A5052610760- Machine Learning in Bioinformatics
- RNA and protein synthesis mechanisms
- Genomics and Phylogenetic Studies
- Computational Drug Discovery Methods
- Protein Structure and Dynamics
- Bacterial Genetics and Biotechnology
- Antibiotic Resistance in Bacteria
- Human Pose and Action Recognition
- Advanced Vision and Imaging
- Generative Adversarial Networks and Image Synthesis
- Bioinformatics and Genomic Networks
- Diabetic Foot Ulcer Assessment and Management
- Monoclonal and Polyclonal Antibodies Research
- Image Enhancement Techniques
- Human Motion and Animation
- Computer Graphics and Visualization Techniques
- Machine Learning in Materials Science
- vaccines and immunoinformatics approaches
- Multimodal Machine Learning Applications
- Circular RNAs in diseases
- Respiratory viral infections research
- 2D Materials and Applications
- Gut microbiota and health
- Genomics and Chromatin Dynamics
- MXene and MAX Phase Materials
Nanjing Agricultural University
2023-2025
Suzhou Municipal Hospital
2024-2025
Shenzhen University
2024-2025
Nanjing University of Science and Technology
2019-2024
First Hospital of Jilin University
2023-2024
Jilin University
2023-2024
Zhejiang University
2024
Zhejiang University of Science and Technology
2023
University of Geneva
2023
Johns Hopkins University
2023
In this work, we propose a new solution to 3D human pose estimation in videos. Instead of directly regressing the joint locations, draw inspiration from skeleton anatomy and decompose task into bone direction prediction length prediction, which locations can be completely derived. Our motivation is fact that lengths remain consistent across time. This promotes us develop effective techniques utilize global information <i>all</i> frames video for high-accuracy prediction. Moreover, network,...
Accurate identification of protein function is critical to elucidate life mechanisms and design new drugs. We proposed a novel deep-learning method, ATGO, predict Gene Ontology (GO) attributes proteins through triplet neural-network architecture embedded with pre-trained language models from sequences. The method was systematically tested on 1068 non-redundant benchmarking 3328 targets the third Critical Assessment Protein Function Annotation (CAFA) challenge. Experimental results showed...
Abstract Efficient and accurate recognition of protein–DNA interactions is vital for understanding the molecular mechanisms related biological processes further guiding drug discovery. Although current experimental protocols are most precise way to determine binding sites, they tend be labor-intensive time-consuming. There an immediate need design efficient computational approaches predicting DNA-binding sites. Here, we proposed ULDNA, a new deep-learning model, deduce sites from protein...
Accurate identification of protein–DNA binding sites is significant for both understanding protein function and drug design. Machine-learning-based methods have been extensively used the prediction sites. However, data imbalance problem, in which number nonbinding residues (negative-class samples) far larger than that (positive-class samples), seriously restricts performance improvements machine-learning-based predictors. In this work, we designed a two-stage imbalanced learning algorithm,...
Blue phosphorene (blue-P), an allotrope of black phosphorene, is prone to oxidize under ambient conditions, which significantly hinders its incorporation in anode for Li/Na ion batteries (LIBs/NIBs). Combining blue-P and hexagonal boron nitride (h-BN) together construct h-BN/blue-P heterostructure (BN/P) can break the limitation restricted properties blue-P. By means first-principles computations, we explored potential using BN/P as material LIBs/NIBs. Our computations show that adsorption...
Accurately identifying DNA-binding proteins (DBPs) from protein sequence information is an important but challenging task for function annotations. In this paper, we establish a novel computational method, named TargetDBP, accurately targeting DBPs primary sequences. four single-view features, i.e., AAC (Amino Acid Composition), PsePSSM (Pseudo Position-Specific Scoring Matrix), PsePRSA Predicted Relative Solvent Accessibility), and PsePPDBS Probabilities of DNA-Binding Sites), are first...
We present a unified and flexible framework to address the generalized problem of 3D motion synthesis that covers tasks prediction, completion, interpolation, spatial-temporal recovery. Since these have different input constraints various fidelity diversity requirements, most existing approaches only cater specific task or use architectures tasks. Here we propose based on Conditional Variational Auto-Encoder (CVAE), where treat any arbitrary as masked series. Notably, by considering this...
In this paper, we explore open-domain sketch-to-photo translation, which aims to synthesize a realistic photo from freehand sketch with its class label, even if the sketches of that are missing in training data. It is challenging due lack supervision and large geometric distortion between domains. To absent photos, propose framework jointly learns photo-to-sketch generation. However, generator trained fake might lead unsatisfying results when dealing classes, domain gap synthesized real...
The accurate identification of protein-nucleotide binding residues is crucial for protein function annotation and drug discovery. Numerous computational methods have been proposed to predict these residues, achieving remarkable performance. However, due the limited availability high variability nucleotides, predicting diverse nucleotides remains a significant challenge. To address these, we propose NucGMTL, new grouped deep multi-task learning approach designed all observed in BioLiP...
Pulmonary hypertension (PH) is a devastating disease marked by elevated pulmonary artery pressure, resulting in right ventricular (RV) failure and mortality. Despite the identification of several dysregulated genes PH, involvement circular RNAs (circRNAs), subset long noncoding RNAs, remains largely unknown. In this study, high-throughput RNA sequencing was performed to analyze genome-wide expression patterns circRNAs arteries from three models PH rats induced hypoxia (Hyp),...
Background: This study aims to analyze the vaccination status and factors influencing delayed among toddlers born hepatitis B surface antigen (HBsAg)-positive mothers. Methods: Data of HBsAg-positive mothers between 1 January 2021 31 December 2022 were provided by Suzhou Maternal Child Health Care Family Planning Service Center. The records obtained from Jiangsu Province Immunization Management Information System. Logistic regression analysis was used vaccination. Results: A total 4250...
Accurately identifying protein functions is essential to understand life mechanisms and thus advance drug discovery. Although biochemical experiments are the gold standard for determining functions, they often time-consuming labor-intensive. Here, we proposed a novel composite deep-learning method, MKFGO, infer Gene Ontology (GO) attributes through integrating five complementary pipelines built on multi-source biological data. MKFGO was rigorously benchmarked 1522 non-redundant proteins,...
Drug discovery faces increasing challenges in identifying novel drug candidates satisfying multiple stringent objectives, such as binding affinity, protein target selectivity, and drug-likeness. Existing optimization methods struggle with the complexity of handling numerous limiting advancements molecular design most algorithms are only effective for up to four objectives. To overcome these limitations, study introduces Pareto Monte Carlo Tree Search Molecular Generation (PMMG) method,...
Antibodies defend our health by binding to antigens with high specificity and potentiality, primarily relying on the Complementarity-Determining Region (CDR). Yet, current experimental methods of discovering new antibody CDRs are heavily time-consuming. Computational design could alleviate this burden; especially, protein language models have proven quite beneficial in many recent studies. However, most existing solely focus potentiality struggle encapsulate diverse range plausible CDR...
Transmembrane proteins have critical biological functions and play a role in multitude of cellular processes including cell signaling, transport molecules ions across membranes. Approximately 60% transmembrane are considered as drug targets. Missense mutations such can lead to many diverse diseases disorders, neurodegenerative cystic fibrosis. However, there limited studies on proteins. In this work, we first design new feature encoding method, termed weight attenuation position-specific...
We propose PAniC-3D, a system to reconstruct stylized 3D character heads directly from illustrated (p)ortraits of (ani)me (c)haracters. Our anime-style domain poses unique challenges single-view reconstruction; compared natural images human heads, portrait illustrations have hair and accessories with more complex diverse geometry, are shaded non-photorealistic contour lines. In addition, there is lack both model illustration data suitable train evaluate this ambiguous reconstruction task....
Protein–DNA interactions exist ubiquitously and play important roles in the life cycles of living cells. The accurate identification DNA-binding proteins (DBPs) is one key steps to understand mechanisms protein–DNA interactions. Although many DBP methods have been proposed, current performance still unsatisfactory. In this study, a new method, called TargetDBP+, developed further enhance identifying DBPs. five convolutional features are first extracted from feature sources, i.e., amino acid...
RNA N6-methyladenosine is a prevalent and abundant type of modification that exerts significant influence on diverse biological processes. To date, numerous computational approaches have been developed for predicting methylation, with most them ignoring the correlations different encoding strategies failing to explore adaptability various attention mechanisms methylation identification. solve above issues, we proposed an innovative framework m6A site, termed BLAM6A-Merge. Specifically, it...
Abstract Objective To investigate the epidemiological characteristics and infections of respiratory syncytial virus (RSV) influenza viruses in hospitalized elderly patients with tract Suzhou City, China, to compare differences clinical economic burden associated these two infections. Methods In this prospective study, pathogenetic testing data for aged 60 years older were collected five hospitals through stratified cluster sampling from December 2023 May 2024. Comparative study on epidemic...
Abstract X-ray crystallography is the major approach for determining atomic-level protein structures. Because not all proteins can be easily crystallized, accurate prediction of crystallization propensity provides critical help in guiding experimental design and improving success rate experiments. This study has developed a new machine-learning-based pipeline that uses newly deep-cascade forest (DCF) model with multiple types sequence-based features to predict propensity. Based on pipeline,...
Missense mutation (MM) may lead to various human diseases by disabling proteins. Accurate prediction of MM is important and challenging for both protein function annotation drug design. Although several computational methods yielded acceptable success rates, there still room further enhancing the performance MM.In present study, we designed a new feature extracting method, which considers impact degree residues in microenvironment range site. Stringent cross-validation independent test on...
Protein fold recognition is a critical step toward protein structure and function prediction, aiming at providing the most likely type of query protein. In recent years, development deep learning (DL) technique has led to massive advances in this important field, accordingly, sensitivity been dramatically improved. Most DL-based methods take an intermediate bottleneck layer as feature representation proteins with new types. However, strategy indirect, inefficient conditional on hypothesis...
Accurate identification of transcription factor binding sites is great significance in understanding gene expression, biological development and drug design. Although a variety methods based on deep-learning models large-scale data have been developed to predict DNA sequences, there room for further improvement prediction performance. In addition, effective interpretation greatly desirable. Here we present MAResNet, new method, predicting 690 ChIP-seq datasets. More specifically, MAResNet...