applications in micro-nanofabrication, and represents a big step towards mass production.

A Lightweight Framework For Chromatin Loop Detection at the Single-Cell Level.

Single-cell Hi-C (scHi-C) has made it possible to analyze chromatin organization at the single-cell level. However, scHi-C experiments generate inherently sparse data, which poses a challenge for loop calling methods. The existing approach performs significance tests across the imputed dense contact maps, leading to substantial computational overhead and loss of information at the single-cell level. To overcome this limitation, a lightweight framework called scGSLoop is proposed, which sets a new paradigm for scHi-C loop calling by adapting the training and inferencing strategies of graph-based deep learning to leverage the sequence features and 1D positional information of genomic loci. With this framework, sparsity is no longer a challenge, but rather an advantage that the model leverages to achieve unprecedented computational efficiency. Compared to existing methods, scGSLoop makes more accurate predictions and is able to identify more loops that have the potential to play regulatory roles in genome functioning. Moreover, scGSLoop preserves single-cell information by identifying a distinct group of loops for each individual cell, which not only enables an understanding of the variability of chromatin looping states between cells, but also allows scGSLoop to be extended for the investigation of multi-connected hubs and their underlying mechanisms.

Universal Strategy to Develop Fluorogenic Probes for Lysine Deacylase/Demethylase Activity and Application in Discriminating Demethylation States.

Dynamically controlling the post-translational modification of the ε-amino groups of lysine residues is critical for regulating many cellular events. Increasing studies have revealed that many important diseases, including cancer and neurological disorders, are associated with the malfunction of lysine deacylases and demethylases. Developing fluorescent probes that are capable of detecting lysine deacylase and demethylase activity is highly useful for interrogating their roles in epigenetic regulation and diseases. Due to the distinct substrate recognition of these epigenetic eraser enzymes, designing a universal strategy for detecting their activity poses substantial difficulty. Moreover, designing activity-based probes for differentiating their demethylation states is even more challenging and still remains largely unexplored. Herein, we report a universal strategy to construct probes that can detect the enzymatic activity of epigenetic "erasers" through NBD-based long-distance intramolecular reactions. The probes can be easily prepared by installing the O-NBD group at the C-terminal residue of specific peptide substrates by click chemistry. Based on this strategy, detecting the activity of lysine deacetylase, desuccinylase, or demethylase with superior sensitivity and selectivity has been successfully achieved through single-step probe development. Furthermore, the demethylase probe based on this strategy is capable of distinguishing different demethylation states by both absorption and fluorescence lifetime readout. We envision that these newly developed probes will provide powerful tools to facilitate drug discovery in epigenetics in the future.

Study of the ordered assembly morphologies of diblock copolymers on the same substrate.

With the development of frontier technology in emerging semiconductor processes, self-assembling (SA) and directed self-assembly (DSA) of block copolymers (BCPs) have attracted great attention from scientific researchers and become promising candidates for advanced photolithography. Using an optimal coating and baking process, highly ordered assembly morphologies (e.g., cylinder and lamella) of two BCPs in thin films were obtained without an additional topcoat material layer. Moreover, the whole experimental study also provides an optimal process for integrating the two BCPs into the same topographic guiding pattern substrate fabricated by electron beam lithography (EBL) to achieve specific self-assembly. This topographic guiding substrate achieves not only lamellar micro-domains aligned perpendicular to the sidewalls of trench edges but also cylindrical micro-domains (PMMA phase in a PS matrix) aligned parallel to trench edges respectively, which provides insights and valuable information for further applications in lithography and electronic devices.

Mini-review: Recent advances in post-translational modification site prediction based on deep learning.

Post-translational modifications (PTMs) are closely linked to numerous diseases, playing a significant role in regulating protein structures, activities, and functions. Therefore, the identification of PTMs is crucial for understanding the mechanisms of cell biology and diseases therapy. Compared to traditional machine learning methods, the deep learning approaches for PTM prediction provide accurate and rapid screening, guiding the downstream wet experiments to leverage the screened information for focused studies. In this paper, we reviewed the recent works in deep learning to identify phosphorylation, acetylation, ubiquitination, and other PTM types. In addition, we summarized PTM databases and discussed future directions with critical insights.

Self-assembled nano-photosensitizer for targeted, activatable, and biosafe cancer phototheranostics.

Cancer treatment currently still faces crucial challenges in therapeutic effectiveness, precision, and complexity. Photodynamic therapy (PDT) as a non-invasive tactic has earned widespread popularity for its excellent therapeutic output, flexibility, and restrained toxicity. Nonetheless, drawbacks, including low efficiency, poor cancer specificity, and limited therapeutic depth, remain considerable during the cancer treatment. Although great effort has been made to improve the performance, the overall efficiency and biosafety are still ambiguous and unable to meet urgent clinical needs. Herein, this study integrates merits from previous PDT strategies and develops a cancer-targeting, activatable, biosafe photosensitizer. Owing to excellent self-assembly ability, this photosensitizer can be conveniently prepared as multifunctional nano-photosensitizers, namely MBNPs, and applied to in vivo cancer phototheranostics in "all-in-one" mode. This study successfully verifies the mechanism of MBNPs, then deploys them to cell-based and in vivo cancer PDT. Based on the unique cancer microenvironment, MBNPs achieve precise distribution, accumulation, and activation towards the tumor, releasing methylene blue as a potent photosensitizer for phototherapy. The PDT outcome demonstrates MBNPs' superior cancer specificity, remarkable PDT efficacy, and negligible toxicity. Meanwhile, in vivo NIR fluorescence and photoacoustic imaging have been utilized to guide the PDT treatment synergistically. Additionally, the biosafety of the MBNPs-based PDT treatment is ensured, thus providing potential for future clinical studies.

Study of the perpendicular self-assembly of a novel high-χ block copolymer without any neutral layer on a silicon substrate.

A novel type of high-χ block copolymer, polystyrene-block-polycarbonate (PS-b-PC), which contains an active -NH- group on the polymer backbone between the PS block and the PC block, has been successfully synthesized. Vertical micro-phase separation can be successfully achieved on Si substrates with neutral-layer-free materials with a pitch of 16.8 nm. Water contact angle experiments indicate that PS and PC have approximate surface energy values on Si substrates. A hydrogen bond mechanism has been proposed for the formation of a periodic and lamella-forming phase separation structure, with the domains oriented perpendicular to the substrate. A combination of both theory and experimental verification proves that the hydrogen bonding plays a dominant role as a real driving force to promote vertical micro-phase separation in the absence of a neutral layer. Subsequently, the study of a novel block copolymer on four different types of substrate without any neutral layer further confirms that the newly synthesized material enables greater flexibility and potential applications for the fabrication of various nanostructures and functional electronic devices in a simple, cost-effective and efficient way, which is of considerable importance to contemporary and emerging technology applications.

Design, synthesis of oleanolic acid-saccharide conjugates using click chemistry methodology and study of their anti-influenza activity.

The development of entry inhibitors is an emerging approach to the inhibition of influenza virus. In our previous research, oleanolic acid (OA) was discovered as a mild influenza hemagglutinin (HA) inhibitor. Herein, as a further study, we report the preparation of a series of OA-saccharide conjugates via the CuAAC reaction, and the anti-influenza activity of these compounds was evaluated in vitro. Among them, compound 11b, an OA-glucose conjugate, showed a significantly increased anti-influenza activity with an IC50 of 5.47 µM, and no obvious cytotoxic effect on MDCK cells was observed at 100 µM. Hemagglutination inhibition assay and docking experiment indicated that 11b might interfere with influenza virus infection by acting on HA protein. Broad-spectrum anti-influenza experiments showed 11b to be robustly potent against 5 different strains, including influenza A and B viruses, with IC50 values at the low-micromole level. Overall, this finding further extends the utility of OA-saccharide conjugates in anti-influenza virus drug design.

A Novel Nanofabrication Technique of Silicon-Based Nanostructures.

A novel nanofabrication technique which can produce highly controlled silicon-based nanostructures in wafer scale has been proposed using a simple amorphous silicon (α-Si) material as an etch mask. SiO2 nanostructures directly fabricated can serve as nanotemplates to transfer into the underlying substrates such as silicon, germanium, transistor gate, or other dielectric materials to form electrically functional nanostructures and devices. In this paper, two typical silicon-based nanostructures such as nanoline and nanofin have been successfully fabricated by this technique, demonstrating excellent etch performance. In addition, silicon nanostructures fabricated above can be further trimmed to less than 10 nm by combing with assisted post-treatment methods. The novel nanofabrication technique will be expected a new emerging technology with low process complexity and good compatibility with existing silicon integrated circuit and is an important step towards the easy fabrication of a wide variety of nanoelectronics, biosensors, and optoelectronic devices.

Novel 14-nm Scallop-Shaped FinFETs (S-FinFETs) on Bulk-Si Substrate.

In this study, novel p-type scallop-shaped fin field-effect transistors (S-FinFETs) are fabricated using an all-last high-k/metal gate (HKMG) process on bulk-silicon (Si) substrates for the first time. In combination with the structure advantage of conventional Si nanowires, the proposed S-FinFETs provide better electrostatic integrity in the channels than normal bulk-Si FinFETs or tri-gate devices with rectangular or trapezoidal fins. It is due to formation of quasi-surrounding gate electrodes on scalloping fins by a special Si etch process. The entire integration flow of the S-FinFETs is fully compatible with the mainstream all-last HKMG FinFET process, except for a modified fin etch process. The drain-induced barrier lowering and subthreshold swing of the fabricated p-type S-FinFETs with a 14-nm physical gate length are 62 mV/V and 75 mV/dec, respectively, which are much better than those of normal FinFETs with a similar process. With an improved short-channel-effect immunity in the channels due to structure modification, the novel structure provides one of possibilities to extend the FinFET scalability to sub-10-nm nodes with little additional process cost.

CMOS-Compatible Top-Down Fabrication of Periodic SiO2 Nanostructures using a Single Mask.

We propose a CMOS-compatible top-down fabrication technique of highly-ordered and periodic SiO2 nanostructures using a single amorphous silicon (α-Si) mask layer. The α-Si mask pattern is precisely transferred into the underlying SiO2 substrate material with a high fidelity by a novel top-down fabrication. It is the first time for α-Si film used as an etch mask to fabricate SiO2 nanostructures including nanoline, nanotrench, and nanohole arrays. It is observed that the α-Si mask can significantly reduce the pattern edge roughness and achieve highly uniform and smooth sidewalls. This behavior may be attributed to the presence of high concentration of dangling bonds in α-Si mask surface. By controlling the process condition, it is possible to achieve a desired vertical etched profile with a controlled size. Our results demonstrate that SiO2 pattern as small as sub-20 nm may be achievable. The obtained SiO2 pattern can be further used as a nanotemplate to produce periodic or more complex silicon nanostructures. Moreover, this novel top-down approach is a potentially universal method that is fully compatible with the currently existing Si-based CMOS technologies. It offers a greater flexibility for the fabrication of various nanoscale devices in a simple and efficient way.