Sci Rep. 2025 Mar 20;15(1):9668. doi: 10.1038/s41598-025-93377-8.

ABSTRACT

Disease classification is important for understanding disease commonalities on both the phenotypical and molecular levels. Based on predicted disease mode of action (MOA) proteins, our algorithm PICMOA (Pan-disease Classification in Mode of Action Protein Space) classifies 3526 diseases across 20 clinically classified classifications (ICD10-CM major classifications). At the top level, all diseases can be classified into "infectious" and "non-infectious" diseases. Non-infectious diseases are classified into 9 classes. To demonstrate the validity of the classifications, for common pathways predicted based on MOA proteins, 77% of the top 10 most frequent pathways have literature evidence of association to their respective disease classes/subclasses. These results indicate that PICMOA will be useful for understanding common disease mechanisms and facilitating the development of drugs for a class of diseases, rather than a single disease. The MOA proteins, molecular functions, pathways for classes, and individual diseases are available at https://sites.gatech.edu/cssb/PICMOA/ .

PMID:40113859 | DOI:10.1038/s41598-025-93377-8

Nat Commun. 2025 Mar 20;16(1):2770. doi: 10.1038/s41467-025-58076-y.

ABSTRACT

Inorganic phosphate (Pi) has essential metabolic and structural roles in living organisms. The Pi exporter, XPR1/SLC53A1, is critical for cellular Pi homeostasis. When intercellular Pi is high, cells accumulate inositol pyrophosphate (1,5-InsP8), a signaling molecule required for XPR1 function. Inactivating XPR1 mutations lead to brain calcifications, causing neurological symptoms including movement disorders, psychosis, and dementia. Here, cryo-electron microscopy structures of dimeric XPR1 and functional characterization delineate the substrate translocation pathway and how InsP8 initiates Pi transport. Binding of InsP8 to XPR1, but not the related inositol polyphosphate InsP6, rigidifies the intracellular SPX domains, with InsP8 bridging the dimers and SPX and transmembrane domains. Locked in this state, the C-terminal tail is sequestered, revealing the entrance to the transport pathway, thus explaining the obligate roles of the SPX domain and InsP8. Together, these findings advance our understanding of XPR1 transport activity and expand opportunities for rationalizing disease mechanisms and therapeutic intervention.

PMID:40113814 | DOI:10.1038/s41467-025-58076-y

ACS Nano. 2025 Mar 20. doi: 10.1021/acsnano.4c15538. Online ahead of print.

ABSTRACT

Bacterial microcompartments (BMCs) are proteinaceous organelles that self-assemble into selectively permeable shells that encapsulate enzymatic cargo. BMCs enhance catalytic pathways by reducing crosstalk among metabolites, preventing harmful intermediates from leaking into the cytosol and increasing reaction efficiency via enzyme colocalization. The intrinsic properties of BMCs make them attractive for biotechnological engineering. However, in vivo expression methods for shell synthesis have significant drawbacks that limit the potential design space for these nanocompartments. Here, we describe the development of an efficient and rapid method for the in vitro assembly of BMC shells from their protein building blocks. Our method enables large-scale construction of BMC shells by utilizing urea as a chaotropic agent to control self-assembly and provides an approach for encapsulation of both biotic and abiotic cargo under a broad range of reaction conditions. We demonstrate an enhanced level of control over the assembly of BMC shells in vitro and expand the design parameter space for engineering BMC systems with specialized and enhanced catalytic properties.

PMID:40113598 | DOI:10.1021/acsnano.4c15538

Rev Cuid. 2024 Oct 11;15(3):e3612. doi: 10.15649/cuidarte.3612. eCollection 2024 Sep-Dec.

ABSTRACT

INTRODUCTION: Different non-pharmacological interventions have been studied to manage symptoms derived from chemotherapy, but their effectiveness is unknown.

OBJECTIVE: To describe non-pharmacological interventions for managing symptoms secondary to antineoplastic chemotherapy in adults.

MATERIALS AND METHODS: Systematic review of analytical experimental and observational studies (2021 to 2023). The studies were selected, and data was extracted in parallel. Discrepancies were resolved with a third reviewer. The risk of bias was assessed using the Risk of Bias (RoB) tool and The Newcastle-Ottawa Scale (NOS). The literature was synthesized descriptively based on prioritized outcomes.

RESULTS: The prioritized outcomes were neutropenia, pain, neuropathy, nausea, vomiting, alopecia, anorexia, and sleep disorders. Out of 7520 references found, 62 were included for analysis. Acupressure showed a possible effect in controlling symptoms such as nausea and vomiting. The intervention with cold on the scalp showed differences in the stages of alopecia severity. Other interventions showed heterogeneity.

DISCUSSION: Non-pharmacological interventions have been widely described in observational and experimental studies in the control of side effects of chemotherapy; however, there is homogeneity and a high risk of bias.

CONCLUSION: Acupressure, muscle massage, music therapy, foot baths, and other interventions have been studied for nausea, vomiting, sleep disorders, neutropenia, alopecia, anorexia, pain, and neuropathy as secondary symptoms prioritized by patients. It is necessary to standardize both the interventions and how measure the outcomes.

PMID:40115309 | PMC:PMC11922587 | DOI:10.15649/cuidarte.3612

J Chem Inf Model. 2025 Mar 20. doi: 10.1021/acs.jcim.5c00275. Online ahead of print.

ABSTRACT

Drug-induced osteotoxicity refers to the harmful effects certain drugs have on the skeletal system, posing significant safety risks. These toxic effects are a key concern in clinical practice, drug development, and environmental management. However, existing toxicity assessment models lack specialized data sets and algorithms for predicting osteotoxicity. In our study, we collected osteotoxic molecules and employed various large language models, including DeepSeek and ChatGPT, alongside traditional machine learning methods to predict their properties. Among these, the DeepSeek R1 and ChatGPT o3 models achieved ACC values of 0.87 and 0.88, respectively. Our results indicate that machine learning methods can assist in evaluating the impact of harmful substances on bone health during drug development, improving safety protocols, mitigating skeletal side effects, and enhancing treatment outcomes and public safety. Furthermore, it highlights the potential of large language models in predicting molecular toxicity and their significance in the fields of health and chemical sciences.

PMID:40114317 | DOI:10.1021/acs.jcim.5c00275

BMC Neurol. 2025 Mar 20;25(1):117. doi: 10.1186/s12883-025-04130-7.

ABSTRACT

BACKGROUND: The second most common cause of death and disability worldwide is stroke. Drug-related problems (DRPs) can arise during any step of the medication process, whether it involves prescribing, transcribing, dispensing, or administering drugs. The purpose of this study was to assess risk factors and associated DRPs in patients with stroke.

METHODS: A cross-sectional study was conducted involving patients who had been diagnosed with stroke for 3 months using a purposive sampling technique at Annapurna Hospital. Data on demographics, comorbidities, and medications were collected through patient medical records, medicine Cardex, and nursing notes. DRPs were identified and classified using the Hepler-Strand classification system. Medscape software was used to assess potential drug-drug interactions (pDDIs). Descriptive statistics, chi-square tests, and binary logistic regression were performed.

RESULTS: Among the 111 patients, the mean age was 58.72 ± 15.68 years. The majority of strokes were ischemic (68.5%), with the middle cerebral artery being the most commonly affected (24.3%). Males were more commonly affected (76.6%) than females (23.4%). Hypertension was the most prevalent comorbidity (61.3%), followed by diabetes mellitus (27.0%) and hyperlipidemia (21.6%). Hyperlipidemia was significantly associated with risk factors for ischemic stroke. The study found that 91.9% of stroke patients experienced DRPs, with pDDIs being the most common type (91.09%). The severity of pDDIs was predominantly categorized as "monitor closely" (73.2%). The use of more than 10 medications was a significant predictor for high-severity pDDIs.

CONCLUSION: The study concludes that polypharmacy is a significant predictor for high-severity pDDIs, highlighting the need for careful consideration when adding new medications to a patient's therapy. The high rate of pDDIs (91%) emphasizes the critical role of clinical pharmacists in identifying and mitigating these interactions to prevent further drug-related complications in stroke patients. Further research is needed to explore interventions to reduce DRPs.

CLINICAL TRIAL NUMBER: Not applicable.

PMID:40114139 | DOI:10.1186/s12883-025-04130-7

Nat Metab. 2025 Mar 20. doi: 10.1038/s42255-025-01252-7. Online ahead of print.

ABSTRACT

Antipsychotic drugs have severe metabolic side effects. Acute use can induce hypothermia, while chronic use often leads to weight gain and associated disorders. However, no treatment is currently available for drug-induced hypothermia, and weight control measures lack evidence for long-term effectiveness. Here we demonstrate that a glucagon-like peptide 2 analogue, teduglutide, effectively prevents olanzapine-induced hypothermia and weight gain, and restores glucose tolerance and insulin sensitivity in mice. Mechanistically, olanzapine suppresses prodynorphin-expressing neurons in the ventromedial hypothalamus (VMHPdyn neurons) via serotonin receptor 2C, while teduglutide activates the same neuron population. Selective ablation of VMHPdyn neurons mimics olanzapine-induced side effects. More importantly, chemogenetic activation of VMHPdyn neurons abolishes olanzapine-induced hypothermia and excessive weight gain, although the psychotropic effects remain intact. Together, our data show that VMHPdyn neurons are the crucial mediator of antipsychotic-induced metabolic dysfunction and glucagon-like peptide 2 receptor agonism may be an effective target to mitigate both acute and chronic side effects.

PMID:40114026 | DOI:10.1038/s42255-025-01252-7

NIH recently launched several enhancements to allow the public to more easily and quickly find funding information for various NIH research areas. The new look and feel of the NIH Research, Condition, and Disease Categorization (RCDC) Categorical Spending webpage adds to NIH’s long-standing efforts to enhance transparency and accountability into NIH funding decisions and the research areas NIH supports.

RCDC launched in 2008 as a tool within NIH’s Research Portfolio Online Reporting Tools (RePORT) suite. It provides estimates of annual support level for more than 300 research, condition, and disease categories based on grants, contracts, and other funding mechanisms used across the NIH, as well as disease burden data published by the CDC National Center for Health Statistics. More on the RCDC process is available on this post from 2018.

The new visual and contextual changes aim to improved usability and understanding of the RCDC categorization process. In particular, the categorical spending page was reorganized so data are more prominent and easier to navigate.

Navigation improvements simplify finding information on FAQs, the Categorization Process, and the biomedical thesaurus. The information in the data tables are more visible due to collapsed textual information  and frozen table headers (Figure 1).

When selecting a particular category, the line graph on the top left of the page will  automatically adjust to reflect funding amounts over time for that topic area (Figure 2, left image). Selecting support data for a given fiscal year will also reveal information about specific projects (Figure 2, right image).

Results can be narrowed further if interested in a particular NIH Institute or Center (Figure 3). To sort in this way, a user would need to click on the dollar amount for a given disease/research area and a given fiscal year in the main table.

Application identification numbers can now be exported for funded awards for any fiscal year in any given category. Previously RCDC only reported the project number for an award. Because application IDs are unique to an individual fiscal year, it is easier to now connect results from RCDC with those obtained in other RePORT tools.

Genome Res. 2025 Apr 14;35(4):914-928. doi: 10.1101/gr.279323.124.

ABSTRACT

Rare structural variants (SVs)-insertions, deletions, and complex rearrangements-can cause Mendelian disease, yet they remain difficult to accurately detect and interpret. We sequenced and analyzed Oxford Nanopore Technologies long-read genomes of 68 individuals from the undiagnosed disease network (UDN) with no previously identified diagnostic mutations from short-read sequencing. Using our optimized SV detection pipelines and 571 control long-read genomes, we detected 716 long-read rare (MAF < 0.01) SV alleles per genome on average, achieving a 2.4× increase from short reads. To characterize the functional effects of rare SVs, we assessed their relationship with gene expression from blood or fibroblasts from the same individuals and found that rare SVs overlapping enhancers were enriched (LOR = 0.46) near expression outliers. We also evaluated tandem repeat expansions (TREs) and found 14 rare TREs per genome; notably, these TREs were also enriched near overexpression outliers. To prioritize candidate functional SVs, we developed Watershed-SV, a probabilistic model that integrates expression data with SV-specific genomic annotations, which significantly outperforms baseline models that do not incorporate expression data. Watershed-SV identified a median of eight high-confidence functional SVs per UDN genome. Notably, this included compound heterozygous deletions in FAM177A1 shared by two siblings, which were likely causal for a rare neurodevelopmental disorder. Our observations demonstrate the promise of integrating long-read sequencing with gene expression toward improving the prioritization of functional SVs and TREs in rare disease patients.

PMID:40113264 | DOI:10.1101/gr.279323.124

Curr Opin Oncol. 2025 May 1;37(3):250-258. doi: 10.1097/CCO.0000000000001141. Epub 2025 Mar 19.

ABSTRACT

PURPOSE OF REVIEW: Rare cancers of the genitourinary (GU) tract are often clinically aggressive yet have few or no standard-of-care treatments. Multiple antibody-drug conjugates (ADCs) have been approved in solid malignancies. This review explores the use of ADCs in rare GU tumors in the context of biological pathways and ongoing research in solid tumors.

RECENT FINDINGS: Few clinical trials of ADCs focus on recruiting participants with rare tumors of the GU tract, including trials testing enfortumab vedotin as monotherapy or combined with pembrolizumab, and sacituzumab govitecan as monotherapy or combined with atezolizumab. We highlight many ongoing trials of novel ADCs for advanced/metastatic solid tumors and emphasize the potential eligibility of patients with rare GU tumors for tumor-agnostic trials.

SUMMARY: ADCs are being tested in multiple solid tumors, including rare GU tumors. Ongoing preclinical research supports the use of some ADCs in several rare GU tumors and improves our understanding of their pathophysiology.

PMID:40110990 | DOI:10.1097/CCO.0000000000001141

Med Image Anal. 2025 Mar 13;102:103539. doi: 10.1016/j.media.2025.103539. Online ahead of print.

ABSTRACT

Pancreatic cancer is a lethal invasive tumor with one of the worst prognosis. Accurate and reliable segmentation for pancreas and pancreatic cancer on computerized tomography (CT) images is vital in clinical diagnosis and treatment. Although certain deep learning-based techniques have been tentatively applied to this task, current performance of pancreatic cancer segmentation is far from meeting the clinical needs due to the tiny size, irregular shape and extremely uncertain boundary of the cancer. Besides, most of the existing studies are established on the black-box models which only learn the annotation distribution instead of the logical thinking and diagnostic experience of high-level medical experts, the latter is more credible and interpretable. To alleviate the above issues, we propose a novel Segment-Like-A-Doctor (SLAD) framework to learn the reliable clinical thinking and experience for pancreas and pancreatic cancer segmentation on CT images. Specifically, SLAD aims to simulate the essential logical thinking and experience of doctors in the progressive diagnostic stages of pancreatic cancer: organ, lesion and boundary stage. Firstly, in the organ stage, an Anatomy-aware Masked AutoEncoder (AMAE) is introduced to model the doctors' overall cognition for the anatomical distribution of abdominal organs on CT images by self-supervised pretraining. Secondly, in the lesion stage, a Causality-driven Graph Reasoning Module (CGRM) is designed to learn the global judgment of doctors for lesion detection by exploring topological feature difference between the causal lesion and the non-causal organ. Finally, in the boundary stage, a Diffusion-based Discrepancy Calibration Module (DDCM) is developed to fit the refined understanding of doctors for uncertain boundary of pancreatic cancer by inferring the ambiguous segmentation discrepancy based on the trustworthy lesion core. Experimental results on three independent datasets demonstrate that our approach boosts pancreatic cancer segmentation accuracy by 4%-9% compared with the state-of-the-art methods. Additionally, the tumor-vascular involvement analysis is also conducted to verify the superiority of our method in clinical applications. Our source codes will be publicly available at https://github.com/ZouLiwen-1999/SLAD.

PMID:40112510 | DOI:10.1016/j.media.2025.103539

Gene. 2025 Mar 18:149421. doi: 10.1016/j.gene.2025.149421. Online ahead of print.

ABSTRACT

BACKGROUND: Pharmacogenomics is used to identify genetic factors that influence drug responses, thereby optimizing therapeutic outcomes and reducing adverse effects. The objective of this study is to identify pharmacogenomic variations and their clinical relevance to drug metabolism and toxicity within the Pumi population.

METHODS: Eighty-two genetic variants in 43 genes were genotyped in 200 unrelated Pumi individuals using the Agena MassARRAY Assay. Chi-square tests, adjusted for multiple comparisons with Bonferroni correction, were used to compare genotype frequency divergences between the Pumi population and 26 other populations. Population genetic structure diversity and pairwise F-statistics (Fst) were assessed across 27 populations using Structure v2.3.1 and Arlequin v3.5 software.

RESULTS: After Bonferroni correction, a number of single nucleotide variations (SNVs) exhibited significant differences in frequency between the Pumi population and other populations. The allele frequencies of ADH1A rs975833, ADH1B rs1229984, TPMT rs1142345, and CYP2A6 rs8192726 in the Pumi population were notably different from the East Asian population or the other 26 populations. PharmGKB data indicate that rs1229984, rs1142345, and rs8192726 are associated with the metabolic efficiency of acetaldehyde, mercaptopurine, and efavirenz, respectively. Additionally, the genetic structure analysis (K = 5) and pairwise Fst calculations revealed that the Pumi population shared a similar genetic background with CHB (Fst = 0.031), JPT (Fst = 0.033), KHV (Fst = 0.035), CHS (Fst = 0.036), and CDX (Fst = 0.037) populations.

CONCLUSION: Our findings reveal unique genetic variations and biomarkers within the Pumi population, which contributes pharmacogenomic insights and theoretical foundations for personalized medicine tailored to the Pumi population.

PMID:40113188 | DOI:10.1016/j.gene.2025.149421

Arch Bronconeumol. 2025 Mar 1:S0300-2896(25)00071-7. doi: 10.1016/j.arbres.2025.02.010. Online ahead of print.

ABSTRACT

BACKGROUND: Cystic fibrosis causes exercise limitation due to impaired lung function and other complications, which in turn increases the chance of mortality. CFTR modulators, particularly the elexacaftor/tezacaftor/ivacaftor (ETI) combination, improve lung function in children older than 6 years in real-life studies.

OBJECTIVE: This study aimed to assess the impact of ETI on aerobic capacity in children with CF aged 6-11 years under real-life conditions and to evaluate whether prior CFTR modulator treatment affects these outcomes.

METHODS: A multicenter, prospective cohort study was conducted with pediatric CF patients. Participants underwent evaluations 6-8 months before ETI (T1), at the start of ETI (T2), and 6-8 months post-treatment (T3). Primary outcomes included cardiorespiratory fitness assessed via peak oxygen consumption (VO2peak) during a cardiopulmonary exercise test (CPET), and secondary outcomes encompassed lung function, quality of life, physical activity, and functional mobility.

RESULTS: Of the 28 patients (mean age 9.02±1.59 years), 19 were ETI-naive, and 9 had prior CFTR modulator treatment. Significant improvements were observed in FEV1 (p<0.001), and several functional mobility tests (30CST, Stair Climb Test, 10MWT). However, VO2peak showed no significant changes between T1 and T3. Quality of life scores improved notably in emotional, eating, and respiratory domains, and a slight improvement was noted in physical activity levels (p=0.037).

CONCLUSIONS: ETI treatment significantly enhances lung function and certain aspects of quality of life and physical fitness in pediatric CF patients. However, it does not significantly alter aerobic capacity (VO2peak) within the observed period.

PMID:40113488 | DOI:10.1016/j.arbres.2025.02.010

Diagn Interv Imaging. 2025 Mar 19:S2211-5684(25)00048-8. doi: 10.1016/j.diii.2025.03.004. Online ahead of print.

ABSTRACT

PURPOSE: The purpose of this study was to validate a deep learning algorithm that generates T2*-weighted images from diffusion-weighted (DW) images and to compare its performance with that of true T2*-weighted images for hemorrhage detection on MRI in patients with acute stroke.

MATERIALS AND METHODS: This single-center, retrospective study included DW- and T2*-weighted images obtained less than 48 hours after symptom onset in consecutive patients admitted for acute stroke. Datasets were divided into training (60 %), validation (20 %), and test (20 %) sets, with stratification by stroke type (hemorrhagic/ischemic). A generative adversarial network was trained to produce generative T2*-weighted images using DW images. Concordance between true T2*-weighted images and generative T2*-weighted images for hemorrhage detection was independently graded by two readers into three categories (parenchymal hematoma, hemorrhagic infarct or no hemorrhage), and discordances were resolved by consensus reading. Sensitivity, specificity and accuracy of generative T2*-weighted images were estimated using true T2*-weighted images as the standard of reference.

RESULTS: A total of 1491 MRI sets from 939 patients (487 women, 452 men) with a median age of 71 years (first quartile, 57; third quartile, 81; range: 21-101) were included. In the test set (n = 300), there were no differences between true T2*-weighted images and generative T2*-weighted images for intraobserver reproducibility (κ = 0.97 [95 % CI: 0.95-0.99] vs. 0.95 [95 % CI: 0.92-0.97]; P = 0.27) and interobserver reproducibility (κ = 0.93 [95 % CI: 0.90-0.97] vs. 0.92 [95 % CI: 0.88-0.96]; P = 0.64). After consensus reading, concordance between true T2*-weighted images and generative T2*-weighted images was excellent (κ = 0.92; 95 % CI: 0.91-0.96). Generative T2*-weighted images achieved 90 % sensitivity (73/81; 95 % CI: 81-96), 97 % specificity (213/219; 95 % CI: 94-99) and 95 % accuracy (286/300; 95 % CI: 92-97) for the diagnosis of any cerebral hemorrhage (hemorrhagic infarct or parenchymal hemorrhage).

CONCLUSION: Generative T2*-weighted images and true T2*-weighted images have non-different diagnostic performances for hemorrhage detection in patients with acute stroke and may be used to shorten MRI protocols.

PMID:40113490 | DOI:10.1016/j.diii.2025.03.004

J Orthop Res. 2025 Mar 20. doi: 10.1002/jor.26071. Online ahead of print.

ABSTRACT

Articular cartilage, essential for smooth joint movement, can sustain micrometer-scale microcracks in its collagen network from low-energy impacts previously considered non-injurious. These microcracks may propagate under cyclic loading, impairing cartilage function and potentially initiating osteoarthritis (OA). Detecting and analyzing microcracks is crucial for understanding early cartilage damage but traditionally relies on manual analyses of second harmonic generation (SHG) images, which are labor-intensive, limit scalability, and delay insights. To address these challenges, we established and validated a YOLOv8-based deep learning model to automate the detection, segmentation, and quantification of cartilage microcracks from SHG images. Data augmentation during training improved model robustness, while evaluation metrics, including precision, recall, and F1-score, confirmed high accuracy and reliability, achieving a true positive rate of 95%. Our model consistently outperformed human annotators, demonstrating superior accuracy, repeatability, all while reducing labor demands. Error analyses indicated precise predictions for microcrack length and width, with moderate variability in estimations of orientation. Our results demonstrate the transformative potential of deep learning in cartilage research, enabling large-scale studies, accelerating analyses, and providing insights into soft tissue damage and engineered material mechanics. Expanding our data set to include diverse anatomical regions and disease stages will further enhance performance and generalization of our YOLOv8-based model. By automating microcrack detection, this study advances understanding of microdamage in cartilage and potential mechanisms of progression of OA. Our publicly available model and data set empower researchers to develop personalized therapies and preventive strategies, ultimately advancing joint health and preserving quality of life.

PMID:40113341 | DOI:10.1002/jor.26071

Environ Pollut. 2025 Mar 18:126083. doi: 10.1016/j.envpol.2025.126083. Online ahead of print.

ABSTRACT

The rapid and accurate identification of antibiotics in environmental samples is critical for addressing the growing concern of antibiotic pollution, particularly in water sources. Antibiotic contamination poses a significant risk to ecosystems and human health by contributing to the spread of antibiotic resistance. SERS, known for its high sensitivity and specificity, is a powerful tool for antibiotic identification. However, its broader application is constrained by the lack of a large-scale antibiotic spectral database crucial for environmental and clinical use. To address this need, we systematically collected 12,800 SERS spectra for 200 environmentally relevant antibiotics and developed an open-access, web-based database at http://sers.test.bniu.net/. We compared six machine learning algorithms with a CNN model, which achieved the highest accuracy at 98.94%, making it the preferred database model. For external validation, CNN demonstrated an accuracy of 82.8%, underscoring its reliability and practicality for real-world applications. The SERS database and CNN prediction model represent a novel resource for environmental monitoring, offering significant advantages in terms of accessibility, speed, and scalability. This study establishes the large-scale, public SERS spectral databases for antibiotics, facilitating the integration of SERS into environmental programs, with the potential to improve antibiotic detection, pollution management, and resistance mitigation.

PMID:40113206 | DOI:10.1016/j.envpol.2025.126083

Cell Syst. 2025 Mar 19;16(3):101229. doi: 10.1016/j.cels.2025.101229.

ABSTRACT

The three-dimensional (3D) morphology of cells emerges from complex cellular and environmental interactions, serving as an indicator of cell state and function. In this study, we used deep learning to discover morphology representations and understand cell states. This study introduced MorphoMIL, a computational pipeline combining geometric deep learning and attention-based multiple-instance learning to profile 3D cell and nuclear shapes. We used 3D point-cloud input and captured morphological signatures at single-cell and population levels, accounting for phenotypic heterogeneity. We applied these methods to over 95,000 melanoma cells treated with clinically relevant and cytoskeleton-modulating chemical and genetic perturbations. The pipeline accurately predicted drug perturbations and cell states. Our framework revealed subtle morphological changes associated with perturbations, key shapes correlating with signaling activity, and interpretable insights into cell-state heterogeneity. MorphoMIL demonstrated superior performance and generalized across diverse datasets, paving the way for scalable, high-throughput morphological profiling in drug discovery. A record of this paper's transparent peer review process is included in the supplemental information.

PMID:40112779 | DOI:10.1016/j.cels.2025.101229

Cell Syst. 2025 Mar 19;16(3):101242. doi: 10.1016/j.cels.2025.101242.

ABSTRACT

One snapshot of the peer review process for "Geometric deep learning and multiple instance learning for 3D cell shape profiling" (De Vries et al., 2025).1.

PMID:40112776 | DOI:10.1016/j.cels.2025.101242

EBioMedicine. 2025 Mar 19;114:105657. doi: 10.1016/j.ebiom.2025.105657. Online ahead of print.

ABSTRACT

BACKGROUND: Heart failure (HF) is a complex syndrome with varied presentations and progression patterns. Traditional classification systems based on left ventricular ejection fraction (LVEF) have limitations in capturing the heterogeneity of HF. We aimed to explore the application of deep learning, specifically a Transformer-based approach, to analyse electronic health records (EHR) for a refined subtyping of patients with HF.

METHODS: We utilised linked EHR from primary and secondary care, sourced from the Clinical Practice Research Datalink (CPRD) Aurum, which encompassed health data of over 30 million individuals in the UK. Individuals aged 35 and above with incident reports of HF between January 1, 2005, and January 1, 2018, were included. We proposed a Transformer-based approach to cluster patients based on all clinical diagnoses, procedures, and medication records in EHR. Statistical machine learning (ML) methods were used for comparative benchmarking. The models were trained on a derivation cohort and assessed for their ability to delineate distinct clusters and prognostic value by comparing one-year all-cause mortality and HF hospitalisation rates among the identified subgroups in a separate validation cohort. Association analyses were conducted to elucidate the clinical characteristics of the derived clusters.

FINDINGS: A total of 379,108 patients were included in the HF subtyping analysis. The Transformer-based approach outperformed alternative methods, delineating more distinct and prognostically valuable clusters. This approach identified seven unique HF patient clusters characterised by differing patterns of mortality, hospitalisation, and comorbidities. These clusters were labelled based on the dominant clinical features present at the initial diagnosis of HF: early-onset, hypertension, ischaemic heart disease, metabolic problems, chronic obstructive pulmonary disease (COPD), thyroid dysfunction, and late-onset clusters. The Transformer-based subtyping approach successfully captured the multifaceted nature of HF.

INTERPRETATION: This study identified seven distinct subtypes, including COPD-related and thyroid dysfunction-related subgroups, which are two high-risk subgroups not recognised in previous subtyping analyses. These insights lay the groundwork for further investigations into tailored and effective management strategies for HF.

FUNDING: British Heart Foundation, European Union - Horizon Europe, and Novo Nordisk Research Centre Oxford.

PMID:40112740 | DOI:10.1016/j.ebiom.2025.105657

Food Chem. 2025 Mar 15;480:143886. doi: 10.1016/j.foodchem.2025.143886. Online ahead of print.

ABSTRACT

In this study, an innovative intelligent system for supervising the quality of fresh produce was proposed, which combined 3D printing technology and deep convolutional neural networks (DCNN). Through 3D printing technology, sensitive, lightweight, and customizable dual-color CO2 monitoring labels were fabricated using bromothymol blue and methyl red as indicators. These labels were applied to sensitively monitor changes in CO2 levels during the storage of vegetables such as green vegetables, cucumbers, okras, plums, and jujubes. The ΔE of the labels was found to have a significant positive correlation with CO2 levels and weight loss rate, while showing a strong inverse relationship with hardness, indirectly reflecting the freshness of the produce. In addition, four lightweight DCNN models (GhostNet, MobileNetv2, ShuffleNet, and Xception) were applied to recognize label images from different storage days, with MobileNetv2 achieving the best performance. The classification accuracy for three freshness levels of okra was 96.06 %, 91.12 %, and 93.86 %, respectively. A mobile application was developed based on this model, which demonstrated excellent performance in recognizing labels at different storage stages, making it suitable for practical applications and effectively distinguishing freshness levels. By combining the novel labels with advanced DCNN models, the accuracy and real-time capabilities of food monitoring can be significantly improved.

PMID:40112721 | DOI:10.1016/j.foodchem.2025.143886