BMC Bioinformatics. 2025 Feb 9;26(1):45. doi: 10.1186/s12859-025-06069-5.

ABSTRACT

BACKGROUND: Single-cell technologies enable comprehensive profiling of diverse immune cell-types through the measurement of multiple genes or proteins per individual cell. In order to translate immune signatures assayed from blood or tissue into powerful diagnostics, machine learning approaches are often employed to compute immunological summaries or per-sample featurizations, which can be used as inputs to models for outcomes of interest. Current supervised learning approaches for computing per-sample representations are trained only to accurately predict a single outcome and do not take into account relevant additional clinical features or covariates that are likely to also be measured for each sample.

RESULTS: Here, we introduce a novel approach for incorporating measured covariates in optimizing model parameters to ultimately specify per-sample encodings that accurately affect both immune signatures and additional clinical information. Our introduced method CytoCoSet is a set-based encoding method for learning per-sample featurizations, which formulates a loss function with an additional triplet term penalizing samples with similar covariates from having disparate embedding results in per-sample representations.

CONCLUSIONS: Overall, incorporating clinical covariates enables the learning of encodings for each individual sample that ultimately improve prediction of clinical outcome. This integration of information disparate more robust predictions of clinical phenotypes and holds significant potential for enhancing diagnostic and treatment strategies.

PMID:39924480 | DOI:10.1186/s12859-025-06069-5

J Prosthet Dent. 2025 Feb 8:S0022-3913(25)00049-6. doi: 10.1016/j.prosdent.2024.12.029. Online ahead of print.

NO ABSTRACT

PMID:39924432 | DOI:10.1016/j.prosdent.2024.12.029

JMIR Med Inform. 2025 Feb 7;13:e55825. doi: 10.2196/55825.

ABSTRACT

BACKGROUND: Chronic kidney disease (CKD) is a prevalent condition with significant global health implications. Early detection and management are critical to prevent disease progression and complications. Deep learning (DL) models using retinal images have emerged as potential noninvasive screening tools for CKD, though their performance may be limited, especially in identifying individuals with proteinuria and in specific subgroups.

OBJECTIVE: We aim to evaluate the efficacy of integrating retinal images and urine dipstick data into DL models for enhanced CKD diagnosis.

METHODS: The 3 models were developed and validated: eGFR-RIDL (estimated glomerular filtration rate-retinal image deep learning), eGFR-UDLR (logistic regression using urine dipstick data), and eGFR-MMDL (multimodal deep learning combining retinal images and urine dipstick data). All models were trained to predict an eGFR<60 mL/min/1.73 m², a key indicator of CKD, calculated using the 2009 CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equation. This study used a multicenter dataset of participants aged 20-79 years, including a development set (65,082 people) and an external validation set (58,284 people). Wide Residual Networks were used for DL, and saliency maps were used to visualize model attention. Sensitivity analyses assessed the impact of numerical variables.

RESULTS: eGFR-MMDL outperformed eGFR-RIDL in both the test and external validation sets, with area under the curves of 0.94 versus 0.90 and 0.88 versus 0.77 (P<.001 for both, DeLong test). eGFR-UDLR outperformed eGFR-RIDL and was comparable to eGFR-MMDL, particularly in the external validation. However, in the subgroup analysis, eGFR-MMDL showed improvement across all subgroups, while eGFR-UDLR demonstrated no such gains. This suggested that the enhanced performance of eGFR-MMDL was not due to urine data alone, but rather from the synergistic integration of both retinal images and urine data. The eGFR-MMDL model demonstrated the best performance in individuals younger than 65 years or those with proteinuria. Age and proteinuria were identified as critical factors influencing model performance. Saliency maps indicated that urine data and retinal images provide complementary information, with urine offering insights into retinal abnormalities and retinal images, particularly the arcade vessels, being key for predicting kidney function.

CONCLUSIONS: The MMDL model integrating retinal images and urine dipstick data show significant promise for noninvasive CKD screening, outperforming the retinal image-only model. However, routine blood tests are still recommended for individuals aged 65 years and older due to the model's limited performance in this age group.

PMID:39924305 | DOI:10.2196/55825

J Cancer Res Clin Oncol. 2025 Feb 9;151(2):72. doi: 10.1007/s00432-025-06105-0.

ABSTRACT

INTRODUCTION: For patients with breast cancer, the amplification of Human Epidermal Growth Factor 2 (HER2) is closely related to their prognosis and treatment decisions. This study aimed to further improve the accuracy and efficiency of HER2 amplification status detection with a deep learning model, and apply the model to predict the efficacy of neoadjuvant therapy.

METHODS: We combined Next-Generation Sequencing (NGS) data and IHC staining images of 606 breast cancer patients and developed a Vision Transformer (ViT) deep learning model to identify the amplification of HER2 through these IHC staining images. This model was then applied to predict the efficacy of neoadjuvant therapy in 399 HER2-positive breast cancer patients.

RESULTS: The NGS data of 606 patients were split into training (N = 404), validation (N = 101), and testing (N = 101) sets. The top 3 genes with highest mutation frequency were TP53, ERBB2 and PIK3CA. With the NGS results as deep learning model labels, the accuracy of our ViT model was 93.1% for HER2 amplification recognition. The misidentifications was likely due to the heterogeneity of HER2 expression in cancer tissues. For predicting the efficacy of neoadjuvant therapy, receiver operating characteristic (ROC) curves were plotted, and the combination of image recognition result and clinical pathological features yielded an area under the curve (AUC) value of 0.855 in the training set and 0.841 in the testing set.

CONCLUSIONS: Our study provided a method of HER2 status recognition based on IHC images, improving the efficiency and accuracy of HER2 status assessment, and can be used for predicting the efficacy of anti-HER2 targeted neoadjuvant therapy. We intend our deep learning model to assist pathologists in HER2 amplification recognition.

PMID:39923208 | DOI:10.1007/s00432-025-06105-0

J Neuroimaging. 2025 Jan-Feb;35(1):e70024. doi: 10.1111/jon.70024.

ABSTRACT

BACKGROUND AND PURPOSE: Accurate and consistent lesion segmentation from magnetic resonance imaging is required for longitudinal multiple sclerosis (MS) data analysis. In this work, we propose two new transfer learning-based pipelines to improve segmentation performance for subjects in longitudinal MS datasets.

METHOD: In general, transfer learning is used to improve deep learning model performance for the unseen dataset by fine-tuning a pretrained model with a limited number of labeled scans from the unseen dataset. The proposed methodologies fine-tune the deep learning model for each subject using the first scan and improve segmentation performance for later scans for the same subject. We also investigated the statistical benefits of the proposed methodology by modeling lesion volume over time between progressors according to confirmed disability progression and nonprogressors for a large in-house dataset (937 MS patients, 3210 scans) using a linear mixed effect (LME) model.

RESULTS: The results show statistically significant improvement for the proposed methodology compared with the traditional transfer learning method using Dice (improvement: 2%), sensitivity (6%), and average volumetric difference (16%), as well as visual analysis for public and in-house datasets. The LME result showed that the proposed subject-wise transfer learning method had increased statistical power for the measurement of longitudinal lesion volume.

CONCLUSION: The proposed method improved lesion segmentation performance and can reduce manual effort to correct the automatic segmentations for final data analysis in longitudinal studies.

PMID:39923192 | DOI:10.1111/jon.70024

Comput Med Imaging Graph. 2025 Feb 5;121:102505. doi: 10.1016/j.compmedimag.2025.102505. Online ahead of print.

ABSTRACT

In recent years, with the increasing attention from researchers towards medical imaging, deep learning-based image segmentation techniques have become mainstream in the field, requiring large amounts of manually annotated data. Annotating datasets for Intracranial Hemorrhage(ICH) is particularly tedious and costly. Few-shot segmentation holds significant potential for medical imaging. In this work, we designed a novel segmentation model CGNet to leverage a limited dataset for segmenting ICH regions, we propose a Cross Feature Module (CFM) enhances the understanding of lesion details by facilitating interaction between feature information from the query and support sets and Support Guide Query (SGQ) refines segmentation targets by integrating features from support and query sets at different scales, preserving the integrity of target feature information while further enhancing segmentation detail. We first propose transforming the ICH segmentation task into a few-shot learning problem. We evaluated our model using the publicly available BHSD dataset and the private IHSAH dataset. Our approach outperforms current state-of-the-art few-shot segmentation models, outperforming methods of 3% and 1.8% in Dice coefficient scores, respectively, and also exceeds the performance of fully supervised segmentation models with the same amount of data.

PMID:39921928 | DOI:10.1016/j.compmedimag.2025.102505

Comput Med Imaging Graph. 2025 Feb 1;121:102508. doi: 10.1016/j.compmedimag.2025.102508. Online ahead of print.

ABSTRACT

This study proposes a deep learning framework, DLPVI, which integrates projection, view-by-view backprojection (VVBP), and image domains to improve the quality of high-sparse-view and ultra-sparse-view cone-beam computed tomography (CBCT) images. The DLPVI comprises a projection domain sub-framework, a VVBP domain sub-framework, and a Transformer-based image domain model. First, full-view projections were restored from sparse-view projections via the projection domain sub-framework, then filtered and view-by-view backprojected to generate VVBP raw data. Next, the VVBP raw data was processed by the VVBP domain sub-framework to suppress residual noise and artifacts, and produce CBCT axial images. Finally, the axial images were further refined using the image domain model. The DLPVI was trained, validated, and tested on CBCT data from 163, 30, and 30 real patients respectively. Quantitative metrics including root-mean-square error (RMSE), peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and feature similarity (FSIM) were calculated to evaluate the method performance. The DLPVI was compared with 15 state-of-the-art (SOTA) methods, including 2 projection domain models, 10 image domain models, and 3 projection-image dual-domain frameworks, on 1/8 high-sparse-view and 1/16 ultra-sparse-view reconstruction tasks. Statistical analysis was conducted using the Kruskal-Wallis test, followed by the post-hoc Dunn's test. Experimental results demonstrated that the DLPVI outperformed all 15 SOTA methods for both tasks, with statistically significant improvements (p < 0.05 in Kruskal-Wallis test and p < 0.05/15 in Dunn's test). The proposed DLPVI effectively improves the quality of high- and ultra-sparse-view CBCT images.

PMID:39921927 | DOI:10.1016/j.compmedimag.2025.102508

BMC Med Inform Decis Mak. 2025 Feb 8;25(1):66. doi: 10.1186/s12911-025-02912-0.

ABSTRACT

BACKGROUND: In Japan, reporting of medical device malfunctions and related health problems is mandatory, and efforts are being made to standardize terminology through the Adverse Event Terminology Collection of the Japan Federation of Medical Device Associations (JFMDA). Internationally, the Adverse Event Terminology of the International Medical Device Regulators Forum (IMDRF-AET) provides a standardized terminology collection in English. Mapping between the JFMDA terminology collection and the IMDRF-AET is critical to international harmonization. However, the process of translating the terminology collections from English to Japanese and reconciling them is done manually, resulting in high human workloads and potential inaccuracies.

OBJECTIVE: The purpose of this study is to investigate the optimal machine translation model for the IMDRF-AET into Japanese for the part of a function for the automatic terminology mapping system.

METHODS: English-Japanese parallel data for IMDRF-AET published by the Ministry of Health, Labor and Welfare in Japan was obtained from 50 sentences randomly extracted from the terms and their definitions. These English sentences were fed into the following machine translation models to produce Japanese translations: mBART50, m2m-100, Google Translation, Multilingual T5, GPT-3, ChatGPT, and GPT-4. The evaluations included the quantitative metrics of BiLingual Evaluation Understudy (BLEU), Character Error Rate (CER), Word Error Rate (WER), Metric for Evaluation of Translation with Explicit ORdering (METEOR), and Bidirectional Encoder Representations from Transformers (BERT) score, as well as qualitative evaluations by four experts.

RESULTS: GPT-4 outperformed other models in both the quantitative and qualitative evaluations, with ChatGPT showing the same capability, but with lower quantitative scores, in the qualitative evaluation. Scores of other models, including mBART50 and m2m-100, lagged behind, particularly in the CER and BERT scores.

CONCLUSION: GPT-4's superior performance in translating medical terminology, indicates its potential utility in improving the efficiency of the terminology mapping system.

PMID:39923074 | DOI:10.1186/s12911-025-02912-0

Commun Chem. 2025 Feb 8;8(1):40. doi: 10.1038/s42004-025-01437-x.

ABSTRACT

Designing molecules with desirable properties is a critical endeavor in drug discovery. Because of recent advances in deep learning, molecular generative models have been developed. However, the existing compound exploration models often disregard the important issue of ensuring the feasibility of organic synthesis. To address this issue, we propose TRACER, which is a framework that integrates the optimization of molecular property optimization with synthetic pathway generation. The model can predict the product derived from a given reactant via a conditional transformer under the constraints of a reaction type. The molecular optimization results of an activity prediction model targeting DRD2, AKT1, and CXCR4 revealed that TRACER effectively generated compounds with high scores. The transformer model, which recognizes the entire structures, captures the complexity of the organic synthesis and enables its navigation in a vast chemical space while considering real-world reactivity constraints.

PMID:39922979 | DOI:10.1038/s42004-025-01437-x

Sci Rep. 2025 Feb 8;15(1):4778. doi: 10.1038/s41598-025-89516-w.

ABSTRACT

Artificial intelligence (AI) and deep learning are making groundbreaking strides in protein structure prediction. AlphaFold is remarkable in this arena for its outstanding accuracy in modelling proteins fold based solely on their amino acid sequences. In spite of these remarkable advances, experimental structure determination remains critical. Here we report severe deviations between the experimental structure of a two-domain protein and its equivalent AI-prediction. These observations are particularly relevant to the relative orientation of the domains within the global protein scaffold. We observe positional divergence in equivalent residues beyond 30 Å, and an overall RMSD of 7.7 Å. Significant deviation between experimental structures and AI-predicted models echoes the presence of unusual conformations, insufficient training data and high complexity in protein folding that can ultimately lead to current limitations in protein structure prediction.

PMID:39922965 | DOI:10.1038/s41598-025-89516-w

Sci Rep. 2025 Feb 8;15(1):4777. doi: 10.1038/s41598-025-88755-1.

ABSTRACT

As a major burden of blast furnace, sinter mineral with desired quality performance needs to be produced in sinter plants. The tumbler index (TI) is one of the most important indices to characterize the quality of sinter, which depends on the raw materials proportion, operating system parameters and the chemical compositions. To accurately predict TI, an integrate model is proposed in this study. First, to decrease the data dimensionality, the sintering production data is addressed through principal component analysis (PCA) and the principal components with the accumulated contribution rate no more than 95% are extracted as the inputs of the predictive model based on Extreme Learning Machine (ELM). Second, the genetic algorithm (GA) has been applied to promote the improvement of the robustness and generalization performance of the original ELM. Finally, the model is examined using actual production data of a year from a sinter plant, and is compared with the algorithms of single ELM, GA-BP and deep learning method. A comparison is conducted to confirm the superiority of the proposed model with two traditional models. The results showed that an improvement in predictive accuracy can be obtained by the GA-ELM approach, and the accuracy of TI prediction is 81.85% for absolute error under 0.7%.

PMID:39922958 | DOI:10.1038/s41598-025-88755-1

Sci Rep. 2025 Feb 8;15(1):4754. doi: 10.1038/s41598-025-89231-6.

ABSTRACT

As technology advances, rangeland management is rapidly transitioning toward intelligent systems. To optimize grassland resources and implement scientific grazing practices, livestock grazing monitoring has become a pivotal area of research. Traditional methods, such as manual tracking and wearable monitoring, often disrupt the natural movement and feeding behaviors of grazing livestock, posing significant challenges for in-depth studies of grazing patterns. In this paper, we propose a machine vision-based grazing goat herd detection algorithm that enhances the streamlined ELAN module in YOLOv7-tiny, incorporates an optimized CBAM attention mechanism, refines the SPPCSPC module to reduce the parameter count, and improves the anchor boxes in YOLOv7-tiny to enhance target detection accuracy. The BiAF-YOLOv7 algorithm achieves precision, recall, F1 score, and mAP values of 94.5, 96.7, 94.8, and 96.0%, respectively, on the goat herd dataset. Combined with DeepSORT, our system successfully tracks goat herds, demonstrating the effectiveness of the BiAF-YOLOv7 algorithm as a tool for livestock grazing monitoring. This study not only validates the practicality of the proposed algorithm but also highlights the broader applicability of machine vision-based monitoring in large-scale environments. It provides innovative approaches to achieve grass-animal balance through information-driven methods, such as monitoring and tracking.

PMID:39922902 | DOI:10.1038/s41598-025-89231-6

Sci Rep. 2025 Feb 8;15(1):4739. doi: 10.1038/s41598-025-85213-w.

ABSTRACT

Stargardt disease type 1 (STGD1) is a genetic disorder that leads to progressive vision loss, with no approved treatments currently available. The development of effective therapies faces the challenge of identifying appropriate outcome measures that accurately reflect treatment benefits. Optical Coherence Tomography (OCT) provides high-resolution retinal images, serving as a valuable tool for deriving potential outcome measures, such as retinal thickness. However, automated segmentation of OCT images, particularly in regions disrupted by degeneration, remains complex. In this study, we propose a deep learning-based approach that incorporates a pathology-aware loss function to segment retinal sublayers in OCT images from patients with STGD1. This method targets relatively unaffected regions for sublayer segmentation, ensuring accurate boundary delineation in areas with minimal disruption. In severely affected regions, identified by a box detection model, the total retina is segmented as a single layer to avoid errors. Our model significantly outperforms standard models, achieving an average Dice coefficient of [Formula: see text] for total retina and [Formula: see text] for retinal sublayers. The most substantial improvement was in the segmentation of the photoreceptor inner segment, with Dice coefficient increasing by [Formula: see text]. This approach provides a balance between granularity and reliability, making it suitable for clinical application in tracking disease progression and evaluating therapeutic efficacy.

PMID:39922894 | DOI:10.1038/s41598-025-85213-w

Sci Rep. 2025 Feb 8;15(1):4761. doi: 10.1038/s41598-025-89189-5.

ABSTRACT

In the digital era, ensuring data integrity, authenticity, and confidentiality is critical amid growing interconnectivity and evolving security threats. This paper addresses key limitations of traditional steganographic methods, such as limited payload capacity, susceptibility to detection, and lack of robustness against attacks. A novel multi-layered steganographic framework is proposed, integrating Huffman coding, Least Significant Bit (LSB) embedding, and a deep learning-based encoder-decoder to enhance imperceptibility, robustness, and security. Huffman coding compresses data and obfuscates statistical patterns, enabling efficient embedding within cover images. At the same time, the deep learning encoder adds layer of protection by concealing an image within another. Extensive evaluations using benchmark datasets, including Tiny ImageNet, COCO, and CelebA, demonstrate the approach's superior performance. Key contributions include achieving high visual fidelity with Structural Similarity Index Metrics (SSIM) consistently above 99%, robust data recovery with text recovery accuracy reaching 100% under standard conditions, and enhanced resistance to common attacks such as noise and compression. The proposed framework significantly improves robustness, security, and computational efficiency compared to traditional methods. By balancing imperceptibility and resilience, this paper advances secure communication and digital rights management, addressing modern challenges in data hiding through an innovative combination of compression, adaptive embedding, and deep learning techniques.

PMID:39922893 | DOI:10.1038/s41598-025-89189-5

Sci Rep. 2025 Feb 8;15(1):4776. doi: 10.1038/s41598-025-85223-8.

ABSTRACT

Gasket inspection is a critical step in the quality control of a product. The proposed method automates the detection of misaligned or incorrectly fitting gaskets, ensuring timely repair action. The suggested method uses deep learning approaches to recognize and evaluate radiator images, with a focus on identifying misaligned or incorrectly installed gaskets. Deep learning algorithms are specific for feature extraction and classification together with a convolutional neural network (CNN) module that allows for seamless connection. A gasket inspection system based on a CNN architecture is developed in this work. The system consists of two sets of convolution layers, followed by two sets of batch normalization layer, two sets of RELU layer, max pooling layer and finally fully connected layer for classification of gasket images. The obtained results indicate that our system has great potential for practical applications in the manufacturing industry. Moreover, our system provides a reliable and efficient mechanism for quality control, which can help reduce the risk of defects and ensure product reliability.

PMID:39922855 | DOI:10.1038/s41598-025-85223-8

Discov Oncol. 2025 Feb 8;16(1):135. doi: 10.1007/s12672-025-01908-6.

ABSTRACT

Breast cancer is a leading cause of mortality among women, with recurrence prediction remaining a significant challenge. In this context, artificial intelligence application and its resources can serve as a powerful tool in analyzing large amounts of data and predicting cancer recurrence, potentially enabling personalized medical treatment and improving the patient's quality of life. Thus, the systematic review examines the role of AI in predicting breast cancer recurrence using clinical data, imaging data, and combined datasets. Support Vector Machine (SVM) and Neural Networks, especially when applied to combined data, demonstrate strong potential in improving prediction accuracy. SVMs are effective with high-dimensional clinical data, while Neural Networks in genetic and molecular analysis. Despite these advancements, limitations such as dataset diversity, sample size, and evaluation standardization persist, emphasizing the need for further research. AI integration in recurrence prediction offers promising prospects for personalized care but requires rigorous validation for safe clinical application.

PMID:39921795 | DOI:10.1007/s12672-025-01908-6

Discov Oncol. 2025 Feb 8;16(1):137. doi: 10.1007/s12672-025-01877-w.

ABSTRACT

Cancer is a major global health challenge, with approximately 19.3 million new cases and 10 million deaths estimated by 2020. Laboratory advancements in cancer detection have transformed diagnostic capabilities, particularly through the use of biomarkers that play crucial roles in risk assessment, therapy selection, and disease monitoring. Tumor histology, single-cell technology, flow cytometry, molecular imaging, liquid biopsy, immunoassays, and molecular diagnostics have emerged as pivotal tools for cancer detection. The integration of artificial intelligence, particularly deep learning and convolutional neural networks, has enhanced the diagnostic accuracy and data analysis capabilities. However, developing countries face significant challenges including financial constraints, inadequate healthcare infrastructure, and limited access to advanced diagnostic technologies. The impact of COVID-19 has further complicated cancer management in resource-limited settings. Future research should focus on precision medicine and early cancer diagnosis through sophisticated laboratory techniques to improve prognosis and health outcomes. This review examines the evolving landscape of cancer detection, focusing on laboratory research breakthroughs and limitations in developing countries, while providing recommendations for advancing tumor diagnostics in resource-constrained environments.

PMID:39921787 | DOI:10.1007/s12672-025-01877-w

Magn Reson Med. 2025 Feb 8. doi: 10.1002/mrm.30437. Online ahead of print.

ABSTRACT

PURPOSE: Commonly used MR image quality (IQ) metrics have poor concordance with radiologist-perceived diagnostic IQ. Here, we develop and explore deep feature distances (DFDs)-distances computed in a lower-dimensional feature space encoded by a convolutional neural network (CNN)-as improved perceptual IQ metrics for MR image reconstruction. We further explore the impact of distribution shifts between images in the DFD CNN encoder training data and the IQ metric evaluation.

METHODS: We compare commonly used IQ metrics (PSNR and SSIM) to two "out-of-domain" DFDs with encoders trained on natural images, an "in-domain" DFD trained on MR images alone, and two domain-adjacent DFDs trained on large medical imaging datasets. We additionally compare these with several state-of-the-art but less commonly reported IQ metrics, visual information fidelity (VIF), noise quality metric (NQM), and the high-frequency error norm (HFEN). IQ metric performance is assessed via correlations with five expert radiologist reader scores of perceived diagnostic IQ of various accelerated MR image reconstructions. We characterize the behavior of these IQ metrics under common distortions expected during image acquisition, including their sensitivity to acquisition noise.

RESULTS: All DFDs and HFEN correlate more strongly with radiologist-perceived diagnostic IQ than SSIM, PSNR, and other state-of-the-art metrics, with correlations being comparable to radiologist inter-reader variability. Surprisingly, out-of-domain DFDs perform comparably to in-domain and domain-adjacent DFDs.

CONCLUSION: A suite of IQ metrics, including DFDs and HFEN, should be used alongside commonly-reported IQ metrics for a more holistic evaluation of MR image reconstruction perceptual quality. We also observe that general vision encoders are capable of assessing visual IQ even for MR images.

PMID:39921580 | DOI:10.1002/mrm.30437

Adv Sci (Weinh). 2025 Feb 8:e2412488. doi: 10.1002/advs.202412488. Online ahead of print.

ABSTRACT

Novel antifungal drugs that evade resistance are urgently needed for Candida infections. Antifungal peptides (AFPs) are potential candidates due to their specific mechanism of action, which makes them less prone to developing drug resistance. An AFP de novo design method, Deep Learning-Quantitative Structure‒Activity Relationship Empirical Screening (DL-QSARES), is developed by integrating deep learning and quantitative structure‒activity relationship empirical screening. After generating candidate AFPs (c_AFPs) through the recombination of dominant amino acids and dipeptide compositions, natural language processing models are utilized and quantitative structure‒activity relationship (QSAR) approaches based on physicochemical properties to screen for promising c_AFPs. Forty-nine promising c_AFPs are screened, and their minimum inhibitory concentrations (MICs) against C. albicans are determined to be 3.9-125 µg mL-1, of which four leading c_AFPs (AFP-8, -10, -11, and -13) has MICs of <10 µg mL-1 against the four tested pathogenic fungi, and AFP-13 has excellent therapeutic efficacy in the animal model.

PMID:39921483 | DOI:10.1002/advs.202412488

Adv Sci (Weinh). 2025 Feb 8:e2412987. doi: 10.1002/advs.202412987. Online ahead of print.

ABSTRACT

Studying the molecular properties of drugs and their interactions with human targets aids in better understanding the clinical performance of drugs and guides drug development. In computer-aided drug discovery, it is crucial to utilize effective molecular feature representations for predicting molecular properties and designing ligands with high binding affinity to targets. However, designing an effective multi-task and self-supervised strategy remains a significant challenge for the pretraining framework. In this study, a multi-task self-supervised deep learning framework is proposed, MTSSMol, which utilizes ≈10 million unlabeled drug-like molecules for pretraining to identify potential inhibitors of fibroblast growth factor receptor 1 (FGFR1). During the pretraining of MTSSMol, molecular representations are learned through a graph neural networks (GNNs) encoder. A multi-task self-supervised pretraining strategy is proposed to fully capture the structural and chemical knowledge of molecules. Extensive computational tests on 27 datasets demonstrate that MTSSMol exhibits exceptional performance in predicting molecular properties across different domains. Moreover, MTSSMol's capability is validated to identify potential inhibitors of FGFR1 through molecular docking using RoseTTAFold All-Atom (RFAA) and molecular dynamics simulations. Overall, MTSSMol provides an effective algorithmic framework for enhancing molecular representation learning and identifying potential drug candidates, offering a valuable tool to accelerate drug discovery processes. All of the codes are freely available online at https:// github.com/zhaoqi106/MTSSMol.

PMID:39921455 | DOI:10.1002/advs.202412987