J Comput Chem. 2025 Jan 30;46(3):e70011. doi: 10.1002/jcc.70011.

ABSTRACT

In the realm of artificial intelligence-driven drug discovery (AIDD), accurately predicting the influence of molecular structures on their properties is a critical research focus. While deep learning models based on graph neural networks (GNNs) have made significant advancements in this area, prior studies have primarily concentrated on molecule-level representations, often neglecting the impact of functional group structures and the potential relationships between fragments on molecular property predictions. To address this gap, we introduce the multi-scale feature attention graph neural network (MfGNN), which enhances traditional atom-based molecular graph representations by incorporating fragment-level representations derived from chemically synthesizable BRICS fragments. MfGNN not only effectively captures both the structural information of molecules and the features of functional groups but also pays special attention to the potential relationships between fragments, exploring how they collectively influence molecular properties. This model integrates two core mechanisms: a graph attention mechanism that captures embeddings of molecules and functional groups, and a feature extraction module that systematically processes BRICS fragment-level features to uncover relationships among the fragments. Our comprehensive experiments demonstrate that MfGNN outperforms leading machine learning and deep learning models, achieving state-of-the-art performance in 8 out of 11 learning tasks across various domains, including physical chemistry, biophysics, physiology, and toxicology. Furthermore, ablation studies reveal that the integration of multi-scale feature information and the feature extraction module enhances the richness of molecular features, thereby improving the model's predictive capabilities.

PMID:39840745 | DOI:10.1002/jcc.70011

Acta Radiol. 2025 Jan 22:2841851241307333. doi: 10.1177/02841851241307333. Online ahead of print.

ABSTRACT

BACKGROUND: Segmentation of the cochlea in temporal bone computed tomography (CT) is the basis for image-guided otologic surgery. Manual segmentation is time-consuming and laborious.

PURPOSE: To assess the utility of deep learning analysis in automatic segmentation of the cochleae in temporal bone CT to differentiate abnormal images from normal images.

MATERIAL AND METHODS: Three models (3D U-Net, UNETR, and SegResNet) were trained to segment the cochlea on two CT datasets (two CT types: GE 64 and GE 256). One dataset included 77 normal samples, and the other included 154 samples (77 normal and 77 abnormal). A total of 20 samples that contained normal and abnormal cochleae in three CT types (GE 64, GE 256, and SE-DS) were tested on the three models. The Dice similarity coefficient (DSC) and Hausdorff distance (HD) were used to assess the models.

RESULTS: The segmentation performances of the three models improved after adding abnormal cochlear images for training. SegResNet achieved the best performance. The average DSC on the test set was 0.94, and the HD was 0.16 mm; the performance was higher than those obtained by the 3D U-Net and UNETR models. The DSCs obtained using the GE 256 CT, SE-DS CT, and GE 64 CT models were 0.95, 0.94, and 0.93, respectively, and the HDs were 0.15, 0.18, and 0.12 mm, respectively.

CONCLUSION: The SegResNet model is feasible and accurate for automated cochlear segmentation of temporal bone CT images.

PMID:39840644 | DOI:10.1177/02841851241307333

Lung. 2025 Jan 22;203(1):27. doi: 10.1007/s00408-024-00783-2.

ABSTRACT

PURPOSE: Pulmonary hypertension (PH) is associated with morbidity and mortality in patients with interstitial lung disease (ILD). Several prediction models have been proposed to predict PH in ILD patients. We sought to discern how previously described prediction models perform in predicting PH in patients with ILD.

METHODS: Patients with ILD who completed a baseline right heart catheterization, from Inova Fairfax Hospital, Northwestern Memorial Hospital, and Asan Medical Center in Korea were enrolled. The performance of various prediction models (FORD model, the FORD calculator, the PH-ILD Detection tool, and the mean pulmonary artery pressure prediction model) were assessed using receiver operating characteristic (ROC) curves and area under the receiver operating characteristic curve (AUROC). There were four definitions of pulmonary hypertension against which the models were evaluated.

RESULTS: There were a total of 192 patients with ILD, of whom 32.8% (n = 63/192) met the modified 5th world symposium on PH definition of precapillary PH. Among the models assessed, the FORD calculator had an AUROC (0.733) that was marginally highest. Subgroup analysis revealed that the FORD index had the highest AUROC (0.817) in patients with idiopathic pulmonary fibrosis, while the FORD calculator had the highest AUROC (0.751) in patients with non-IPF ILD.

CONCLUSION: The FORD model can be used to predict group 3 PH in both IPF patients and non-IPF ILD patients. It could serve as a tool for ILD patient selection for right heart catheterization as well as an enrichment tool for clinical trials targeting the pulmonary vasculature.

PMID:39841267 | DOI:10.1007/s00408-024-00783-2

Elife. 2025 Jan 22;12:RP90841. doi: 10.7554/eLife.90841.

ABSTRACT

Audiovisual information reaches the brain via both sustained and transient input channels, representing signals' intensity over time or changes thereof, respectively. To date, it is unclear to what extent transient and sustained input channels contribute to the combined percept obtained through multisensory integration. Based on the results of two novel psychophysical experiments, here we demonstrate the importance of the transient (instead of the sustained) channel for the integration of audiovisual signals. To account for the present results, we developed a biologically inspired, general-purpose model for multisensory integration, the multisensory correlation detectors, which combines correlated input from unimodal transient channels. Besides accounting for the results of our psychophysical experiments, this model could quantitatively replicate several recent findings in multisensory research, as tested against a large collection of published datasets. In particular, the model could simultaneously account for the perceived timing of audiovisual events, multisensory facilitation in detection tasks, causality judgments, and optimal integration. This study demonstrates that several phenomena in multisensory research that were previously considered unrelated, all stem from the integration of correlated input from unimodal transient channels.

PMID:39841060 | DOI:10.7554/eLife.90841

Adv Biol (Weinh). 2025 Jan 22:e2400713. doi: 10.1002/adbi.202400713. Online ahead of print.

ABSTRACT

Polycystic ovary syndrome is one of the most common endocrine disorders in women of reproductive age, characterized by functional and structural alterations of the female reproductive organs. Due to the unknown underlying molecular mechanisms, in vivo murine models and in vitro human cellular models are developed to study the syndrome. These models are used to analyze various aspects of the pathology by replicating the conditions of the syndrome. Even though the complexity of polycystic ovary syndrome and the challenge of reproducing all its features leave several questions unanswered, studies conducted to date have elucidated some of the alterations in ovarian follicle molecular and cellular mechanisms involved in the syndrome, and do not require the employment of complex and invasive techniques on human patients. This review examines ovarian functions and their alterations in polycystic ovary syndrome, explores preclinical in vivo and in vitro models, and highlights emerging research and medical perspectives. It targets researchers, healthcare professionals, and academics, including endocrinologists, cell biologists, and reproductive medicine specialists, studying the molecular and cellular mechanisms of the syndrome.

PMID:39840999 | DOI:10.1002/adbi.202400713

Protein Sci. 2025 Feb;34(2):e70028. doi: 10.1002/pro.70028.

ABSTRACT

Organisms from all kingdoms of life depend on Late Embryogenesis Abundant (LEA) proteins to survive desiccation. LEA proteins are divided into broad families distinguished by the presence of family-specific motif sequences. The LEA_4 family, characterized by 11-residue motifs, plays a crucial role in the desiccation tolerance of numerous species. However, the role of these motifs in the function of LEA_4 proteins is unclear, with some studies finding that they recapitulate the function of full-length LEA_4 proteins in vivo, and other studies finding the opposite result. In this study, we characterize the ability of LEA_4 motifs to protect a desiccation-sensitive enzyme, citrate synthase (CS), from loss of function during desiccation. We show here that LEA_4 motifs not only prevent the loss of function of CS during desiccation but also that they can do so more robustly via synergistically interactions with cosolutes. Our analysis further suggests that cosolutes induce synergy with LEA_4 motifs in a manner that correlates with transfer free energy. This research advances our understanding of LEA_4 proteins by demonstrating that during desiccation their motifs can protect specific clients to varying degrees and that their protective capacity is modulated by their chemical environment. Our findings extend beyond the realm of desiccation tolerance, offering insights into the interplay between IDPs and cosolutes. By investigating the function of LEA_4 motifs, we highlight broader strategies for understanding protein stability and function.

PMID:39840786 | DOI:10.1002/pro.70028

Sci Prog. 2025 Jan-Mar;108(1):368504241301535. doi: 10.1177/00368504241301535.

ABSTRACT

The recent severe acute respiratory syndrome coronavirus 2 pandemic has clearly exemplified the need for broad-spectrum antiviral (BSA) medications. However, previous outbreaks show that about one year after an outbreak, interest in antiviral research diminishes and the work toward an effective medication is left unfinished. Martin et al. endeavored to support the early stages of focused BSA development by creating the Small Molecule Antiviral Compound Collection (SMACC), which is publicly available online at https://smacc.mml.unc.edu. SMACC is a highly curated database with over 32,500 entries of chemical compounds tested in both phenotypic and target-based assays across 13 viruses from the NIAID's list of emerging infectious diseases/pathogens. The authors advise judicious use of knowledge collections such as SMACC and recommend users critically evaluate retrieved data and resulting hypotheses prior to experimental testing. When used correctly, SMACC-like databases may serve as a reference for medicinal chemists and virologists working to develop novel BSA medications. To summarize, we emphasize the importance of data curation for both novel outbreak prediction and development of BSAs against these outbreaks.

PMID:39840476 | DOI:10.1177/00368504241301535

Alzheimers Dement (N Y). 2025 Jan 21;11(1):e70037. doi: 10.1002/trc2.70037. eCollection 2025 Jan-Mar.

ABSTRACT

ABSTRACT: Recent clinical trials on slowing dementia progression have led to renewed focus on finding safer, more effective treatments. One approach to identify plausible candidates is to assess whether existing medications for other conditions may affect dementia risk. We conducted a systematic review to identify studies adopting a data-driven approach to investigate the association between a wide range of prescribed medications and dementia risk. We included 14 studies using administrative or medical records data for more than 130 million individuals and 1 million dementia cases. Despite inconsistencies in identifying specific drugs that may modify Alzheimer's or dementia risk, some themes emerged for drug classes with biological plausibility. Antimicrobials, vaccinations, and anti-inflammatories were associated with reduced risk, while diabetes drugs, vitamins and supplements, and antipsychotics were associated with increased risk. We found conflicting evidence for antihypertensives and antidepressants. Drug repurposing for use in dementia is an urgent priority. Our findings offer a basis for prioritizing candidates and exploring underlying mechanisms.

HIGHLIGHTS: ·We present a systematic review of studies reporting association between drugs prescribed for other conditions and risk of dementia including 139 million people and 1 million cases of dementia.·Our work supports some previously reported associations, for example, showing decreased risk of dementia with drugs to treat inflammatory disease and increased risk with antipsychotic treatment.·Antimicrobial treatment was perhaps more surprisingly associated with decreased risk, supportive of recent increased interest in this potential therapeutic avenue.·Our work should help prioritize drugs for entry into adaptive platform trials in Alzheimer's disease and will serve as a useful resource for those investigating drugs or classes of drugs and risk of dementia.

PMID:39839078 | PMC:PMC11747987 | DOI:10.1002/trc2.70037

Clin Pract Epidemiol Ment Health. 2024 Dec 3;20:e17450179337258. doi: 10.2174/0117450179337258241031035148. eCollection 2024.

ABSTRACT

BACKGROUND: Many pharmacological treatments are considered effective in the treatment of panic disorder (PD), however, about 20 to 40% of the patients have treatment-resistant PD. Pharmacogenetics could explain why some patients are treatment-resistant.

OBJECTIVE: Our objective was to gather preliminary data on the clinical usefulness of pharmacogenetic testing in this disorder.

METHODS: Twenty patients with treatment-resistant PD were included in this observational study and submitted to commercial pharmacogenetic testing. Testing panel included gene polymorphisms related to CYP, genes EPHX1, UGT1A4, UGT2B15, ABCB1, ADRA2A, ANKK1, COMT, DRD2, FKBP5, GRIK4, GSK3B, HTR1A, HTR2A, HTR2C, MC4R, OPRM1, SCN1A, SLC6A4 and MTHFR. Participants received treatment-as-usual for PD before being enrolled in this study, including first-line and second-line medications for PD.

RESULTS: In 30% of the patients, the tests indicated reduced chance of response to the prescribed drug, while they indicated very low serum levels of the prescribed drug in 20% of the subjects. The pharmacogenetic tests predicted reduction of MTHFR enzyme activity in 74% of the patients. ABCB1 gene alleles associated to drug resistance were found in 90% of the samples.

CONCLUSION: Commercial pharmacogenetic testing failed to predict negative treatment outcome in most patients with PD. The association between treatment-resistance in PD and the genes CYP2C19, MTHFR and ABCB1 deserves further study.

PMID:39839219 | PMC:PMC11748058 | DOI:10.2174/0117450179337258241031035148

Front Psychiatry. 2025 Jan 7;15:1497565. doi: 10.3389/fpsyt.2024.1497565. eCollection 2024.

ABSTRACT

The application of personalized medicine in patients with first-episode psychosis (FEP) requires tools for classifying patients according to their response to treatment, considering both treatment efficacy and toxicity. However, several limitations have hindered its translation into clinical practice. Here, we describe the rationale, aims and methodology of Applied Pharmacogenetics to Predict Response to Treatment of First Psychotic Episode (the FarmaPRED-PEP project), which aims to develop and validate predictive algorithms to classify FEP patients according to their response to antipsychotics, thereby allowing the most appropriate treatment strategy to be selected. These predictors will integrate, through machine learning techniques, pharmacogenetic (measured as polygenic risk scores) and epigenetic data together with clinical, sociodemographic, environmental, and neuroanatomical data. To do this, the FarmaPRED-PEP project will use data from two already recruited cohorts: the PEPS cohort from the "Genotype-Phenotype Interaction and Environment. Application to a Predictive Model in First Psychotic Episodes" study (the PEPs study from the Spanish abbreviation) (N=335) and the PAFIP cohort from "Clinical Program on Early Phases of Psychosis" (PAFIP from the Spanish abbreviation) (N = 350). These cohorts will be used to create the predictor, which will then be validated in a new cohort, the FarmaPRED cohort (N = 300). The FarmaPRED-PEP project has been designed to overcome several of the limitations identified in pharmacogenetic studies in psychiatry: (1) the sample size; (2) the phenotype heterogeneity and its definition; (3) the complexity of the phenotype and (4) the gender perspective. The global reach of the FarmaPRED-PEP project is to facilitate the effective deployment of precision medicine in national health systems.

PMID:39839139 | PMC:PMC11747510 | DOI:10.3389/fpsyt.2024.1497565

Curr Hypertens Rev. 2025 Jan 20. doi: 10.2174/0115734021342501250107052350. Online ahead of print.

ABSTRACT

INTRODUCTION: Hypertension is a chronic medical state and a major determining factor for cardiovascular and renal diseases. Both genetic and non-genetic factors contribute to hypertensive conditions among individuals. The renin-angiotensin-aldosterone system (RAAS) is a major genetic target for the anti-hypertension approach.

PURPOSE OF THE STUDY: The majority of classical antihypertensive drugs were mainly focused on the RAAS signaling pathways. Though these antihypertensive drugs control blood pressure (BP), they have mild to severe life-threatening effects. Unrevealing effective hypertensive targets for BP management is essential. The effective targets could emerge either from RAAS-dependent or RAAS-independent pathways and/or through the cross-talks among them.

RESULTS: Analyzing the physiopathological mechanisms of hypertension has the benefit of understanding the interactions between these systems which helps in better understanding of drug targets and the importance of emergence of novel therapeutics.

CONCLUSION: This review is about the signaling pathways involved in hypertension pathogenesis and their cross-talks and it contributes to a better understanding of the etiology of hypertension.

PMID:39838689 | DOI:10.2174/0115734021342501250107052350

Recent Pat Biotechnol. 2025 Jan 20. doi: 10.2174/0118722083359334250116063638. Online ahead of print.

ABSTRACT

Personalized medicine is an evolving paradigm that aims to tailor therapeutic interventions to individual patient characteristics. With a growing understanding of the genetic, epigenetic, and molecular mechanisms underlying diseases, tailored therapies are becoming more feasible and effective. This review highlights the significant advancements in personalized medicine, focusing specifically on pharmacological strategies. The article explores the integration of genomics, transcriptomics, proteomics, and metabolomics in drug development and therapy optimization. Pharmacogenomics, the customization of drug therapy based on an individual's genetic makeup, receives particular emphasis. This leads to the identification of specific biomarkers that can predict therapeutic response, drug toxicity, and susceptibility to various diseases. Together with computational tools and artificial intelligence, these advancements contribute to tailored treatment plans for patients with conditions such as cancer, cardiovascular diseases, and neurological disorders. We also highlight the challenges and ethical considerations in implementing personalized medicine, such as data privacy, cost-effectiveness, and accessibility. We outline future prospects and ongoing research in this field, highlighting the importance of collaborative efforts between researchers, clinicians, pharmacists, and regulatory authorities.

PMID:39838664 | DOI:10.2174/0118722083359334250116063638

Front Immunol. 2025 Jan 7;15:1466802. doi: 10.3389/fimmu.2024.1466802. eCollection 2024.

ABSTRACT

BACKGROUND AND AIM: NK cells and NK-cell-derived cytokines were shown to regulate neutrophil activation in acute lung injury (ALI). However, the extent to which ALI regulates lung tissue-resident NK (trNK) activity and their molecular phenotypic alterations are not well defined. We aimed to assess the impact of 1,25-hydroxy-vitamin-D3 [1,125(OH)2D] on ALI clinical outcome in a mouse model and effects on lung trNK cell activations.

METHODS: Oleic acid (OA)-induced ALI in C57BL/6J mice and 1,25(OH)2D treatment 2×/2 weeks were performed. Lung tissue was harvested to assess alveolar I/II cell apoptosis and lung injury marker of Surfactant-Protein-D (SP-D). Pulmonary edema markers of epithelial sodium channel, cystic fibrosis transmembrane conductance regulator, and aquaporin 5 were assessed by RT-PCR. Lung trNK cells were assessed for activation markers of CD107a and NKp46, vitamin D receptor (VDR), and programmed cell death protein-1 (PD-1) via flow cytometry. The bronchoalveolar lavage fluid (BALF) obtained was investigated for soluble receptor for advanced glycation end products (sRAGE), inflammatory cytokines, soluble 1,25(OH)2D, and PDL-1. Naïve mice treated with DMSO (vehicle) were used as a control.

RESULTS: Flow cytometry analysis displayed a high apoptotic rate in alveolar I/II cells of threefold in ALI mice as compared to naïve mice. These findings were accompanied by elevated markers of pulmonary edema as well as lung injury markers of SP-D. Isolated lung trNK cells of the ALI mice exhibited reduced CD107a and NKp46 markers and cytotoxicity potentials and were correlated through significantly 2.1-fold higher levels of PD-1 and diminished VDR expressions as compared to naïve mice. BALF samples of ALI mice displayed high soluble PDL-1 and reduced soluble 1,25(OH)2D levels compared to naïve mice. 1,25(OH)2D treatment alongside OA led to a significant fourfold increase in the CD107a and NKp46 expressions to levels higher than the mice treated with the vehicle. Furthermore, 1,25(OH)2D ameliorates free radical scavengers of GSH, GPX, CAT, and GPx-1; decreased pro-inflammatory cytokines and soluble PDL-1; and increased soluble 1,25(OH)2D with amelioration in pulmonary edema markers and alveolar I/II apoptosis.

CONCLUSION: Our results indicate 1,25(OH)2D's potential therapeutic effect in preventing clinical outcomes associated with ALI via regulating NK cells through inhibiting inflammatory cytokines and alleviating levels of PDL-1 and 1,25(OH)2D released by lung tissue.

PMID:39840066 | PMC:PMC11746039 | DOI:10.3389/fimmu.2024.1466802

MDM Policy Pract. 2025 Jan 17;10(1):23814683241305106. doi: 10.1177/23814683241305106. eCollection 2025 Jan-Jun.

ABSTRACT

Introduction. Diminishing marginal lifespan utility (DMLU) implies that a particular lifespan increment (e.g., 1 life-year) confers lesser marginal utility if added to longer lifespans (e.g., 90 y to 91 y) than to shorter lifespans (e.g., 60 y to 61 y) if quality of life is unchanged. Because DMLU is difficult to disambiguate from discounting, risk attitude, and other elements of utility "curvature," it is poorly characterized. However, the imperative to consider equity in cost-effectiveness analysis (CEA) renders its characterization more important. Methods. I add certainty to the characterization of DMLU through literature review and illustrative example. The literature review synthesizes stated preference studies of utility curvature that exclude risk or probability. The example compares alternative valuations of approaches to reduce inequality in cystic fibrosis outcomes between US centers serving mostly White patients and centers serving mostly non-Black Hispanic patients, with versus without DMLU. Results. The existence of DMLU is likely, and empirical data support its relevance over typical CEA time horizons. The imperative to consider equity in CEA magnifies the importance of DMLU for several reasons. First, intergenerational CEAs require lower discount rates that are less likely to incidentally absorb DMLU. Second, DMLU is incompatible with the use of absolute measures of inequality aversion. Third, DMLU may bias the interpretation of relative measures of inequality aversion toward prioritarianism. Finally, not considering DMLU implicitly biases life-year-based metrics against equity. Conclusion. DMLU is likely to exist, can benefit from additional characterization, and may merit inclusion in CEA alongside discounting. Omitting consideration of DMLU will sometimes confer an antiequity bias and may affect the interpretation of CEAs incorporating inequality aversion.

HIGHLIGHTS: Diminishing marginal lifespan utility (DMLU) means that the value of extending lifespan may differ based on the duration of life already lived.DMLU is not typically considered in cost-effectiveness analyses.Not considering DMLU may bias cost-effectiveness analyses against equity.Not considering DMLU may reduce the accuracy of distributive cost-effectiveness analyses and other approaches to consider equity along with efficiency.

PMID:39839684 | PMC:PMC11748391 | DOI:10.1177/23814683241305106

Quant Imaging Med Surg. 2025 Jan 2;15(1):189-202. doi: 10.21037/qims-24-989. Epub 2024 Dec 30.

ABSTRACT

BACKGROUND: Clinical severity and progression of lung disease in cystic fibrosis (CF) are significantly influenced by the degree of lung inflammation. Non-invasive quantitative diagnostic tools are desirable to differentiate structural and inflammatory lung changes in order to help prevent chronic airway disease. This might also be helpful for the evaluation of longitudinal effects of novel therapeutics. Therefore, the present study assesses the quantification of inflammatory lung changes using positron emission tomography/magnetic resonance imaging (PET/MRI) of the lung in children and adolescents with CF and evaluates the possible impact of PET/MRI on individualized therapy management.

METHODS: This monocentric, retrospective cohort study included 19 PET/MRI of the lung performed between 2014 and 2021 in 11 patients (16±4.5 years, 8-22 years; 7 females). PET acquisition was performed at least 20 minutes after i.v. application of a weight-adjusted dose of fluor-18-fluorodeoxyglucose (18F-FDG) of 1 MBq/kgBW (mean effective dose, 1.3±0.4 mSv). Lesions of increased uptake were quantified based on standardized uptake values (SUV) and compared to background activity, liver and blood pool. Pulmonary changes were assessed using the established magnetic resonance imaging-CF (MR-CF) score and correlated to inflammatory lesions. Results were correlated to changes in therapy (initiation, modification or discontinuation of therapy after baseline-PET/MRI) based on the electronic medical records.

RESULTS: Uptake was highly increased in 5 cases, moderate in 4 cases, low in 7 cases, no uptake in 3 cases. Most MR-CF score points were assigned to peribronchitis (23%) and air trapping (23%). Metabolically increased lesions were mainly interpreted as consolidations (59%; P<0.001) and mucus plugging (19%, P=0.024). There was a decrease in mean number and volumes of inflammatory lesions (P=0.016 each) and MR-CF score (P=0.047) between baseline and follow-up. After PET/MRI, therapy changed in 18 cases (95%; new medication: 58%, n=11; termination of therapy: 16%, n=3; modification of therapy: 21%, n=4).

CONCLUSIONS: In selected cases, pulmonary FDG-PET/MRI can help guide therapeutic decision-making and provide complementary information on CF-related lung changes to conventional MRI at a low radiation exposure.

PMID:39838989 | PMC:PMC11744157 | DOI:10.21037/qims-24-989

Clin Oral Implants Res. 2025 Jan 22. doi: 10.1111/clr.14406. Online ahead of print.

ABSTRACT

OBJECTIVE: This study evaluated ResNet-50 and U-Net models for detecting and segmenting vertical misfit in dental implant crowns using periapical radiographic images.

METHODS: Periapical radiographs of dental implant crowns were classified by two experts based on the presence of vertical misfit (reference group). The misfit area was manually annotated in images exhibiting vertical misfit. The resulting datasets were utilized to train the ResNet-50 and U-Net deep learning models. Then, 70% of the images were allocated for training, while the remaining 30% were used for validation and testing. Five general dentists categorized the testing images as "misfit" or "fit." Inter-rater reliability with Cohen's kappa index and performance metrics were calculated. The average performance metrics of dentists and artificial intelligence (AI) were compared using the paired-samples t test.

RESULTS: A total of 638 radiographs were collected. The kappa values between dentists and AI ranged from 0.93 to 0.98, indicating perfect agreement. The ResNet-50 model achieved accuracy and precision of 92.7% and 87.5%, respectively, whereas dentists had a mean accuracy of 93.3% and precision of 89.6%. The sensitivity and specificity for AI were 90.3% and 93.8%, respectively, compared to 90.1% and 95.1% for dentists. The Dice coefficient yielded 88.9% for the ResNet-50 and 89.5% among the dentists. The U-Net algorithm produced a loss of 0.01 and an accuracy of 0.98. No significant difference was found between the average performance metrics of dentists and AI (p > 0.05).

CONCLUSION: AI can detect and segment vertical misfit of implant prosthetic crowns in periapical radiographs, comparable to clinician performance.

PMID:39840554 | DOI:10.1111/clr.14406

Adv Sci (Weinh). 2025 Jan 22:e2409835. doi: 10.1002/advs.202409835. Online ahead of print.

ABSTRACT

Leaky and structurally abnormal blood vessels and increased pressure in the tumor interstitium reduce the infiltration of CAR-T cells in solid tumors, including triple-negative breast cancer (TNBC). Furthermore, high burden of tumor cells may cause reduction of infiltrating CAR-T cells and their functional exhaustion. In this study, various effector-to-target (E:T) ratio experiments are established to model the treatment using CAR-T cells in leukemia (high E:T ratio) and solid tumor (low E:T ratio). It is found that the antitumor immune response is decreased in solid tumors with low E:T ratio. Furthermore, single cell sequencing is performed to investigate the functional exhaustion at a low ratio. It is revealed that the inhibition of mitophagy-mediated mitochondrial dysfunction diminished the antitumor efficacy of CAR-T-cell therapy. The mitophagy agonist BC1618 is screened via AI-deep learning and cytokine detection, in vivo and in vitro studies revealed that BC1618 significantly strengthened the antitumor response of CAR-T cells via improving mitophagy. Here, injection hydrogels are engineered for the controlled co-delivery of CAR-T cells and BC1618 that improves the treatment of TNBC. Local delivery of hydrogels creates an inflammatory and mitophagy-enhanced microenvironment at the tumor site, which stimulates the CAR-T cells proliferation, provides antitumor ability persistently, and improves the effect of treatment.

PMID:39840546 | DOI:10.1002/advs.202409835

Front Plant Sci. 2025 Jan 7;15:1503033. doi: 10.3389/fpls.2024.1503033. eCollection 2024.

ABSTRACT

INTRODUCTION: Accurate detection and recognition of tea bud images can drive advances in intelligent harvesting machinery for tea gardens and technology for tea bud pests and diseases. In order to realize the recognition and grading of tea buds in a complex multi-density tea garden environment.

METHODS: This paper proposes an improved YOLOv7 object detection algorithm, called YOLOv7-DWS, which focuses on improving the accuracy of tea recognition. First, we make a series of improvements to the YOLOv7 algorithm, including decouple head to replace the head of YOLOv7, to enhance the feature extraction ability of the model and optimize the class decision logic. The problem of simultaneous detection and classification of one-bud-one-leaf and one-bud-two-leaves of tea was solved. Secondly, a new loss function WiseIoU is proposed for the loss function in YOLOv7, which improves the accuracy of the model. Finally, we evaluate different attention mechanisms to enhance the model's focus on key features.

RESULTS AND DISCUSSION: The experimental results show that the improved YOLOv7 algorithm has significantly improved over the original algorithm in all evaluation indexes, especially in the R Tea (+6.2%) and mAP@0.5 (+7.7%). From the results, the algorithm in this paper helps to provide a new perspective and possibility for the field of tea image recognition.

PMID:39840356 | PMC:PMC11747160 | DOI:10.3389/fpls.2024.1503033

Front Plant Sci. 2025 Jan 8;15:1435301. doi: 10.3389/fpls.2024.1435301. eCollection 2024.

ABSTRACT

Smart farming is a hot research area for experts globally to fulfill the soaring demand for food. Automated approaches, based on convolutional neural networks (CNN), for crop disease identification, weed classification, and monitoring have substantially helped increase crop yields. Plant diseases and pests are posing a significant danger to the health of plants, thus causing a reduction in crop production. The cotton crop, is a major cash crop in Asian and African countries and is affected by different types of weeds leading to reduced yield. Weeds infestation starts with the germination of the crop, due to which diseases also invade the field. Therefore, proper monitoring of the cotton crop throughout the entire phases of crop development from sewing to ripening and reaping is extremely significant to identify the harmful and undesired weeds timely and efficiently so that proper measures can be taken to eradicate them. Most of the weeds and pests attack cotton plants at different stages of growth. Therefore, timely identification and classification of such weeds on virtue of their symptoms, apparent similarities, and effects can reduce the risk of yield loss. Weeds and pest infestation can be controlled through advanced digital gadgets like sensors and cameras which can provide a bulk of data to work with. Yet efficient management of this extraordinarily bulging agriculture data is a cardinal challenge for deep learning techniques too. In the given study, an approach based on deep CNN-based architecture is presented. This work covers identifying and classifying the cotton weeds efficiently alongside a comparison of other already existing CNN models like VGG-16, ResNet, DenseNet, and Xception Model. Experimental results indicate the accuracy of VGG-16, ResNet-101, DenseNet-121, XceptionNet as 95.4%, 97.1%, 96.9% and 96.1%, respectively. The proposed model achieved an accuracy of 98.3% outperforming other models.

PMID:39840351 | PMC:PMC11750437 | DOI:10.3389/fpls.2024.1435301

Front Genet. 2025 Jan 7;15:1494474. doi: 10.3389/fgene.2024.1494474. eCollection 2024.

ABSTRACT

Recent advancements in deep learning, particularly large language models (LLMs), made a significant impact on how researchers study microbiome and metagenomics data. Microbial protein and genomic sequences, like natural languages, form a language of life, enabling the adoption of LLMs to extract useful insights from complex microbial ecologies. In this paper, we review applications of deep learning and language models in analyzing microbiome and metagenomics data. We focus on problem formulations, necessary datasets, and the integration of language modeling techniques. We provide an extensive overview of protein/genomic language modeling and their contributions to microbiome studies. We also discuss applications such as novel viromics language modeling, biosynthetic gene cluster prediction, and knowledge integration for metagenomics studies.

PMID:39840283 | PMC:PMC11747409 | DOI:10.3389/fgene.2024.1494474