IEEE Trans Neural Netw Learn Syst. 2025 Mar 20;PP. doi: 10.1109/TNNLS.2025.3543926. Online ahead of print.

ABSTRACT

This article presents a unified framework for dynamics modeling and control design using deep learning, focusing on incorporating prior side information on stabilizability. Control theory provides systematic techniques for designing feedback systems while ensuring fundamental properties such as stabilizability, which are crucial for practical control applications. However, conventional data-driven approaches often overlook or struggle to explicitly incorporate such control properties into learned models. To address this, we introduce a novel neural network (NN)-based approach that concurrently learns the system dynamics, a stabilizing feedback controller, and a Lyapunov function for the closed-loop system, thus explicitly guaranteeing stabilizability in the learned model. Our proposed deep learning framework is versatile and applicable across a wide range of control problems, including safety control, -gain control, passivation, and solutions to Hamilton-Jacobi inequalities. By embedding stabilizability as a core property within the learning process, our method allows for the development of learned models that are not only data-driven but also grounded in control-theoretic guarantees, greatly enhancing their utility in real-world control applications. This article includes examples that demonstrate the effectiveness of this approach, showcasing the stability and control performance improvements achieved in various control scenarios. The methods proposed in this article can be easily applied to modeling without control design. The code has been open-sourced and is available at https://github.com/kashctrl/Deep_Stabilizable_Models.

PMID:40111782 | DOI:10.1109/TNNLS.2025.3543926

IEEE J Biomed Health Inform. 2025 Mar 20;PP. doi: 10.1109/JBHI.2025.3553217. Online ahead of print.

ABSTRACT

Deep learning offers efficient solutions for drug-target interaction prediction, but current methods often fail to capture the full complexity of multi-modal data (i.e. sequence, graphs, and three-dimensional structures), limiting both performance and generalization. Here, we present UnitedDTA, a novel explainable deep learning framework capable of integrating multi-modal biomolecule data to improve the binding affinity prediction, especially for novel (unseen) drugs and targets. UnitedDTA enables automatic learning unified discriminative representations from multi-modality data via contrastive learning and cross-attention mechanisms for cross-modality alignment and integration. Comparative results on multiple benchmark datasets show that UnitedDTA significantly outperforms the state-of-the-art drug-target affinity prediction methods and exhibits better generalization ability in predicting unseen drug-target pairs. More importantly, unlike most "black-box" deep learning methods, our well-established model offers better interpretability which enables us to directly infer the important substructures of the drug-target complexes that influence the binding activity, thus providing the insights in unveiling the binding preferences. Moreover, by extending UnitedDTA to other downstream tasks (e.g. molecular property prediction), we showcase the proposed multi-modal representation learning is capable of capturing the latent molecular representations that are closely associated with the molecular property, demonstrating the broad application potential for advancing the drug discovery process.

PMID:40111772 | DOI:10.1109/JBHI.2025.3553217

Adv Sci (Weinh). 2025 Mar 20:e2415007. doi: 10.1002/advs.202415007. Online ahead of print.

ABSTRACT

Idiopathic pulmonary fibrosis (IPF) is characterized by excessive fibroblast recruitment and persistent extracellular matrix deposition at sites of tissue injury, leading to severe morbidity and mortality. However, the precise mechanisms by which fibroblasts contribute to IPF pathogenesis remain poorly understood. The study reveals that Sema3E and its receptor Plexin D1 are significantly overexpressed in the lungs of IPF patients and bleomycin (BLM)-induced lung fibrotic mice. Elevated plasma levels of Sema3E in IPF patients are negatively correlated with lung function. Importantly, Sema3E in IPF lungs predominantly exists as the P61-Sema3E. The knockdown of Sema3E or Plexin D1 effectively inhibits fibroblast activation, proliferation, and migration. Mechanistically, Furin-mediated cleavage of P87-Sema3E into P61-Sema3E drives these pro-fibrotic activities, with P61-Sema3E-PlexinD1 axis promoting fibroblast activation, proliferation, and migration by affecting the phosphorylation of ErbB2, which subsequently activates the ErbB2 pathways. Additionally, Furin inhibition reduces fibroblast activity by decreasing P61-Sema3E production. In vivo, both whole-lung Sema3E knockdown and fibroblast-specific Sema3E knockout confer protection against BLM-induced lung fibrosis. These findings underscore the crucial role of the P61-Sema3E-Plexin D1 axis in IPF pathogenesis and suggest that targeting this pathway may hold promise for the development of novel therapeutic strategies for IPF treatment.

PMID:40112179 | DOI:10.1002/advs.202415007

Am J Physiol Cell Physiol. 2025 Mar 20. doi: 10.1152/ajpcell.00374.2024. Online ahead of print.

ABSTRACT

Idiopathic pulmonary fibrosis (IPF) is a progressive and degenerative interstitial lung disease characterized by complex etiology, unclear pathogenesis, and high mortality. Long non-coding RNAs (lncRNAs) have been identified as key regulators in modulating the initiation, maintenance, and progression of pulmonary fibrosis. However, the precise pathological mechanisms through which lncRNAs are involved in IPF remain limited and require further elucidation. A novel lncABCE1-5 was identified as significantly decreased by an ncRNA microarray analysis in our eight IPF lung samples compared with three donor tissues and validated by qRT-PCR analysis in clinical lung samples. To investigate the biological function of ABCE1-5, we performed loss-and gain-of-function experiments in vitro and in vivo. LncABCE1-5 silencing promoted A549 cell migration and A549 and BEAS-2B cell apoptosis, while enhancing the expression of proteins associated with extracellular matrix deposition, whereas overexpression of ABCE1-5 partially attenuated TGF-β-induced fibrogenesis. Forced ABCE1-5 expression by intratracheal injection of adeno-associated virus 6 (AAV6) revealing the anti-fibrotic effect of ABCE1-5 in BLM-treated mice. Mechanistically, RNA pull-down-mass spectrometry and RIP assay demonstrated that ABCE1-5 directly binds to keratin14 (krt14) sequences, potentially impeding its expression by perturbing mRNA stability. Furthermore, decreased ABCE1-5 levels can promote krt14 expression and enhance the phosphorylation of both mTOR and Akt; overexpression of ABCE1-5 in BLM mouse lung tissue significantly attenuated the elevated levels of p-mTOR and p-AKT. Knockdown of krt14 reversed the activation of mTOR signaling mediated by ABCE1-5 silencing. Collectively, the downregulation of ABCE1-5 mediated krt14 activation, thereby activating mTOR/AKT signaling, to facilitate pulmonary fibrosis progression in IPF.

PMID:40111939 | DOI:10.1152/ajpcell.00374.2024

Plant Physiol. 2025 Mar 20:kiaf109. doi: 10.1093/plphys/kiaf109. Online ahead of print.

ABSTRACT

Herbivores sharing host plants are often temporally and spatially separated, limiting direct interactions between them. Nevertheless, as observed in numerous aboveground study systems, they can reciprocally influence each other via systemically induced plant responses. In contrast, examples of such plant-mediated interactions between belowground herbivores are scarce; however, we postulated that they similarly occur, given the large diversity of root-interacting soil organisms. To test this hypothesis, we analyzed the performance of cabbage root fly (Delia radicum) larvae feeding on the main roots of field mustard (Brassica rapa) plants whose fine roots were infected by the root-knot nematode (Meloidogyne incognita). Simultaneously, we studied the effects of M. incognita on D. radicum-induced defense responses and the accumulation of primary metabolites in the main root. We observed that almost 1.5 times as many D. radicum adults emerged from nematode-infected plants, indicating a facilitation effect of M. incognita infection. Although we observed increases in the accumulation of proteins and two essential amino acids, the strongest effect of nematode infection was visible in the defense response to D. radicum. We observed a 1.5 times higher accumulation of the defense-related phytohormone JA-Ile in response to D. radicum on nematode-infected plants, coinciding with a 75% increase in indole glucosinolate concentrations. Contrastingly, concentrations of aliphatic glucosinolates, secondary metabolites negatively affecting D. radicum, were 10-25% lower in nematode-infected plants. We hypothesize that the attenuated aliphatic glucosinolate concentrations result from antagonistic interactions between biosynthetic pathways of both glucosinolate classes, which was reflected in the expression of key biosynthesis genes. Our results provide explicit evidence of plant-mediated interactions between belowground organisms, likely via systemically induced responses in roots.

PMID:40112263 | DOI:10.1093/plphys/kiaf109

Plant Physiol. 2025 Mar 20:kiaf105. doi: 10.1093/plphys/kiaf105. Online ahead of print.

ABSTRACT

Abiotic and biotic stresses pose serious threats to plant productivity. Elucidating the gene regulatory networks involved in plant stress responses is essential for developing future breeding programs and innovative agricultural products. Here, we introduce the AtSRGA (Arabidopsis thaliana Stress Responsive Gene Atlas), a user-friendly application facilitating the retrieval of stress-responsive genes in Arabidopsis (Arabidopsis thaliana). The application was developed using 1,131 microarray and 1,050 RNA sequencing datasets obtained from public databases. These datasets correspond to 11 stress-related conditions, namely abscisic acid, cold, drought, heat, high light, hypoxia, osmotic stress, oxidative stress, salt, wounding, and Pseudomonas syringae pv. tomato DC3000. Using a modified meta-analysis technique known as the vote-counting method, we computed integrated scores to evaluate stress responsiveness for each condition across multiple studies. AtSRGA visualizes gene behavior under 11 stress conditions and offers an interactive, user-friendly interface accessible to all researchers. It presents a comprehensive heatmap of stress-responsive genes, facilitating the comparative analysis of individual stress responses and groups of genes responding to multiple stresses. We validated the expression patterns of several high-scoring genes of unknown function under cold and heat stress using RT-qPCR, thus demonstrating that our application helps select targets to understand stress-responsive gene networks in Arabidopsis. AtSRGA will improve the screening of stress-responsive genes in Arabidopsis, thereby supporting the advancement of plant science toward a sustainable society.

PMID:40112239 | DOI:10.1093/plphys/kiaf105

J Proteome Res. 2025 Mar 20. doi: 10.1021/acs.jproteome.5c00092. Online ahead of print.

ABSTRACT

Proteomics data-dependent acquisition data sets collected with high-resolution mass-spectrometry (MS) can achieve very high-quality results, but nearly every analysis yields results that are thresholded at some accepted false discovery rate, meaning that a substantial number of results are incorrect. For study conclusions that rely on a small number of peptide-spectrum matches being correct, it is thus important to examine at least some crucial spectra to ensure that they are not one of the incorrect identifications. We present Quetzal, a peptide fragment ion spectrum annotation tool to assist researchers in annotating and examining such spectra to ensure that they correctly support study conclusions. We describe how Quetzal annotates spectra using the new Human Proteome Organization (HUPO) Proteomics Standards Initiative (PSI) mzPAF standard for fragment ion peak annotation, including the Python-based code, a web-service end point that provides annotation services, and a web-based application for annotating spectra and producing publication-quality figures. We illustrate its functionality with several annotated spectra of varying complexity. Quetzal provides easily accessible functionality that can assist in the effort to ensure and demonstrate that crucial spectra support study conclusions. Quetzal is publicly available at https://proteomecentral.proteomexchange.org/quetzal/.

PMID:40111914 | DOI:10.1021/acs.jproteome.5c00092

J Comp Physiol B. 2025 Mar 20. doi: 10.1007/s00360-025-01608-1. Online ahead of print.

ABSTRACT

Widespread species often display traits of generalists, yet local adaptations may limit their ability to cope with diverse environmental conditions. With climate change being a pressing issue, distinguishing between the general ecological and physiological capacities of a species and those of individual populations is vital for assessing the capability to adapt rapidly to changing habitats. Despite its importance, physiological variation across broad range distributions, particularly among free-ranging bats in natural environments, has rarely been assessed. Studies focusing on physiological variation among different populations across seasons are even more limited. We investigated physiological variation in the Malagasy Trident Bat Triaenops menamena across three different roost types in Madagascar during the wet and dry season, examining aspects such as energy regimes, body temperature, and roost microclimates. We focused on patterns of torpor in relation to roosting conditions. We hypothesized that torpor occurrence would be higher during the colder, more demanding dry season. We predicted that populations roosting in more variable microclimates would expend less energy than those in mores stable ones due to more frequent use of torpor and greater metabolic rate reductions. Our findings highlight complex thermoregulatory strategies, with varying torpor expression across seasons and roosts. We observed an overall higher energy expenditure during the wet season but also greater energy savings during torpor in that season, regardless of roost type. We found that reductions in metabolic rate were positively correlated with greater fluctuations in ambient conditions, demonstrating these bats' adaptability to dynamic environments. Notably, we observed diverse torpor patterns, indicating the species' ability to use prolonged torpor under extreme conditions. This individual-level variation is crucial for adaptation to changing environmental conditions. Moreover, the flexibility in body temperature during torpor suggests caution in relying solely on it as an indicator for torpor use. Our study emphasizes the necessity to investigate thermoregulatory responses across different populations in their respective habitats to fully understand a species' adaptive potential.

PMID:40111435 | DOI:10.1007/s00360-025-01608-1

Microbiol Spectr. 2025 Mar 20:e0298724. doi: 10.1128/spectrum.02987-24. Online ahead of print.

ABSTRACT

Candida albicans filamentation and biofilm formation are key virulence factors tied to tissue invasion and antifungal tolerance. Pilocarpine hydrochloride (PHCl), a muscarinic receptor agonist, inhibits biofilm maturation, although its mechanism remains unclear. We explored PHCl effects by analyzing sphingolipid and lipid composition and proteomics in treated C. albicans SC5314 biofilms. PHCl significantly decreased polar lipid and ergosterol levels in biofilms while inducing phytoceramide and glucosylceramide accumulation. PHCl also induced reactive oxygen species and early apoptosis. Proteomic analysis revealed that PHCl treatment downregulated proteins associated with metabolism, cell wall remodeling, and DNA repair in biofilms to levels comparable to those observed in planktonic cells. Consistent with ergosterol reduction, Erg2 was found to be reduced. Overall, PHCl disrupts key pathways essential for biofilm integrity, decreasing its stability and promoting surface detachment, underscoring its potential as a versatile antifungal compound.

IMPORTANCE: Candida albicans filamentation and biofilm formation represent crucial virulence factors promoting fungus persistence and drug tolerance. The common eukaryotic nature of mammalian cells poses significant limitations to the development of new active nontoxic compounds. Understanding the mechanism underlying PHCl inhibitory activity on yeast-hypha transition, biofilm adhesion, and maturation can pave the way to efficient drug repurposing in a field where pharmaceutical investment is lacking.

PMID:40111054 | DOI:10.1128/spectrum.02987-24

Cytoskeleton (Hoboken). 2025 Mar 20. doi: 10.1002/cm.22020. Online ahead of print.

ABSTRACT

Polo-like kinase 4 (PLK4) is a centrosome-specific kinase aberrantly expressed in cancers. Drugs inhibiting its catalytic kinase domain are under clinical phase-1/2 trials in patients with different leukemia types. However, the kinase domain of PLK4 shows structural similarity with other kinases. Therefore, drugs targeting the unique C-terminal polo-box domain (PBD) of PLK4 could provide better specificity. The knowledge of domain orientation in a full-length PLK4 structure is imperative for drug discovery. In this work, we utilized ab initio and threading approaches to predict the full-length structure of human PLK4, which was employed for virtually screening the ChEMBL library. Among the hit compounds targeting the unique regions in PLK4, we identified Alectinib, which affects centrosome numbers corresponding to PLK4 levels at centrosomes. The FT-IR analysis also confirmed Alectinib interaction with the PBD. Therefore, this work identifies a chemical scaffold that could be repurposed to target the unique regions of PLK4.

PMID:40110897 | DOI:10.1002/cm.22020

Pharmacogenomics. 2025 Mar 20:1-13. doi: 10.1080/14622416.2025.2481016. Online ahead of print.

ABSTRACT

AIM: To assess patient perspectives following evaluation in a multidisciplinary pharmacogenomics clinic run by a clinical pharmacist, genetic counselor, and physician.

METHODS: A survey was distributed to 187 adults seen in the Brigham and Women's Hospital Pharmacogenomics Clinic. Participants who completed the survey were invited to complete a semi-structured interview. Interview subjects were selected based on order of responses, scheduling availability, and range of participant experiences with testing and the clinic process. Surveys were analyzed with descriptive statistics, and interview transcripts were analyzed with thematic analysis.

RESULTS: Forty-two survey responses were received; 13 participants were interviewed. Quantitative data demonstrated high satisfaction with the multidisciplinary clinic model and belief that pharmacogenomic testing has value. Qualitative analysis identified four themes: 1) Self-Advocacy as a Patient Responsibility in the Utilization of Pharmacogenomic Results, 2) High Satisfaction with Multidisciplinary Pharmacogenomics Clinic Model and Team, 3) Utility of Pharmacogenomics, and 4) Desire for Pharmacogenomics Resources.

CONCLUSION: Patients value the care provided by a multidisciplinary pharmacogenomics clinic team, but they need to advocate for the use of their results with other healthcare professionals.

PMID:40111244 | DOI:10.1080/14622416.2025.2481016

Otolaryngol Head Neck Surg. 2025 Mar 20. doi: 10.1002/ohn.1231. Online ahead of print.

ABSTRACT

OBJECTIVE: To differentiate pediatric chronic rhinosinusitis (CRS) into clinically relevant primary and secondary phenotypes based on clinical, radiographic, and laboratory findings.

STUDY DESIGN: Retrospective chart review of patients with CRS who underwent endoscopic sinus surgery over a 5-year period.

SETTING: Tertiary referral children's hospital.

METHODS: Relevant medical and surgical history inclusive of disease onset, clinical and radiographic findings, laboratory data, and operative culture results was recorded. Data analysis resulted, where appropriate, in phenotype and endotype characterization.

RESULTS: In total, 94 patients, aged 6.8 to 22.0 years, with a mean age of 15.4 years, satisfied the inclusion criteria. Eosinophilic CRS was the most common primary phenotype (n = 19, 20.2%), and this group was the most likely to have recurrent disease requiring revision surgery. Additional primary phenotypes identified included allergic fungal rhinosinsusitis (n = 10, 10.6%) and central compartment atopic disease (n = 2, 2.1%). CRS associated with cystic fibrosis was the most common secondary CRS category (n = 13, 13.8%). Based on available data, over one-third of patients (n = 36, 38.2%) could not be categorized into a specific phenotype based on current clinical and radiologic criteria.

CONCLUSION: A phenotype and endotype-based classification system for CRS is evolving for the adult population. This study highlights that such a classification system is possible in the pediatric and adolescent age group, facilitating targeted biologic therapies at the underlying inflammatory mechanism. Further investigation is clearly required given an etiologic source of paranasal sinus inflammation could not be identified in approximately one-third of patients.

PMID:40111215 | DOI:10.1002/ohn.1231

Waste Manag Res. 2025 Mar 20:734242X241313286. doi: 10.1177/0734242X241313286. Online ahead of print.

ABSTRACT

As the Internet of things (IoT) continues to transform modern technologies, innovative applications in waste management and air pollution monitoring are becoming critical for sustainable development. In this manuscript, a novel smart waste management (SWM) and air pollution forecasting (APF) system is proposed by leveraging IoT sensors and the fully Elman neural network (FENN) model, termed as SWM-APF-IoT-FENN. The system integrates real-time data from waste and air quality sensors including weight, trash level, odour and carbon monoxide (CO) that are collected from smart bins connected to a Google Cloud Server. Here, the MaxAbsScaler is employed for data normalization, ensuring consistent feature representation. Subsequently, the atmospheric contaminants surrounding the waste receptacles were observed using a FENN model. This model is utilized to predict the atmospheric concentration of CO and categorize the bin status as filled, half-filled and unfilled. Moreover, the weight parameter of the FENN model is tuned using the secretary bird optimization algorithm for better prediction results. The implementation of the proposed methodology is done in Python tool, and the performance metrics are analysed. Experimental results demonstrate significant improvements in performance, achieving 15.65%, 18.45% and 21.09% higher accuracy, 18.14%, 20.14% and 24.01% higher F-Measure, 23.64%, 24.29% and 29.34% higher False Acceptance Rate (FAR), 25.00%, 27.09% and 31.74% higher precision, 20.64%, 22.45% and 28.64% higher sensitivity, 26.04%, 28.65% and 32.74% higher specificity, 9.45%, 7.38% and 4.05% reduced computational time than the conventional approaches such as Elman neural network, recurrent artificial neural network and long short-term memory with gated recurrent unit, respectively. Thus, the proposed method offers a streamlined, efficient framework for real-time waste management and pollution forecasting, addressing critical environmental challenges.

PMID:40111379 | DOI:10.1177/0734242X241313286

Med Phys. 2025 Mar 20. doi: 10.1002/mp.17763. Online ahead of print.

ABSTRACT

BACKGROUND: Accurate segmentation of diffuse large B-cell lymphoma (DLBCL) lesions is challenging due to their complex patterns in medical imaging. Traditional methods often struggle to delineate these lesions accurately.

OBJECTIVE: This study aims to develop a precise segmentation method for DLBCL using 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) and computed tomography (CT) images.

METHODS: We propose a 3D segmentation method based on an encoder-decoder architecture. The encoder incorporates a dual-branch design based on the shifted window transformer to extract features from both PET and CT modalities. To enhance feature integration, we introduce a multi-scale information fusion (MSIF) module that performs multi-scale feature fusion using cross-attention mechanisms with a shifted window framework. A gated neural network within the MSIF module dynamically adjusts feature weights to balance the contributions from each modality. The model is optimized using the dice similarity coefficient (DSC) loss function, minimizing discrepancies between the model prediction and ground truth. Additionally, we computed the total metabolic tumor volume (TMTV) and performed statistical analyses on the results.

RESULTS: The model was trained and validated on a private dataset of 165 DLBCL patients and a publicly available dataset (autoPET) containing 145 PET/CT scans of lymphoma patients. Both datasets were analyzed using five-fold cross-validation. On the private dataset, our model achieved a DSC of 0.7512, sensitivity of 0.7548, precision of 0.7611, an average surface distance (ASD) of 3.61 mm, and a Hausdorff distance at the 95th percentile (HD95) of 15.25 mm. On the autoPET dataset, the model achieved a DSC of 0.7441, sensitivity of 0.7573, precision of 0.7427, ASD of 5.83 mm, and HD95 of 21.27 mm, outperforming state-of-the-art methods (p < 0.05, t-test). For TMTV quantification, Pearson correlation coefficients of 0.91 (private dataset) and 0.86 (autoPET) were observed, with R2 values of 0.89 and 0.75, respectively. Extensive ablation studies demonstrated the MSIF module's contribution to enhanced segmentation accuracy.

CONCLUSION: This study presents an effective automatic segmentation method for DLBCL that leverages the complementary strengths of PET and CT imaging. The method demonstrates robust performance on both private and publicly available datasets, ensuring its reliability and generalizability. Our method provides clinicians with more precise tumor delineation, which can improve the accuracy of diagnostic interpretations and assist in treatment planning for DLBCL patients. The code for the proposed method is available at https://github.com/chenzhao2023/lymphoma_seg.

PMID:40111352 | DOI:10.1002/mp.17763

Med Phys. 2025 Mar 20. doi: 10.1002/mp.17752. Online ahead of print.

ABSTRACT

BACKGROUND: Deep learning (DL)-based systems have not yet been broadly implemented in clinical practice, in part due to unknown robustness across multiple imaging protocols.

PURPOSE: To this end, we aim to evaluate the performance of several previously developed DL-based models, which were trained to distinguish idiopathic pulmonary fibrosis (IPF) from non-IPF among interstitial lung disease (ILD) patients, under standardized reference CT imaging protocols. In this study, we utilized CT scans from non-IPF ILD subjects, acquired using various imaging protocols, to assess the model performance.

METHODS: Three DL-based models, including one 2D and two 3D models, have been previously developed to classify ILD patients into IPF or non-IPF based on chest CT scans. These models were trained on CT image data from 389 IPF and 700 non-IPF ILD patients, retrospectively, obtained from five multicenter studies. For some patients, multiple CT scans were acquired (e.g., one at inhalation and one at exhalation) and/or reconstructed (e.g., thin slice and/or thick slice). Thus, for each patient, one CT image dataset was selected to be used in the construction of the classification model, so the parameters of that data set serve as the reference conditions. In one non-IPF ILD study, due to its specific study protocol, many patients had multiple CT image data sets that were acquired under both prone and supine positions and/or reconstructed under different imaging parameters. Therefore, to assess the robustness of the previously developed models under different (e.g., non-reference) imaging protocols, we identified 343 subjects from this study who had CT data from both the reference condition (used in model construction) and non-reference conditions (e.g., evaluation conditions), which we used in this model evaluation analysis. We reported the specificities from three model under the non-reference conditions. Generalized linear mixed effects model (GLMM) was utilized to identify the significant CT technical and clinical parameters that were associated with getting inconsistent diagnostic results between reference and evaluation conditions. Selected parameters include effective tube current-time product (known as "effective mAs"), reconstruction kernels, slice thickness, patient orientation (prone or supine), CT scanner model, and clinical diagnosis. Limitations include the retrospective nature of this study.

RESULTS: For all three DL models, the overall specificity of the previously trained IPF diagnosis model decreased (p < 0.05 for two out of three models). GLMM further suggests that for at least one out of three models, mean effective mAs across the scan is the key factor that leads to the decrease in model predictive performance (p < 0.001); the difference of mean effective mAs between the reference and evaluation conditions (p = 0.03) and slice thickness (3 mm; p = 0.03) are flagged as significant factors for one out of three models; other factors are not statistically significant (p > 0.05).

CONCLUSION: Preliminary findings demonstrated the lack of robustness of IPF diagnosis model when the DL-based model is applied to CT series collected under different imaging protocols, which indicated that care should be taken as to the acquisition and reconstruction conditions used when developing and deploying DL models into clinical practice.

PMID:40111345 | DOI:10.1002/mp.17752

J Phys Chem Lett. 2025 Mar 20:3165-3176. doi: 10.1021/acs.jpclett.4c03650. Online ahead of print.

ABSTRACT

Interpretability is fundamental in the precise identification of single-molecule charge transport, and its absence in deep learning models is currently the major barrier to the usage of such powerful algorithms in the field. Here, we have pioneered a novel identification method employing fusion attention-based deep learning technologies. Central to our approach is the innovative neural network architecture, SingleFACNN, which integrates convolutional neural networks with a fusion of multihead self-attention and spatial attention mechanisms. Our findings demonstrate that SingleFACNN accurately classifies the three-type and four-type STM-BJ data sets, leveraging the convolutional layers' robust feature extraction and the attention layers' capacity to capture long-range interactions. Through comprehensive gradient-weighted class activation mapping and ablation studies, we identified and analyzed the critical features impacting classification outcomes with remarkable accuracy, thus enhancing the interpretability of our deep learning model. Furthermore, SingleFACNN's application was extended to mixed samples with varying proportions, achieving commendable prediction performance at low computational cost. Our study underscores the potential of SingleFACNN in advancing the interpretability and credibility of deep learning applications in single-molecule charge transport, opening new avenues for single-molecule detection in complex systems.

PMID:40111072 | DOI:10.1021/acs.jpclett.4c03650

J Chem Inf Model. 2025 Mar 20. doi: 10.1021/acs.jcim.5c00130. Online ahead of print.

ABSTRACT

Accurately predicting the effect of mutations on protein-protein interactions (PPIs) is essential for understanding the protein structure and function, as well as providing insights into disease-causing mechanisms. Many recent popular approaches based on the three-dimensional structure of proteins have been proposed to predict the changes in binding affinity caused by mutations, i.e. ΔΔG. However, how to effectively use the structural information to comprehensively exploit complex interactions within proteins and integrate multisource features remains a significant challenge. In this study, we propose SFM-Net, a powerful deep learning model constructed with GNN-based multiway feature extractors and a new context-aware selective fusion module that jointly leverages the sequence, structural, and evolutionary information. Such design enables SFM-Net to effectively and selectively use features from different sources to facilitate binding affinity change prediction. Benchmarking experiments and targeted ablation studies illustrate the effectiveness and robustness of our method for improving the binding affinity change prediction.

PMID:40111004 | DOI:10.1021/acs.jcim.5c00130

Aust J Rural Health. 2025 Apr;33(2):e70031. doi: 10.1111/ajr.70031.

ABSTRACT

OBJECTIVE: Diabetic retinopathy (DR) screening rates are poor in remote Western Australia where communities rely on outdated primary care-based retinal cameras. Deep learning systems (DLS) may improve access to screening, however, require validation in real-world settings. This study describes and evaluates the implementation of a new, mobile DR screening model that incorporates artificial intelligence (AI) into routine care.

DESIGN: Prospective, population-based study.

SETTING: The model was co-designed with local Aboriginal communities and implemented in the remote, Pilbara region of Western Australia. A research officer without formal healthcare qualification performed retinal screening aboard a Mercedes Sprinter Van using an automated retinal camera with integrated AI diagnostics. Patients received their diagnosis on-the-spot and completed an evaluation survey. A remote clinician provided supervision and on-the-spot telehealth consultation for referable disease.

PARTICIPANTS: People with diabetes from the Pilbara region.

MAIN OUTCOME MEASURE(S): Number of people screened, acceptability of AI to patients.

RESULTS: From February to August 2024, DR screening was provided to 9 communities across the Pilbara region. 78 patients provided research consent, of which 56.4% were Aboriginal or Torres Strait Islanders. 10.3% of retinal photos had referable DR and 8.4% of photos were ungradable. 96% of patients were 'Happy with the use of AI'.

CONCLUSION: Our new model for AI-assisted DR screening was culturally safe, acceptable to patients and effective, demonstrating an 11-fold increase in screening rates compared to 2023 Pilbara data. In remote Australian settings, AI-assisted DR screening may overcome historical barriers to service provision and improve minimisation of preventable blindness.

PMID:40110918 | DOI:10.1111/ajr.70031

SAR QSAR Environ Res. 2025 Mar 20:1-26. doi: 10.1080/1062936X.2025.2475407. Online ahead of print.

ABSTRACT

P38α has been identified as a key target for drug design to treat a wide range of diseases. In this study, multiple independent Gaussian accelerated molecular dynamics (GaMD) simulations, deep learning (DL), and the molecular mechanics generalized Born surface area (MM-GBSA) method were used to investigate the binding mechanism of inhibitors (SB2, SK8, and BMU) to DFG-in and DFG-out P38α and clarify the effect of conformational differences in P38α on inhibitor binding. GaMD trajectory-based DL effectively identified important functional domains, such as the A-loop and N-sheet. Post-processing analysis on GaMD trajectories showed that binding of the three inhibitors profoundly affected the structural flexibility and dynamical behaviour of P38α situated at the DFG-in and DFG-out states. The MM-GBSA calculations not only revealed that differences in the binding ability of inhibitors are affected by DFG-in and DFG-out conformations of P38α, but also confirmed that van der Waals interactions are the primary force driving inhibitor-P38α binding. Residue-based free energy estimation identifies hot spots of inhibitor-P38α binding across DFG-in and DFG-out conformations, providing potential target sites for drug design towards P38α. This work is expected to offer valuable theoretical support for the development of selective inhibitors of P38α family members.

PMID:40110797 | DOI:10.1080/1062936X.2025.2475407

Oncologist. 2025 Mar 10;30(3):oyaf013. doi: 10.1093/oncolo/oyaf013.

ABSTRACT

BACKGROUND: Accurate prediction of early recurrence (ER) is essential to improve the prognosis of patients with hepatocellular carcinoma (HCC) underwent thermal ablation (TA). Therefore, a deep learning model system using longitudinal magnetic resonance imaging (MRI) was developed to predict ER of patients with HCC.

METHODS: From 2014, April to 2017, May, a total of 289 eligible patients with HCC underwent TA were retrospectively enrolled from 3 hospitals and assigned into one training cohort (n = 254) and one external testing cohort (n = 35). Two deep learning models (Pre and PrePost) were developed using the pre-operative MRI and longitudinal MRI (pre- and post-operative) to predict ER for the patients with HCC after TA, respectively. Then, an integrated model (DL_Clinical) incorporating PrePost model signature and clinical variables was built for post-ablation ER risk stratification for the patients with HCC.

RESULTS: In the external testing cohort, the area under the receiver operating characteristic curve (AUC) of the DL_Clinical model was better than that of the Clinical (0.740 vs 0.571), Pre (0.740 vs 0.648), and PrePost model (0.740 vs 0.689). Additionally, there was a significant difference in RFS between the high- and low-risk groups which were divided by the DL_Clinical model (P = .04).

CONCLUSIONS: The PrePost model developed using longitudinal MRI showed outstanding performance for predicting post-ablation ER of HCC. The DL_Clinical model could stratify the patients into high- and low-risk groups, which may help physicians in treatment and surveillance strategy selection in clinical practice.

PMID:40110765 | DOI:10.1093/oncolo/oyaf013