The accurate prediction of drug absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties represents a crucial step in early drug development for reducing failure risk. Current deep learning approaches face challenges with data sparsity and information loss due to single-molecule representation limitations and isolated predictive tasks. This research proposes molecular properties prediction with parallel-view and collaborative learning (MolP-PC), a multi-view fusion and multi-task deep learning framework that integrates 1D molecular fingerprints (MFs), 2D molecular graphs, and 3D geometric representations, incorporating an attention-gated fusion mechanism and multi-task adaptive learning strategy for precise ADMET property predictions. Experimental results demonstrate that MolP-PC achieves optimal performance in 27 of 54 tasks, with its multi-task learning (MTL) mechanism significantly enhancing predictive performance on small-scale datasets and surpassing single-task models in 41 of 54 tasks. Additional ablation studies and interpretability analyses confirm the significance of multi-view fusion in capturing multi-dimensional molecular information and enhancing model generalization. A case study examining the anticancer compound Oroxylin A demonstrates MolP-PC's effective generalization in predicting key pharmacokinetic parameters such as half-life (T0.5) and clearance (CL), indicating its practical utility in drug modeling. However, the model exhibits a tendency to underestimate volume of distribution (VD), indicating potential for improvement in analyzing compounds with high tissue distribution. This study presents an efficient and interpretable approach for ADMET property prediction, establishing a novel framework for molecular optimization and risk assessment in drug development.
Deep Learning

OBJECTIVE: To assess the short-term lag effects of climate and air pollution on hospital admissions for cardiovascular and respiratory diseases, and to develop deep learning-based models for daily hospital admission prediction. METHODS: A multi-city study was conducted in Tokyo's 23 wards, Osaka City, and Nagoya City. Random forest models were employed to assess the synergistic short-term lag effects (lag0, lag3, and lag7) of climate and air pollutants on hospitalization for five cardiovascular diseases (CVDs) and two respiratory diseases (RDs). Furthermore, we developed hybrid deep learning models that integrated an autoencoder (AE) with a Long Short-Term Memory network (AE+LSTM) to predict daily hospital admissions. RESULTS: On the day of exposure (lag0), air pollutants, particularly nitrogen oxides (NO x), exhibited the strongest influence on hospital admissions for CVD and RD, with pronounced effects observed for hypertension (I10-I15), ischemic heart disease (I20), arterial and capillary diseases (I70-I79), and lower respiratory infections (J20-J22 and J40-J47). At longer lags (lag3 and lag7), temperature and precipitation were more influential predictors. The AE+LSTM model outperformed the standard LSTM, improving the prediction accuracy by 32.4% for RD in Osaka and 20.94% for CVD in Nagoya. CONCLUSION Our findings reveal the dynamic, time-varying health risks associated with environmental exposure and demonstrate the utility of deep learnings in predicting short-term hospital admissions. This framework can inform early warning systems, enhance healthcare resource allocation, and support climate-adaptive public health strategies.
Humans ; Hospitalization/statistics & numerical data* ; Cardiovascular Diseases/epidemiology* ; Japan/epidemiology* ; Air Pollutants/analysis* ; Air Pollution/adverse effects* ; Cities/epidemiology* ; Climate ; Respiratory Tract Diseases/epidemiology* ; Deep Learning ; Male

Artificial intelligence (AI) is rapidly advancing in periodontology, bringing new opportunities to clinical diagnosis, risk assessment, personalized treatment planning, and remote patient care. Leveraging core technologies such as deep learning, machine learning, and natural language processing, AI significantly enhances the sensitivity of early periodontal disease detection and provides precise quantification of alveolar bone loss and soft tissue damage. AI facilitates multimodal data integration by synthesizing medical history, lifestyle factors, and imaging data, thereby offering enhanced accurate risk prediction and personalized therapeutic recommendations. By integrating remote monitoring with tailored health counseling, AI helps patients maintain adherence to self-care protocols, significantly improving their oral health-related quality of life and treatment satisfaction. Moreover, AI demonstrates considerable potential in periodontal research and education, particularly in large-scale data mining, virtual clinical case simulations, and natural language processing-assisted literature management. Nevertheless, challenges remain concerning model generalizability, data quality, ethical concerns, and interpretability. The advancement of multi-center big-data platforms is expected to foster a profound integration of AI and periodontology, propelling precision medicine and digital healthcare, enabling holistic management from prevention to long-term care, and enhancing diagnostic efficiency and patient health outcomes.
Humans ; Artificial Intelligence ; Periodontics/methods* ; Periodontal Diseases/therapy* ; Deep Learning ; Precision Medicine ; Quality of Life

Predicting protein mutation effects is a key challenge in bioinformatics and protein engineering. Recent advancements in deep learning, particularly the development of protein language models (PLMs), have brought new opportunities to this field. This review summarizes the application of PLMs in predicting protein mutation effects, focusing on three main types of models: sequence-based models, structure-based models, and models that combine sequence and structural information. We analyze in detail the principles, advantages, and limitations of these models and discuss the application of unsupervised and supervised learning in model training. Furthermore, this paper discusses the main challenges currently faced, including the acquisition of high-quality datasets and the handling of data noise. Finally, we look ahead to future research directions, including the application prospects of emerging technologies such as multimodal fusion and few-shot learning. This review aims to provide researchers with a comprehensive perspective to further advance the prediction of protein mutation effects.
Mutation ; Proteins/chemistry* ; Computational Biology/methods* ; Deep Learning ; Protein Engineering

Natural components serve the survival instincts of cells that are obtained through long-term evolution, while they often fail to meet the demands of engineered cells for efficiently performing biological functions in special industrial environments. Enzymes, as biological catalysts, play a key role in biosynthetic pathways, significantly enhancing the rate and selectivity of biochemical reactions. However, the catalytic efficiency, stability, substrate specificity, and tolerance of natural enzymes often fall short of industrial production requirements. Therefore, exploring and modifying enzymes to suit specific biomanufacturing processes has become crucial. In recent years, artificial intelligence (AI) has played an increasingly important role in the discovery, evaluation, engineering, and de novo design of proteins. AI can accelerate the discovery and optimization of proteins by analyzing large amounts of bioinformatics data and predicting protein functions and characteristics by machine learning and deep learning algorithms. Moreover, AI can assist researchers in designing new protein structures by simulating and predicting their performance under different conditions, providing guidance for protein design. This paper reviews the latest research advances in protein discovery, evaluation, engineering, and de novo design for biomanufacturing and explores the hot topics, challenges, and emerging technical methods in this field, aiming to provide guidance and inspiration for researchers in related fields.
Protein Engineering/methods* ; Artificial Intelligence ; Proteins/genetics* ; Computational Biology ; Machine Learning ; Data Mining ; Algorithms ; Deep Learning

Synthetic promoters are novel promoters artificially designed and do not exist in nature. They can initiate the expression of target genes with specific regulatory modes, offering advantages such as high expression efficiency, precise regulation, and modularity. These features endow synthetic promoters with significant application potential in fields such as industrial production, environmental monitoring, and disease diagnosis and treatment. This paper reviews the basic structures, functions, and classification of promoters, discusses various regulatory elements that influence promoter functions, including enhancers, signal response elements, and transcription factors. Additionally, the conventional and deep learning-based strategies for designing synthetic promoters are summarized. Finally, the theoretical significance of synthetic promoters is emphasized, which is followed by an overview of their current applications, along with a rational discussion on the challenges and future development directions of synthetic promoters. Given the critical role of promoters in gene regulation, this article provides a review and outlook on the research progress of synthetic promoters, which holds reference value for the design of cellular gene circuits.
Promoter Regions, Genetic/genetics* ; Transcription Factors/genetics* ; Synthetic Biology/methods* ; Gene Expression Regulation ; Humans ; Deep Learning

Testicular histology based on testicular biopsy is an important factor for determining appropriate testicular sperm extraction surgery and predicting sperm retrieval outcomes in patients with azoospermia. Therefore, we developed a deep learning (DL) model to establish the associations between testicular grayscale ultrasound images and testicular histology. We retrospectively included two-dimensional testicular grayscale ultrasound from patients with azoospermia (353 men with 4357 images between July 2017 and December 2021 in The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China) to develop a DL model. We obtained testicular histology during conventional testicular sperm extraction. Our DL model was trained based on ultrasound images or fusion data (ultrasound images fused with the corresponding testicular volume) to distinguish spermatozoa presence in pathology (SPP) and spermatozoa absence in pathology (SAP) and to classify maturation arrest (MA) and Sertoli cell-only syndrome (SCOS) in patients with SAP. Areas under the receiver operating characteristic curve (AUCs), accuracy, sensitivity, and specificity were used to analyze model performance. DL based on images achieved an AUC of 0.922 (95% confidence interval [CI]: 0.908-0.935), a sensitivity of 80.9%, a specificity of 84.6%, and an accuracy of 83.5% in predicting SPP (including normal spermatogenesis and hypospermatogenesis) and SAP (including MA and SCOS). In the identification of SCOS and MA, DL on fusion data yielded better diagnostic performance with an AUC of 0.979 (95% CI: 0.969-0.989), a sensitivity of 89.7%, a specificity of 97.1%, and an accuracy of 92.1%. Our study provides a noninvasive method to predict testicular histology for patients with azoospermia, which would avoid unnecessary testicular biopsy.
Humans ; Male ; Azoospermia/diagnostic imaging* ; Deep Learning ; Testis/pathology* ; Retrospective Studies ; Adult ; Ultrasonography/methods* ; Sperm Retrieval ; Sertoli Cell-Only Syndrome/diagnostic imaging*

OBJECTIVE: To develop an original-mirror alignment associated deep learning algorithm for intelligent registration of three-dimensional maxillofacial point cloud data, by utilizing a dynamic graph-based registration network model (maxillofacial dynamic graph registration network, MDGR-Net), and to provide a valuable reference for digital design and analysis in clinical dental applications. METHODS: Four hundred clinical patients without significant deformities were recruited from Peking University School of Stomatology from October 2018 to October 2022. Through data augmentation, a total of 2 000 three-dimensional maxillofacial datasets were generated for training and testing the MDGR-Net algorithm. These were divided into a training set (1 400 cases), a validation set (200 cases), and an internal test set (200 cases). The MDGR-Net model constructed feature vectors for key points in both original and mirror point clouds (X, Y), established correspondences between key points in the X and Y point clouds based on these feature vectors, and calculated rotation and translation matrices using singular value decomposition (SVD). Utilizing the MDGR-Net model, intelligent registration of the original and mirror point clouds were achieved, resulting in a combined point cloud. The principal component analysis (PCA) algorithm was applied to this combined point cloud to obtain the symmetry reference plane associated with the MDGR-Net methodology. Model evaluation for the translation and rotation matrices on the test set was performed using the coefficient of determination (R²). Angle error evaluations for the three-dimensional maxillofacial symmetry reference planes were constructed using the MDGR-Net-associated method and the "ground truth" iterative closest point (ICP)-associated method were conducted on 200 cases in the internal test set and 40 cases in an external test set. RESULTS: Based on testing with the three-dimensional maxillofacial data from the 200-case internal test set, the MDGR-Net model achieved an R² value of 0.91 for the rotation matrix and 0.98 for the translation matrix. The average angle error on the internal and external test sets were 0.84°±0.55° and 0.58°±0.43°, respectively. The construction of the three-dimensional maxillofacial symmetry reference plane for 40 clinical cases took only 3 seconds, with the model performing optimally in the patients with skeletal Class Ⅲ malocclusion, high angle cases, and Angle Class Ⅲ orthodontic patients. CONCLUSION This study proposed the MDGR-Net association method based on intelligent point cloud registration as a novel solution for constructing three-dimensional maxillofacial symmetry reference planes in clinical dental applications, which can significantly enhance diagnostic and therapeutic efficiency and outcomes, while reduce expert dependence.
Humans ; Deep Learning ; Algorithms ; Imaging, Three-Dimensional/methods* ; Male ; Female ; Maxilla/diagnostic imaging* ; Adult

With the development of artificial intelligence technology and the growing demand for minimally invasive surgery, the intelligentization of minimally invasive surgery has become a current research hotspot. Surgical instrument segmentation is a highly promising technology that can enhance the performance of minimally invasive endoscopic imaging systems, surgical video analysis systems, and other related systems. This article summarizes the semantic and instance segmentation methods of minimally invasive surgical instruments based on deep learning, deeply analyzes the supervision methods of training algorithms, network structure improvements, and attention mechanisms, and then discusses the methods based on the Segment Anything Model. Given that deep learning methods have extremely high requirements for data, current data augmentation methods have also been explored. Finally, a summary and outlook on instrument segmentation technology are provided.
Artificial Intelligence ; Minimally Invasive Surgical Procedures/instrumentation* ; Algorithms ; Deep Learning ; Humans ; Image Processing, Computer-Assisted

The clinical application of binocular endoscope relies primarily on the visual system of physicians to create a three-dimensional effect, but it cannot provide accurate depth information. The utilization of 3D reconstruction technology in binocular endoscopy can facilitate the recovery of image depth information, and the application of deep learning-based 3D reconstruction technology can significantly improve the accuracy and real-time performance of reconstruction results, making it widely applicable in the realm of minimally invasive surgery. This paper aims to explore the key technologies and implementation methods of deep learning based 3D reconstruction for binocular endoscopic images, and seeks to outline strategies for enhancing the quality of 3D reconstruction in endoscopic images, providing guidance for sustainable development of binocular endoscopic image reconstruction technology in clinical settings. This will assist in the application of minimally invasive surgery and contribute to meeting the demands of precision medicine.
Deep Learning ; Imaging, Three-Dimensional/methods* ; Humans ; Endoscopy/methods* ; Image Processing, Computer-Assisted/methods* ; Minimally Invasive Surgical Procedures