Nat Commun. 2025 May 30;16(1):5041. doi: 10.1038/s41467-025-59435-5.

ABSTRACT

Transfer RNA (tRNA) is the most abundant cellular RNA family in terms of copy numbers. It not only folds into defined structures but also has complex cellular interaction networks involving aminoacyl-tRNA synthetases, translation factors, and ribosomes. The human tRNAome is comprised of chromosomal-encoded tRNAs with a large sequence diversity and mitochondrial-encoded tRNAs with A/U-rich sequences and noncanonical tertiary interactions. How tRNA folding and interactions in a eukaryotic cell respond to stress is poorly understood. Here, we develop DM-DMS-MaPseq, which utilizes in vivo dimethyl-sulfate (DMS) chemical probing and mutational profiling (MaP) coupled with demethylase (DM) treatment in transcriptome-wide tRNA sequencing to profile structures and the cellular interactions of human chromosomal and mitochondrial-encoded tRNAs. We found that tRNAs maintain stable structures in vivo, but the in vivo DMS profiles are vastly different from those in vitro, which can be explained by their interactions with cellular proteins and the ribosome. We also identify cytosolic and mitochondrial tRNA structure and interaction changes upon arsenite treatment, a type of oxidative stress that induces translational reprogramming, which is consistent with global translation repression in both compartments. Our results reveal variations of tRNA structurome and dynamic interactome that have functional consequences in translational regulation.

PMID:40447571 | DOI:10.1038/s41467-025-59435-5

Stem Cell Reports. 2025 May 20:102511. doi: 10.1016/j.stemcr.2025.102511. Online ahead of print.

ABSTRACT

Human embryonic stem cells (hESCs) are notable for their ability to self-renew and to differentiate into all tissue types in the body. NANOG is a core regulator of hESC identity, and dynamic control of its expression is crucial to maintain the balance between self-renewal and differentiation. Transcriptional regulation depends on enhancers, but NANOG enhancers in hESCs are not well characterized. Here, we report two NANOG enhancers discovered from a CRISPR interference screen in hESCs. Deletion of a single copy of either enhancer significantly reduced NANOG expression, compromising self-renewal and increasing differentiation propensity. Interestingly, these two NANOG enhancers are involved in a tandem duplication event found in certain primates including humans but not in mice. However, the duplicated counterparts do not regulate NANOG expression. This work expands our knowledge of functional enhancers in hESCs and highlights the sensitivity of the hESC state to the dosage of core regulators and their enhancers.

PMID:40446796 | DOI:10.1016/j.stemcr.2025.102511

Thorax. 2025 May 30:thorax-2024-222242. doi: 10.1136/thorax-2024-222242. Online ahead of print.

ABSTRACT

BACKGROUND: Despite significant clinical improvements, there is evidence of persisting airway inflammation in people with cystic fibrosis (CF) established on elexacaftor/tezacaftor/ivacaftor (ETI) therapy. As CF is a multi-system disease, systemic immune profiles can reflect local inflammation within the lungs and other organs. Understanding systemic inflammation after ETI therapy may reveal important translational insights. This study aims to profile systemic inflammatory changes and relate these to the well-documented improvements observed with ETI therapy.

METHODS: We conducted a single-centre longitudinal study with 57 CF subjects initiating ETI therapy. All participants were Phe508del homozygous or Phe508del/minimal function. Blood samples were collected pre-ETI and 3-12 months post-therapy initiation. Analyses included mass spectrometry-based proteomics, a multiplex immunoassay, and flow cytometry for peripheral immune cell counts and phenotype. Controls samples were provided by 29 age-matched healthy controls.

RESULTS: Systemic inflammation reduced with ETI therapy; however, the immune profile remained distinct from healthy controls. ETI reduced neutrophil counts and was associated with a more mature, less inflammatory phenotype, as well as a shift towards an immune resolving state associated with increased CD206 expression. Cytokines known to influence neutrophil levels reduced with therapy. Despite ETI therapy, neutrophil and monocyte counts remained elevated compared with healthy controls. There was no obvious association between the ETI-related improvements in systemic inflammation and lung function.

CONCLUSIONS: Patients with CF showed evidence of persisting systemic inflammation despite ETI therapy, which may have long-term potentially adverse effects on respiratory and other organ systems.

PMID:40447326 | DOI:10.1136/thorax-2024-222242

Cancer Cell. 2025 May 28:S1535-6108(25)00211-9. doi: 10.1016/j.ccell.2025.05.002. Online ahead of print.

ABSTRACT

Mycobacterium bovis Bacillus Calmette-Guérin (BCG) is the vaccine against tuberculosis and an immunotherapy for bladder cancer. When administered intravenously, BCG reprograms bone marrow hematopoietic stem and progenitor cells (HSPCs), leading to heterologous protection against infections. Whether HSPC reprogramming contributes to the anti-tumor effects of BCG administered into the bladder is unknown. We demonstrate that BCG administered in the bladder colonizes the bone marrow and, in both mice and humans, reprograms HSPCs to alter and amplify myelopoiesis. BCG-reprogrammed HSPCs are sufficient to confer augmented anti-tumor immunity through production of neutrophils, monocytes, and dendritic cells that broadly remodel the tumor microenvironment, drive T cell-dependent anti-tumor responses, and synergize with checkpoint blockade. We conclude that bladder BCG acts systemically through hematopoiesis, highlighting the broad potential of HSPC reprogramming to enhance the innate drivers of T cell-dependent tumor immunity.

PMID:40446799 | DOI:10.1016/j.ccell.2025.05.002

Lancet Haematol. 2025 Jun;12(6):e451-e462. doi: 10.1016/S2352-3026(25)00035-3.

ABSTRACT

As the therapeutic landscape in haematological malignancies has evolved from traditional chemotherapies to novel biological, targeted, and cellular therapies, adverse event profiles have accordingly shifted with emerging and newly described chronic, cumulative, and delayed symptomatic adverse events. The current standard of toxicity reporting in clinical trials, centred on maximum-grade adverse events, is wholly inadequate for characterising the tolerability of therapies in the modern era. As such, the science of adverse event measurement, analysis, and reporting in clinical trials needs to evolve with our ever-growing repertoire of therapeutics to facilitate more comprehensive and accurate toxicity assessment for treatment decision making. In this first paper in the Adverse Event Reporting Series, a follow-up of a 2018 Lancet Haematology Commission, we review advances in the reporting of newly described adverse events and toxicity domains in haematological malignancies, emerging clinical trial designs to more accurately identify optimal dosing strategies through enhanced adverse event measurement, and novel analytic and visualisation tools to facilitate interpretation of trial adverse event data.

PMID:40447353 | DOI:10.1016/S2352-3026(25)00035-3

Prim Care Companion CNS Disord. 2025 May 29;27(3):24r03857. doi: 10.4088/PCC.24r03857.

ABSTRACT

Objective: To conduct a systematic review of the available evidence on hydroxychloroquine (HCQ)-induced psychiatric side effects and their management.

Data sources: A literature search was conducted in PubMed, MEDLINE, PsycINFO, and Cochrane collaboration databases from 2000 to 2024 using the keywords "hydroxychloroquine" AND "psychiatry" OR "psychosis" OR "depression" OR "anxiety" OR "bipolar disorder" OR "delirium" OR "psychotic disorders" OR "psychiatric side effects" OR "psychiatric disorders."

Study selection: English-language articles with studies reporting HCQ-induced psychiatric/neuropsychiatric side effects were included. Duplicate records and studies reporting only chloroquine side effects were excluded.

Results: The review included 16 case reports, 8 original articles, and 3 review articles. HCQ was found to trigger symptoms of psychosis, depression, suicidal ideation, mania/hypomania, anxiety, sleep disturbances, and cognitive impairments. The onset of these psychiatric side effects varied, appearing shortly after starting the medication to a more extended period.

Conclusion: Based on the literature, HCQ may be associated with short-term psychiatric adverse effects. A psychiatric consultation for a thorough clinical and risk factor evaluation to differentiate a primary psychiatric disorder from a drug induced adverse effect would help guide the management. Dosage adjustments, discontinuing HCQ if feasible, and psychotropic medications like olanzapine or risperidone may be necessary when psychiatric side effects are secondary to HCQ. Further studies are needed to validate these findings.

Prim Care Companion CNS Disord 2025;27(3):24r03857.

Author affiliations are listed at the end of this article.

PMID:40446824 | DOI:10.4088/PCC.24r03857

PLoS One. 2025 May 30;20(5):e0324146. doi: 10.1371/journal.pone.0324146. eCollection 2025.

ABSTRACT

Drug-target interactions (DTIs) play a critical role in drug discovery and repurposing. Deep learning-based methods for predicting drug-target interactions are more efficient than wet-lab experiments. The extraction of original and substructural features from drugs and proteins plays a key role in enhancing the accuracy of DTI predictions, while the integration of multi-feature information and effective representation of interaction data also impact the precision of DTI forecasts. Consequently, we propose a drug-target interaction prediction model, SSCPA-DTI, based on substructural subsequences and a cross co-attention mechanism. We use drug SMILES sequences and protein sequences as inputs for the model, employing a Multi-feature information mining module (MIMM) to extract original and substructural features of DTIs. Substructural information provides detailed insights into molecular local structures, while original features enhance the model's understanding of the overall molecular architecture. Subsequently, a Cross-public attention module (CPA) is utilized to first integrate the extracted original and substructural features, then to extract interaction information between the protein and drug, addressing issues such as insufficient accuracy and weak interpretability arising from mere concatenation without interactive integration of feature information. We conducted experiments on three public datasets and demonstrated superior performance compared to baseline models.

PMID:40445972 | DOI:10.1371/journal.pone.0324146

Depress Anxiety. 2025 May 22;2025:9002779. doi: 10.1155/da/9002779. eCollection 2025.

ABSTRACT

Background: People living with a rare disease are a vulnerable patient group and experience challenges in participation and healthcare. Due to changes in healthcare and threat of the infection during coronavirus disease 2019 (COVID-19) pandemic, people living with rare diseases have been particularly affected. Therefore, this study aimed to investigate depressive symptoms and symptoms of anxiety during the COVID-19 pandemic and identify factors associated with symptom levels. Methods: One-hundred and seventy-two people living with a rare disease were recruited from centers for rare diseases and patient organizations in Germany from January 2021 to January 2022. In addition to descriptive analyses and group comparisons, we applied multiple linear regression models to identify factors associated with outcome variables of interest (depressive and anxiety symptoms, assessed by the Hospital Anxiety and Depression Scale [HADS]). Results: For the depressive symptoms, 14% of the participants reached the cutoff for moderate and 14.5% for a high level of depressive symptoms. Concerning anxiety symptoms, 22% reported moderate levels of anxiety and 13.4% reported high levels of anxiety. Higher depressive symptoms were significantly associated with older age, lower socioeconomic status, having severe or varying symptoms compared to low symptom severity, lower treatment satisfaction, lower social support, and more unmet needs. Higher anxiety levels were associated with more unmet needs and more intense COVID-19-related concerns. Conclusions: The findings indicate red flags of high symptoms that should be considered during routine care of patients with rare diseases. Healthcare providers should be sensitized for the need for psychosocial support and use a quick assessment to assign patients in need to specific support programs. Trial Registration: German Clinical Trials Registry: DRKS00020488.

PMID:40444181 | PMC:PMC12122157 | DOI:10.1155/da/9002779

PLoS One. 2025 May 30;20(5):e0309875. doi: 10.1371/journal.pone.0309875. eCollection 2025.

ABSTRACT

Kidney transplantation recipients (KTRs) represent a vulnerable population for COVID-19 infection and severe disease. Nirmatrelvir-ritonavir has demonstrated efficacy in treating COVID-19 among KTRs, and interacts with tacrolimus leading to a precipitous increase in tacrolimus blood levels when co-administered, which may potentially result in toxicity. To explore a safe strategy for the combination of nirmatrelvir-ritonavir and tacrolimus, we established a new administration strategy to restore tacrolimus after the discontinuation of nirmatrelvir-ritonavir and conducted a real-world retrospective observational cohort study to evaluate its clinical efficacy. In the experimental group, tacrolimus was initiated at 20-25% of the baseline dose 48 hours after the discontinuation of nirmatrelvir-ritonavir, with daily increments of 20-25% until the baseline dose was restored. The patients who did not follow the experimental protocol were included in the control group. Results showed that withholding tacrolimus 12 hours before starting nirmatrelvir-ritonavir maintained tacrolimus blood levels above 83% of the baseline throughout the nirmatrelvir-ritonavir treatment period. Compared with the control group, the experimental group achieved target trough concentrations of tacrolimus more quickly and maintained a higher proportion within the therapeutic range (p = 0.029), and had significantly lower rates of adverse events (p = 0.002, OR = 0.308, 95%CI:0.136-0.695). This study provides a safe and effective pharmacological strategy for KTRs infected with COVID-19, allowing the safe co-administration of nirmatrelvir-ritonavir and tacrolimus.

PMID:40445936 | DOI:10.1371/journal.pone.0309875

Alcohol Clin Exp Res (Hoboken). 2025 May 30. doi: 10.1111/acer.70052. Online ahead of print.

ABSTRACT

BACKGROUND: In two 12-week, randomized, placebo-controlled trials (RCTs) in individuals with alcohol use disorder (AUD), topiramate significantly reduced heavy drinking days (HDDs), and alcohol-related problems. In a secondary analysis of those findings, we examined four broad measures of genetic risk-polygenic scores (PGS)-of problematic alcohol use (PAU), drinks per week (DPW), and time to relapse to any drinking (TR) and heavy drinking (THR) as moderators of topiramate's effect on HDDs and alcohol-related problems.

METHODS: We analyzed data from 285 individuals with AUD (65.6% male) of European-like ancestry, who were treated with either topiramate (49.1%) or placebo (50.9%). All patients underwent genome-wide array genotyping, and PGS were calculated using summary statistics from genome-wide association studies of PAU, DPW, and TR and THR (two time-to-event outcomes among patients treated in AUD pharmacotherapy trials). We hypothesized an interaction effect in which greater genetic risk-particularly for PAU-would be associated with a greater therapeutic response to topiramate than placebo.

RESULTS: As shown previously, topiramate significantly reduced both HDDs (odds ratio [OR] = 0.50, p < 0.001) and Short Index of Problems (SIP) scores (b = -3.04, p < 0.001) more than placebo. There were nonsignificant associations of higher PGS with more HDDs (OR = 1.17, 95% CI = 0.98-1.41, p = 0.091) and a greater reduction in HDDs in the topiramate group (OR = 0.80, 95% CI = 0.62-1.03, p = 0.089). There were also significant interaction effects with treatment on SIP score by PGS for PAU (b = -1.64, SE = 0.78, p = 0.033), TR (b = -2.16, SE = 0.72, p = 0.003), and TRH (b = -2.17, SE = 0.72, p = 0.003).

CONCLUSIONS: These findings provide proof of principle for the use of alcohol-related PGS as moderators of the effects of topiramate for treating AUD. Larger RCTs of topiramate are needed to provide adequate statistical power to validate this pharmacogenetic approach to precision AUD treatment.

PMID:40445294 | DOI:10.1111/acer.70052

PLoS Pathog. 2025 May 30;21(5):e1013190. doi: 10.1371/journal.ppat.1013190. Online ahead of print.

ABSTRACT

Mycobacterium abscessus (MAB) is a clinically significant multidrug-resistant (MDR) pathogen, particularly implicated in pulmonary infections among cystic fibrosis (CF) patients. Tedizolid (TZD), an oxazolidinone-class antibacterial drug, has been recommended as an alternative treatment for MAB-infected patients who are intolerant to or whose isolate is resistant to first-line drugs including linezolid (LZD). To investigate the TZD resistance mechanisms in MAB, we isolated 23 TZD-resistant MAB mutants and performed whole-genome sequencing (WGS) to identify resistance-associated genes. Frequent mutations were identified in MAB_2885, encoding a putative TetR transcriptional regulator, and MAB_2303, encoding a putative mycobacterial membrane protein large (MmpL). Drug susceptibility testing confirmed that MAB_2885 mutations contribute to both TZD and LZD resistance in MAB. RNA-seq analysis revealed that restoring wild-type MAB_2885 in mutants downregulated the MAB_2302-MAB_2303. Electrophoretic mobility shift assay (EMSA) showed the MAB_2885 protein binds to its target sequence upstream of MAB_2302-MAB_2303, further confirming their regulatory relationship. The W91R mutation in the MAB_2885 protein was found to impair its DNA-binding activity compared to the wild-type. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis confirmed that MAB_2302-MAB_2303 functions as a TZD efflux pump. Additionally, overexpression of MAB_2885 in M. abscessus subsp. bolletii and M. abscessus subsp. massiliense also increased their TZD susceptibility and downregulated their respective MmpS-MmpL orthologs. Overall, our study demonstrates that mutations in MAB_ 2885 contribute to TZD and LZD resistance by disrupting the negative regulation of the downstream MAB_2302-MAB_2303, which functions as a direct efflux pump for TZD. These findings provide new insights into oxazolidinone resistance mechanisms in MAB and identify potential biomarkers for detecting drug resistance.

PMID:40445981 | DOI:10.1371/journal.ppat.1013190

Nanomedicine (Lond). 2025 May 30:1-17. doi: 10.1080/17435889.2025.2509477. Online ahead of print.

ABSTRACT

Messenger RNA (mRNA)-based therapeutics offer the potential to treat a variety of pulmonary disorders that arise due to genetics, infectious diseases, and chronic respiratory conditions. However, various physiological barriers in the lungs, such as mucociliary clearance, macrophage phagocytosis, and lung surfactant interference, present challenges for efficient mRNA delivery. Polymeric nanoparticles (NPs) have emerged as a therapeutic platform for delivering mRNA therapeutics due to their stability, tunability, and controlled release properties, making them suitable and potentially ideal for encapsulating and protecting mRNA molecules for delivery in vivo. Continued advances in polymer and NP design have improved mucus penetration and cellular uptake upon lung delivery; further, administration via local and systemic routes enable modulation of NP biodistribution. These advancements benefit the potential treatment of a range of pulmonary diseases, including viral infections, cystic fibrosis (CF), asthma, and lung cancer, by facilitating immune modulation and genetic therapy delivery. In this review, we explore how polymeric NPs address disease-specific requirements and physiological challenges to expand the potential for therapeutic mRNAs in the lung.

PMID:40445199 | DOI:10.1080/17435889.2025.2509477

J Investig Med. 2025 May 30:10815589251348918. doi: 10.1177/10815589251348918. Online ahead of print.

ABSTRACT

Bacterial extracellular vesicles (EVs) are important mediators of host infection. Persons with cystic fibrosis (CF) often suffer from chronic infection with Pseudomonas aeruginosa, an opportunistic pathogen. However, the relative abundance of P. aeruginosa is not associated with the onset of increased pulmonary symptoms, known as a pulmonary exacerbation. We hypothesized that the cargo of P. aeruginosa EVs is different at times of baseline wellness and pulmonary exacerbation onset in persons with CF. This is the first study to characterize and compare P. aeruginosa EVs at these two time points, using a novel series of steps to isolate the P. aeruginosa EVs directly from the sputum of persons with CF. Our study found a differential packaging of P. aeruginosa proteins at baseline wellness and pulmonary exacerbation, with six proteins being more frequently present in pulmonary exacerbation samples. Additionally, we were able to demonstrate the P. aeruginosa EVs isolated from the sputum of persons with CF at the time of pulmonary exacerbation induced an inflammatory response in CF human bronchial epithelial (HBE) cells. These data, while preliminary, support the clinical relevance of P. aeruginosa EVs in influencing gene regulation and lung inflammation and immunity in persons with CF.

PMID:40444885 | DOI:10.1177/10815589251348918

JMIR Form Res. 2025 May 30;9:e56057. doi: 10.2196/56057.

ABSTRACT

BACKGROUND: Conventional approaches for major depressive disorder (MDD) screening rely on two effective but subjective paradigms: self-rated scales and clinical interviews. Artificial intelligence (AI) can potentially contribute to psychiatry, especially through the use of objective data such as objective audiovisual signals.

OBJECTIVE: This study aimed to evaluate the efficacy of different paradigms using AI analysis on audiovisual signals.

METHODS: We recruited 89 participants (mean age, 37.1 years; male: 30/89, 33.7%; female: 59/89, 66.3%), including 41 patients with MDD and 48 asymptomatic participants. We developed AI models using facial movement, acoustic, and text features extracted from videos obtained via a tool, incorporating four paradigms: conventional scale (CS), question and answering (Q&A), mental imagery description (MID), and video watching (VW). Ablation experiments and 5-fold cross-validation were performed using two AI methods to ascertain the efficacy of paradigm combinations. Attention scores from the deep learning model were calculated and compared with correlation results to assess comprehensibility.

RESULTS: In video clip-based analyses, Q&A outperformed MID with a mean binary sensitivity of 79.06% (95%CI 77.06%-83.35%; P=.03) and an effect size of 1.0. Among individuals, the combination of Q&A and MID outperformed MID alone with a mean extent accuracy of 80.00% (95%CI 65.88%-88.24%; P= .01), with an effect size 0.61. The mean binary accuracy exceeded 76.25% for video clip predictions and 74.12% for individual-level predictions across the two AI methods, with top individual binary accuracy of 94.12%. The features exhibiting high attention scores demonstrated a significant overlap with those that were statistically correlated, including 18 features (all Ps<.05), while also aligning with established nonverbal markers.

CONCLUSIONS: The Q&A paradigm demonstrated higher efficacy than MID, both individually and in combination. Using AI to analyze audiovisual signals across multiple paradigms has the potential to be an effective tool for MDD screening.

PMID:40446148 | DOI:10.2196/56057

PLoS One. 2025 May 30;20(5):e0321559. doi: 10.1371/journal.pone.0321559. eCollection 2025.

ABSTRACT

Deep learning has revolutionized the classification of land cover objects in hyperspectral images (HSIs), particularly by managing the complex 3D cube structure inherent in HSI data. Despite these advances, challenges such as data redundancy, computational costs, insufficient sample sizes, and the curse of dimensionality persist. Traditional 2D Convolutional Neural Networks (CNNs) struggle to fully leverage the interconnections between spectral bands in HSIs, while 3D CNNs, which capture both spatial and spectral features, require more sophisticated design. To address these issues, we propose a novel multilayered, multi-branched 2D-3D CNN model in this paper that integrates Segmented Principal Component Analysis (SPCA) and the minimum-Redundancy-Maximum-Relevance (mRMR) technique. This approach explores the local structure of the data and ranks features by significance. Our approach then hierarchically processes these features: the shallow branch handles the least significant features, the deep branch processes the most critical features, and the mid branch deals with the remaining features. Experimental results demonstrate that our proposed method outperforms most of the state-of-the-art techniques on the Salinas Scene, University of Pavia, and Indian Pines hyperspectral image datasets achieving 100%, 99.94%, and 99.12% Overall Accuracy respectively.

PMID:40446012 | DOI:10.1371/journal.pone.0321559

PLoS One. 2025 May 30;20(5):e0322405. doi: 10.1371/journal.pone.0322405. eCollection 2025.

ABSTRACT

Arsenic contamination of drinking water is a significant health risk. Countries such as Bangladesh's rural areas and regions are in the red alert zone because groundwater is the only primary source of drinking. Early detection of arsenic disease is critical for mitigating long-term health issues. However, these approaches are not widely accepted. In this study, we proposed a fusion approach for the detection of arsenic skin disease. The proposed model is a combination of the Xception model with the Inception module in a deep learning architecture named "ArsenicNet." The model was trained and tested on a publicly available image dataset named "ArsenicSkinImageBD" which contains only 1287 samples and is based on Bangladeshi people. The proposed model achieved the best accuracy through proper experimentation compared to several state-of-the-art deep learning models, including InceptionV3, VGG19, EfficientNetV2B0, ResNet152V2, ViT, and Xception. The proposed model achieved an accuracy of 97.69% and an F1 score of 97.63%, demonstrating superior performance. This research indicates that our proposed model can detect complex patterns in which arsenic skin disease is present, leading to a superior detection performance. Moreover, data augmentation techniques and earlystoping function were used to prevent models overfitting. This study highlights the potential of sophisticated deep learning methodologies to enhance the accuracy of arsenic detection and prevent premature interventions in the diagnosis of arsenic-related illnesses in people. This research contributes to ongoing efforts to develop robust and scalable solutions to monitor and manage arsenic contamination-related health issues.

PMID:40446004 | DOI:10.1371/journal.pone.0322405

PLoS One. 2025 May 30;20(5):e0321770. doi: 10.1371/journal.pone.0321770. eCollection 2025.

ABSTRACT

Accurate diagnosis of apple diseases is vital for tree health, yield improvement, and minimizing economic losses. This study introduces a deep learning-based model to tackle issues like limited datasets, small sample sizes, and low recognition accuracy in detecting apple leaf diseases. The approach begins with enhancing the CycleGAN-M network using a multi-scale attention mechanism to generate synthetic samples, improving model robustness and generalization by mitigating imbalances in disease-type representation. Next, an improved YOLOv8s-KEF model is introduced to overcome limitations in feature extraction, particularly for small lesions and complex textures in natural environments. The model's backbone replaces the standard C2f structure with C2f-KanConv, significantly enhancing disease recognition capabilities. Additionally, we optimize the detection head with Efficient Multi-Scale Convolution (EMS-Conv), improving the model's ability to detect small targets while maintaining robustness and generalization across diverse disease types and conditions. Incorporating Focal-EIoU further reduces missed and false detections, enhancing overall accuracy. The experiment results demonstrate that the YOLOv8s-KEF model achieves 95.0% in accuracy, 93.1% in recall, 95.8% in precision, and an F1-score of 94.5%. Compared to the original YOLOv8s model, the proposed model improves accuracy by 7.2%, precision by 6.5%, and F1-score by 5.0%, with only a modest 6MB increase in model size. Furthermore, compared to Faster RCNN, ResNet50, SSD, YOLOv3-tiny, YOLOv6, YOLOv9s, and YOLOv10m, our model demonstrates substantial improvements, with up to 30.2% higher precision and 18.0% greater accuracy. This study used CycleGAN-M and YOLOv8s-KEF methods to enhance the detection capability of apple leaf diseases.

PMID:40445983 | DOI:10.1371/journal.pone.0321770

PLoS One. 2025 May 30;20(5):e0324146. doi: 10.1371/journal.pone.0324146. eCollection 2025.

ABSTRACT

Drug-target interactions (DTIs) play a critical role in drug discovery and repurposing. Deep learning-based methods for predicting drug-target interactions are more efficient than wet-lab experiments. The extraction of original and substructural features from drugs and proteins plays a key role in enhancing the accuracy of DTI predictions, while the integration of multi-feature information and effective representation of interaction data also impact the precision of DTI forecasts. Consequently, we propose a drug-target interaction prediction model, SSCPA-DTI, based on substructural subsequences and a cross co-attention mechanism. We use drug SMILES sequences and protein sequences as inputs for the model, employing a Multi-feature information mining module (MIMM) to extract original and substructural features of DTIs. Substructural information provides detailed insights into molecular local structures, while original features enhance the model's understanding of the overall molecular architecture. Subsequently, a Cross-public attention module (CPA) is utilized to first integrate the extracted original and substructural features, then to extract interaction information between the protein and drug, addressing issues such as insufficient accuracy and weak interpretability arising from mere concatenation without interactive integration of feature information. We conducted experiments on three public datasets and demonstrated superior performance compared to baseline models.

PMID:40445972 | DOI:10.1371/journal.pone.0324146

PLoS One. 2025 May 30;20(5):e0320362. doi: 10.1371/journal.pone.0320362. eCollection 2025.

ABSTRACT

This study aimed to compare image quality and solid focal liver lesion (FLL) assessments between free-breathing, diffusion-weighted imaging using deep learning reconstruction (FB-DL-DWI) and conventional DWI (FB-C-DWI) in patients undergoing clinically indicated liver MRIs. Our retrospective study included 199 patients who underwent 3 T-liver MRIs with FB-DL-DWI and FB-C-DWI. DWI was performed using a single-shot, spin-echo, echo-planar, fat suppression technique during free-breathing with matching parameters. Three radiologists independently evaluated subjective image quality across two sequences. The apparent diffusion coefficient (ADC) was measured in 15 liver regions. Four radiologists analyzed 138 solid FLLs from 60 patients for the presence of diffusion restriction, lesion conspicuity, and sharpness. Among the 199 patients, 110 (55.3%) had underlying chronic liver disease (CLD). FB-DL-DWI was found to be 43.0% faster than FB-C-DWI (119.4 ± 2.2 sec vs. 209.6 ± 3.7 sec). Furthermore, FB-DL-DWI scored higher than FB-C-DWI for all subjective image quality parameters (all, P < 0.001); however, FB-DL-DWI exhibited greater artificial sensation than FB-C-DWI (P < 0.001). In patients with CLD, FB-DL-DWI exhibited a better subjective image quality (all, P < 0.001) than FB-C-DWI. ADC values ranged from 1.06-1.12 × 10-3 mm2/sec in FB-DL-DWI and 1.06-1.20 × 10-3 mm2/sec in FB-C-DWI. Among the 138 lesions analyzed, 116 malignancies (61 hepatocellular carcinomas, 3 cholangiocarcinomas, 52 metastases) and 22 benignities were included. Four readers identified 88, 93, 93, and 105 diffusion-restricted FLLs in FB-DL-DWI and 84, 80, 98, and 95 in FB-C-DWI. FB-DL-DWI (75.9-90.5%) demonstrated comparable or superior diffusion restriction rates for malignant FLLs compared to FB-C-DWI (68.1-82.8%). Furthermore, FB-DL-DWI presented higher lesion-edge sharpness and lesion-conspicuity compared to FB-C-DWI. Overall, FB-DL-DWI provided better image quality, lesion sharpness, and conspicuity for solid FLLs, with a shorter acquisition time than FB-C-DWI. Therefore, FB-DL-DWI may replace FB-C-DWI as the preferred imaging method for liver evaluations.

PMID:40445963 | DOI:10.1371/journal.pone.0320362

PLoS One. 2025 May 30;20(5):e0322738. doi: 10.1371/journal.pone.0322738. eCollection 2025.

ABSTRACT

Outlier detection is essential for identifying unusual patterns or observations that significantly deviate from the normal behavior of a dataset. With the rapid growth of data science, the prevalence of anomalies and outliers has increased, which can disrupt system modeling and parameter estimation, leading to inaccurate results. Recently, deep learning-based outlier detection methods have gained significant attention, but their performance is often limited by challenges in parameter selection and the nearest neighbor search. To overcome these limitations, we propose a three-stage Efficient Outlier Detection Approach (named EODA), that not only detects outliers with high accuracy but also emphasizes dataset characteristics. In the first stage, we apply a feature selection algorithm based on the Boruta method and Random Forest to reduce the data size by selecting the most relevant attributes and calculating the highest Z-score of shadow features. In the second stage, we improve the K-nearest neighbors algorithm to enhance the accuracy of nearest neighbor identification in the clustering phase. Finally, the third stage efficiently identifies the most significant outliers within clustered datasets. We evaluate the proposed EODA algorithm across eight UCI machine-learning repository datasets. The results demonstrate the effectiveness of our EODA approach, achieving a Precision of 63.07%, Recall of 82.49%, and an F1-Score of 64.53%, outperforming the existing techniques in the field.

PMID:40445940 | DOI:10.1371/journal.pone.0322738