Mol Oncol. 2025 Aug 11. doi: 10.1002/1878-0261.70046. Online ahead of print.

ABSTRACT

Acute lymphoblastic leukemia (ALL), the most common cancer in children, is overall divided into two subtypes, B-cell precursor ALL (B-ALL) and T-cell ALL (T-ALL), which have different molecular characteristics. Despite massive progress in understanding the disease trajectories of ALL, ALL remains a major cause of death in children. Thus, further research exploring the biological foundations of ALL is essential. Here, we examined the diagnostic, prognostic, and therapeutic potential of gene expression data in pediatric patients with ALL. We discovered a subset of expression markers differentiating B- and T-ALL: CCN2, VPREB3, NDST3, EBF1, RN7SKP185, RN7SKP291, SNORA73B, RN7SKP255, SNORA74A, RN7SKP48, RN7SKP80, LINC00114, a novel gene (ENSG00000227706), and 7SK. The expression level of these markers all demonstrated significant effects on patient survival, comparing the two subtypes. We also discovered four expression subgroups in the expression data with eight genes driving separation between two of these predicted subgroups. A subset of the 14 markers could distinguish B- and T-ALL in an independent cohort of patients with ALL. This study can enhance our knowledge of the transcriptomic profile of different ALL subtypes.

PMID:40788820 | DOI:10.1002/1878-0261.70046

Indian J Cancer. 2025 Apr 1;62(2):213-219. doi: 10.4103/ijc.ijc_24_23. Epub 2025 Aug 8.

ABSTRACT

BACKGROUND: Cancer patients are administered various chemotherapeutic agents along with supportive management, which increases the risk of potential drug-drug interactions (pDDIs). We aimed to assess the pDDIs in In-patient Department (IPD) patients receiving cancer chemotherapy in a tertiary care hospital.

METHODS: A prospective, observational study was conducted in the oncology department of a tertiary care hospital for a period of 6 months. Patient information was noted in the data collection form, and pDDIs were assessed using the Micromedex® database. Mann-Whitney U, Chi-square, and Spearman's correlation were the tests used for statistical analysis. P value of less than 0.05 was considered statistically significant.

RESULTS: In a sample size of 145 patients having a confidence interval of 95% and response distribution of 50%, the margin of error was found to be 8.11%. Male predominance (57.2%) was seen in the study. Although the adult patient population (56.5%) dominated the study, pediatric (26.9%) and geriatric (16.6%) patients were also included. A total of 115 pDDIs were found in 41% of the total patient population, out of which 56% (n = 64) were moderate and 44% (n = 51) were major in severity. The number of drug interactions was found to have correlation with the number of drugs (rho = 0.2, P = 0.01) prescribed during hospital stay.

CONCLUSION: The present study shows that cancer patients are relatively at risk for drug-drug interactions. To avoid Adverse Drug Reaction (ADRs), harmful effects, and other undesirable clinical manifestations shown by drug-drug interactions, screening for pDDIs is required.

PMID:40788735 | DOI:10.4103/ijc.ijc_24_23

Naunyn Schmiedebergs Arch Pharmacol. 2025 Aug 11. doi: 10.1007/s00210-025-04508-x. Online ahead of print.

ABSTRACT

Drug repurposing (DR) is a strategy to develop novel drugs from pre-approved drugs. To overcome the limitations of conventional drug development pathways, DR is a novel method that is cost-effective, with minimal side effects, and less time-consuming. Contrary to de novo drug discovery, DR eliminates the need for FDA approval. Fluoroquinolones (FQs) are potentially antibiotics, but recently, it has been discovered that FQs are a potential "treasure trove" to circumvent cancer drug resistance. Various in vitro studies on FQs showed their anticancer functionality against various cancer cell lines. Moreover, FQs are found to enhance the therapeutic efficacy of various clinical drugs by acting synergistically or additively in combination therapy. FQs such as ciprofloxacin, moxifloxacin, and levofloxacin are found to enhance the anti-tumor effect of a wide range of FDA-approved chemotherapeutics through multiple mechanisms including cell cycle arrest, apoptosis induction, anti-proliferation, and modulation of EMT. This review will comprehensively focus on the chemosensitization mechanism and cellular targets of FQs. This study provides a scientific basis to explore FQs as potential chemotherapeutic agents for combinational therapy to tackle the current scenario of drug resistance.

PMID:40788483 | DOI:10.1007/s00210-025-04508-x

Nat Prod Bioprospect. 2025 Aug 11;15(1):52. doi: 10.1007/s13659-025-00533-8.

ABSTRACT

Tuberculosis (TB) is a leading infectious disease killer and one of the major causes of deaths worldwide. Although TB is a curable and preventable disease, in 2023, approximately 10.8 million people fell ill with TB and there were an estimated 1.25 million of deaths worldwide. Despite some research progress for new drug candidates, drug repurposing, and new regimens, there is still an urgent need for the new medicins to treat TB, especially due to the growing cases of multidrug and extensively drug-resistant (MDR/XDR) strains. Drug resistance is a challenging obstacle to TB care and prevention globally, making TB harder and longer to treat, often with poorer outcomes for patients. The Actinomycetota encompass Gram-positive bacteria that produce a milieu of bioactive metabolites, including antibiotics, antiproliferative drugs, immunosuppressive agents, and other important medical molecules. Actinomycetota have a special place in the therapeutic arsenal to fight TB, as rifamycins, aminoglycosides, and cycloserine are derived from Streptomyces species, one of the most important genera in this phylum. Furthermore, hundreds of antimycobacterial metabolites have been isolated from Actinomycetota and can serve as effective drugs or useful agents for the discovery of new lead compounds to combat TB. The present review covers more than 171 isolated substances as potential antimycobacterial agents discovered between the years 1972 to 2024. Among the most potent compounds, with MIC in the submicromolar range, steffimycins, ilamycins/rufomycins, nosiheptide, actinomycins, lassomycin and boromycin are the most promising compounds. These compounds represent highly promising candidates for development of new antitubercular drugs. Additionally, some of these substances also demonstrated activity against resistant Mycobacterium tuberculosis (Mtb) strains, which is particularly relevant given the difficulty of treating MDR and XDR strains. Thus, actinobacteria have played and continue to play an important role in fight TB, remaining a promising source of antibiotic metabolites. Their unique metabolic diversity enables the production of metabolites with innovative mechanisms of action, making them a strategic reservoir for discovering therapies against untreatable forms of the disease.

PMID:40788464 | DOI:10.1007/s13659-025-00533-8

F1000Res. 2024 Dec 20;13:1546. doi: 10.12688/f1000research.157890.1. eCollection 2024.

ABSTRACT

BACKGROUND: Cerebral small vessel disease (cSVD) is a progressive neurovascular-degenerative condition without specific treatment that causes lacunar stroke, most intracerebral haemorrhage, vascular cognitive impairment (VCI) and several neuropsychiatric conditions.

OBJECTIVES: To conduct a rapid multi-stage scoping review to identify licensed interventions that could be repurposed for testing in cSVD at phase-3.

METHODS: First, we screened preclinical studies of potential relevance to cSVD and used a drug dictionary to identify studies of potential interventions. Separately, we screened clinical studies of relevance to cSVD and VCI. Following merging, we removed drugs that were unsuitable or impractical to assess long-term in the UK. We then performed mini-meta-analyses for shortlisted interventions assessing effects on cognition and scored these for their relevance to cSVD.

RESULTS: The preclinical review created a long-list of 1,757 deduplicated interventions. Those that were not available in the UK, not expensive or impractical to administer long-term were merged with 62 interventions identified from 75 relevant clinical studies to create a medium-list of 52 interventions. Focussed literature review short-listed ten interventions for review by an independent scientific advisory group; they ranked three as most suitable for immediate testing: metformin, tadalafil and isosorbide mononitrate.

CONCLUSION: This rapid review identified three interventions that are suitable for testing in a late phase-3 (platform) trial involving patients with cSVD. The approach could be improved with partial automation, text mining and generative pre-trained transformer approaches which would help manage the large data volumes. Further, our data-driven approach could be combined with genetic or other mechanistic methods to further de-risk future trials.

PMID:40786096 | PMC:PMC12335727 | DOI:10.12688/f1000research.157890.1

Front Pharmacol. 2025 Jul 25;16:1631419. doi: 10.3389/fphar.2025.1631419. eCollection 2025.

ABSTRACT

BACKGROUND: Cancer drug resistance significantly reduces the effectiveness of current anticancer treatments. Multiple dysregulated signaling pathways drive cancer initiation, progression, and related drug resistance. This highlights the need for developing new multi-targeting drugs that are more cost-effective, have fewer side effects, and remain effective against cancer. Drug repurposing offers a promising solution to expensive targeted therapies and helps overcome drug resistance. Mebendazole (MBZ), albendazole, flubendazole, and oxfendazole are broad-spectrum anti-helminthic drugs from the benzimidazole family.

PURPOSE: Therefore, MBZ demonstrated potential in suppressing the growth of various cancer cells, both in vitro and in vivo. Consequently, we thoroughly reviewed MBZ as a therapeutic option against cancer and related drug resistance.

RESULTS AND DISCUSSION: In this study, we identified MBZ as a promising cancer treatment that works through multiple mechanisms such as regulating tumor angiogenesis, autophagy, and apoptosis, modulating key signaling pathways, boosting antitumor immune responses, and inhibiting matrix metalloproteinases activity-all of which are major factors in cancer drug resistance. Additionally, the development of new MBZ delivery systems aims to address its pharmacokinetic limitations. While the anticancer effects of MBZ are encouraging, further research is needed before it can be used clinically.

CONCLUSION: Extensive data from in vitro, in vivo, and clinical trials support MBZ's anticancer potential and highlight the need for innovative delivery methods, including polymeric nanoparticles, nanostructured lipid formulations, micelles, nanosuspensions, and beyond.

PMID:40786044 | PMC:PMC12331676 | DOI:10.3389/fphar.2025.1631419

Front Pharmacol. 2025 Jul 25;16:1592011. doi: 10.3389/fphar.2025.1592011. eCollection 2025.

ABSTRACT

We describe a drug repurposing treatment involving the use of nicardipine in a young patient with Pitt-Hopkins syndrome (a rare neurodevelopmental disorder that results from variants of TCF4 gene) as a bench-to-bedside approach. Loss of TCF4 function in Pitt-Hopkins syndrome leads to increased excitability of Nav1.8 in neurons. Nicardipine is normally used alone or together with other medicines to treat severe chest pain (angina) or high blood pressure (hypertension), and can also be used in children to treat hypertension. Nicardipine was shown to have an inhibitory effect on Nav1.8 in vitro as well as in Tcf4 +/- mice, showing promising effects on behavior, learning and memory. In this study, nicardipine was given orally for 7 months (starting dose 0.2 mg/kg/d, maximum dose 1.7 mg/kg/d). There were no significant side effects. The patient showed mild to moderate improvement in all developmental trajectories as well as in her restlessness. Repurposing nicardipine in Pitt-Hopkins syndrome patients could be a promising approach to enhance development in these often severely affected patients.

PMID:40786031 | PMC:PMC12331575 | DOI:10.3389/fphar.2025.1592011

Clin J Am Soc Nephrol. 2025 Aug 11. doi: 10.2215/CJN.0000000858. Online ahead of print.

ABSTRACT

Lupus nephritis (LN) is seeing more and more enriching immunotherapies, but important unmet needs remain. Here we discuss how to focus on histological signs of immunological activity triggering immunotherapy versus signs of irreversible kidney injury requiring care for chronic kidney disease. Also, the correct interpretation of residual proteinuria requires dissecting immunological activity from glomerular hyperfiltration, e.g., by repeat biopsy. Despite modern triple immunotherapy, per-protocol biopsies still document irreversible injury to occur in the first year. Immediate inhibition of the complement system may address this unmet need and may even help to ultimately replace early glucocoorticoid therapy. We advocate the concept of a clone-directed therapy to sufficiently suppress the autoreactive clones of memory B and T cells inside the lymphoid tissues as well as the long-lived plasma cells in the bone marrow that maintain activity of systemic lupus erythematosus (SLE) and drive disease flares. Numerous B cell- and plasma cell-targeting therapies are gradually becoming available and their parenteral route of application may also avoid oral drug non-adherence. Replacing oral and toxic medications such as steroids, mycophenolate, and calcineurin inhibitors is now a goal for the next decade. Obtaining orphan disease designation for LN would accelerate progress and is supported by latest data on LN prevalence. With these conceptual and management improvements, LN, once "complex" and frequently fatal, may become easy-to-manage as other autoimmune diseases.

PMID:40788686 | DOI:10.2215/CJN.0000000858

Front Oncol. 2025 Jul 25;15:1589724. doi: 10.3389/fonc.2025.1589724. eCollection 2025.

ABSTRACT

BACKGROUND: Colorectal cancer is the second most prevalent cancer in Chile, affecting both sexes. Late-stage diagnosis occurs in approximately 25% of cases, with a five-year survival rate of only 14%. Standard treatment involves surgical resection followed by 5-fluorouracil-based chemotherapy, often combined with oxaliplatin or irinotecan. However, patient responses vary significantly due to genetic polymorphisms affecting drug metabolism, including variants in TYMS, DPYD, GSTs, and DNA repair enzymes. While genetic factors influencing chemotherapy outcomes have been studied, their impact remains unclear and varies across populations. No predictive model integrating genetic and clinical variables for chemotherapy safety in Chilean colorectal cancer patients has been established.

OBJECTIVE: This study aimed to identify relevant genetic variants in TYMS, TYMP, DPYD, GSTP1, MTHFR, ERCC2, ABCB1, ABCC2, ABCC4, and ABCG2 genes, which, combined with clinical factors, could contribute to a predictive model for 5-FU-based chemotherapy safety in advanced colorectal cancer patients.

METHODS: A retrospective nested case-control study was conducted on 82 advanced colorectal cancer patients. Sixteen genetic variants were analyzed to assess their association with adverse reactions and their severity using logistic regression. Multivariate models were developed to predict chemotherapy safety.

RESULTS: Among the 16 variants analyzed in 82 patients, key findings included: The G allele of GSTP1 (rs1695) was protective against neuropathy (OR = 0.147; p = 0.012) but increased mucositis risk (OR = 2.27; p = 0.036). The C allele of DPYD (rs1801265) was linked to a higher neuropathy risk (OR = 4.58; p = 0.05). The TYMS deletion genotype (rs11280056) conferred protection against hematological adverse reactions (OR = 0.029; p = 0.001). On the other hand, the 3R genotype of TYMS 5'UTR (rs45445694) is associated as a risk factor for skin and subcutaneous tissue disorders (OR = 6.40; p = 0.029). Two multivariate models were developed to predict anemia (p = 0.027) and pain (p = 0.01) development.

CONCLUSIONS: This study provides a foundation for developing pharmacogenetic-based predictive models for adverse reactions associated with 5-FU, including neuropathy, mucositis, and hematological and skin toxicities. Future research may refine these models to enable personalized dose adjustments, improving chemotherapy safety in Chilean colorectal patients.

PMID:40786508 | PMC:PMC12331468 | DOI:10.3389/fonc.2025.1589724

Pancreas. 2025 Aug 12. doi: 10.1097/MPA.0000000000002544. Online ahead of print.

ABSTRACT

OBJECTIVE: This study aimed to describe the frequency of exocrine pancreatic disease in youth with diabetes.

METHODS: We conducted a retrospective chart review on data that was obtained from a single center prospectively collected database of patients with diabetes. Patients were categorized as having type 1 diabetes, type 2 diabetes, or "Other" diabetes if they had cystic fibrosis-related diabetes, maturity onset diabetes of the young, or drug/chemical induced diabetes. All patients'charts were reviewed for exocrine pancreas disease, inclusive of pancreatitis or exocrine pancreatic insufficiency.

RESULTS: Nine-hundred and eighty-eight patients with a diabetes diagnosis were included. Thirty five out of 988 (3.5%) were diagnosed with pancreatic exocrine disease. Diabetes patients with exocrine disease compared to the ones without were significantly older (13.1 years, IQR 9.8-15.3 vs. 11.7 years, P= 0.04). Those with exocrine disease were more likely to have "Other" diabetes (P<0.0001). The exocrine group had a lower median hemoglobin A1c at diabetes diagnosis (7%, IQR 5.8-9.2% vs. 11.3%, 8.9-13.8%; P<0.0001). Out of the 988 patients, 18 patients had pancreatitis diagnosed, which was 2% of the overall cohort. Nine of the 18 patients were found to have developed pancreatitis after diabetes diagnosis, or 1% of the entire diabetes cohort (9/988).

CONCLUSIONS: The co-existence of exocrine and endocrine pancreatic disease occurred in 3.5% of diabetes patients. The risk of pancreatitis occurring after diabetes was 1%, a rate 100 times higher than the general pediatric population (0.01%). Future studies are needed to determine the specific mechanisms involved in the connection between endocrine and exocrine pancreatic disease in children.

PMID:40788280 | DOI:10.1097/MPA.0000000000002544

Clin Transplant. 2025 Aug;39(8):e70270. doi: 10.1111/ctr.70270.

ABSTRACT

BACKGROUND: En bloc heart-lung transplantation (HLTx) has been utilized for the past 50 years for the treatment of end-stage heart and lung disease, with significant evolution in the field over that time. This is a systematic review of HLTx and a description of the evolution and outcomes in this patient population.

METHODS: Pubmed and Embase were searched for all articles on HLTx from the time of database inception. A total of 1513 articles were screened, and after exclusion, 29 were included in this systematic review.

RESULTS: Reported cases of HLTx were more common in the early era (before 2000), for the indications of cystic fibrosis, Eisenmenger's syndrome, and pulmonary hypertension. In the contemporary era (2000-present), patients were not as commonly transplanted for cystic fibrosis, with pulmonary hypertension and congenital heart disease comprising the majority of cases. Rates of short-term mortality tended to be lower in more recent studies, with only recent studies reporting long-term survival.

DISCUSSION: HLTx has evolved substantially. In tandem with isolated heart and lung transplantation, the indications for transplant, medical therapy, and outcomes have changed over time. While HLTx is used less frequently in contemporary times compared to the early days of cardiothoracic transplantation, indications for HLTx continue to exist, and the use of HLTx will continue to be indicated. Centers with experience in HLTx should continue to report trends in patient management and outcomes, to continue to guide continued refinement in the field of HLTx.

PMID:40788177 | DOI:10.1111/ctr.70270

AACE Endocrinol Diabetes. 2025 Apr 11;12(2):67-70. doi: 10.1016/j.aed.2025.03.011. eCollection 2025 Jul-Aug.

ABSTRACT

BACKGROUND/OBJECTIVE: Cystic fibrosis (CF)-related diabetes (CFRD) is a common extrapulmonary complication of CF, with increasing prevalence. As individuals with CF live longer, obesity rates are increasing, leading to an emerging phenotype called CFRD with metabolic syndrome. The objective of this report is to describe the use of semaglutide in an adult with CFRD, obesity, and clinical insulin resistance.

CASE REPORT: A 32-year-old man with CF, pancreatic insufficiency, obesity, and poorly controlled CFRD presented with worsening blood sugar control, increasing insulin requirements, and a strong family history of metabolic syndrome. His body mass index was 38.5 kg/m2, and his hemoglobin A1c level ranged from 9.4% to 11.4%. He reported difficulty adhering to insulin therapy and concerns regarding weight and body image. A continuous glucose monitor was initiated; however, it did not significantly improve glycemic control. Given his metabolic profile and desire to lose weight, semaglutide was introduced and gradually increased over 5 months. This improved the hemoglobin A1c level by 5.7%, lowered the mean glucose levels, reduced the body mass index to 33.4 kg/m2, and decreased insulin requirements without adverse effects.

DISCUSSION: Although insulin is the primary treatment for CFRD, glucagon-like peptide 1 receptor agonists may provide additional benefits in carefully selected patients.

CONCLUSION: This case highlights the potential benefits of glucagon-like peptide 1 receptor agonists in CFRD with metabolic syndrome and emphasizes the need for further investigation.

PMID:40786988 | PMC:PMC12332434 | DOI:10.1016/j.aed.2025.03.011

Abdom Radiol (NY). 2025 Aug 11. doi: 10.1007/s00261-025-05149-7. Online ahead of print.

ABSTRACT

BACKGROUND: Fully automated AI-based algorithms can quantify adipose tissue on abdominal CT images. The aim of this study was to investigate the clinical value of these biomarkers by determining the association between adipose tissue measures and all-cause mortality.

METHODS: This retrospective study included 151,141 patients who underwent abdominal CT for any reason between 2000 and 2021. A validated AI-based algorithm quantified subcutaneous (SAT) and visceral (VAT) adipose tissue cross-sectional area. A visceral-to-subcutaneous adipose tissue area ratio (VSR) was calculated. Clinical data (age at the time of CT, sex, date of death, date of last contact) was obtained from a database search of the electronic health record. Hazard ratios (HR) and Kaplan-Meier curves assessed the relationship between adipose tissue measures and mortality. The endpoint of interest was all-cause mortality, with additional subgroup analysis including age and gender.

RESULTS: 138,169 patients were included in the final analysis. Higher VSR was associated with increased mortality; this association was strongest in younger women (highest compared to lowest risk quartile HR 3.32 in 18-39y). Lower SAT was associated with increased mortality regardless of sex or age group (HR up to 1.63 in 18-39y). Higher VAT was associated with increased mortality in younger age groups, with the trend weakening and reversing with age; this association was stronger in women.

CONCLUSION: AI-based CT measures of SAT, VAT, and VSR are predictive of mortality, with VSR being the highest performing fat area biomarker overall. These metrics tended to perform better for women and younger patients. Incorporating AI tools can augment patient assessment and management, improving outcome.

PMID:40788576 | DOI:10.1007/s00261-025-05149-7

Indian Dermatol Online J. 2025 Aug 11. doi: 10.4103/idoj.idoj_1055_24. Online ahead of print.

ABSTRACT

Psoriasis is a chronic inflammatory skin condition affecting millions of people globally, with prevalence varying significantly between countries. Conventional treatments, including topical agents, phototherapy, and systemic medications, often fail to account for individual variability, leading to suboptimal outcomes and potential adverse effects. Artificial intelligence (AI) has emerged as a promising approach to enhance precision and personalization in psoriasis management, potentially transforming diagnostic accuracy and treatment selection. This review examines the integration of AI across multiple domains of psoriasis treatment: (1) machine learning algorithms for phototherapy outcome prediction, (2) deep learning techniques for lesion segmentation and severity assessment, (3) AI-enhanced remote photographic monitoring systems, and (4) predictive modeling for response to systemic therapies and biologics. The analysis encompasses various AI methodologies, including random forest classifiers, convolutional neural networks, multiscale superpixel clustering, and gradient-boosted decision trees applied to clinical datasets, imaging analysis, and multi-omic patient data. AI-driven models demonstrate significant clinical utility with phototherapy outcome prediction, achieving high sensitivity (>84%) and accuracy (75-85%). Automated lesion segmentation reaches 86.99%-pixel accuracy, while remote AI assessments strongly correlate with clinical evaluations (Intraclass Correlation Coefficient [ICC] = 0.78-0.99). Notably, predictive models can forecast biologic therapy responses with > 95% accuracy within 2-4 weeks of treatment initiation, substantially reducing evaluation timelines from the conventional 12-week assessment period. AI technologies offer transformative potential in psoriasis management by enabling precise diagnosis, outcome prediction, and personalized therapy selection. Current implementations show promising results across diverse clinical applications, from phototherapy optimization to biologic response prediction. While challenges in dataset diversity, standardization, and validation remain, these represent opportunities for further advancement toward precision medicine in dermatology.

PMID:40788101 | DOI:10.4103/idoj.idoj_1055_24

Adv Sci (Weinh). 2025 Aug 11:e04079. doi: 10.1002/advs.202504079. Online ahead of print.

ABSTRACT

The precise matching of medication regimens to individual patients, known as personalized medication, is critical for the effective management of chronic diseases. Traditional machine learning-based models for personalized medication regimens typically rely solely on either clinical macro-phenotypes or molecular-level drug characteristics. It remains challenging to capture the patient-medication relationship from a comprehensive perspective that integrates individual patient characteristics with macro- and micro-level properties of the medication. Determining patient-medication relationships constitutes a three-stage sequential decision process from a clinical decision-making perspective. Therefore, inspired by Chain-of-Thought prompting, which simulates the decision-making process of human experts, a Multimodal Data-Driven Chain-of-Decisions (MDD-CoD) framework is proposed, where three-stage deep learning tasks are sequentially organized to reflect upstream-downstream logical dependencies, thereby forming a coherent clinical decision-making process. The model incorporates multimodal clinical phenotype data, multi-attribute medication data, and insights from clinical experts. Performance evaluation of the model involved comprehensive experiments utilizing five datasets covering four chronic diseases sourced from three hospitals. The dataset comprises information from chronic kidney disease (CKD), membranous nephropathy (MN), rheumatoid arthritis (RA), colorectal cancer (CRC), and knee osteoarthritis (KOA), totaling 3173 unimodal, 502 multimodal, and 2187 medication records from 3675 patients. Experimental results demonstrate that the framework achieves enhanced predictive performance in personalized medication decision-making based on individual patient disease characteristics, surpassing the strongest baseline across all tasks. This framework serves as a foundational model for clinical mixed data, with improved generalization and interpretability in cross-disease personalized decision-making tasks. It offers a scalable solution for the implementation of personalized medication regimens for chronic diseases.

PMID:40788064 | DOI:10.1002/advs.202504079

Adv Sci (Weinh). 2025 Aug 11:e07189. doi: 10.1002/advs.202507189. Online ahead of print.

ABSTRACT

Surface-Enhanced Raman Scattering (SERS) shows promise for monitoring health during space missions, particularly in assessing the effects of microgravity and radiation. However, traditional SERS sensors struggle with precise interfacial engineering, leading to a relatively poor assembly efficiency, and are unable to meet the practical needs of extreme spaceflight environments. To address this, it is designed and fabricated precision-arranged DNA origami plasmonic nanoantennas. By leveraging DNA origami's addressability, it is built a 3 × 4 antenna array with a controlled spacing of 21.76 nm, enhancing assembly efficiency fourfold compared to disordered systems. The ordered system enabled accurate detection of calcium ions, interleukin-6, and microRNA-214 in serum from mice exposed to microgravity and radiation, with intraclass correlation coefficients > 0.75, comparable to ELISA and qPCR. More importantly, integrating the system with a convolutional neural network enabled precise bone health prediction. This platform provides a promising tool for astronaut health monitoring.

PMID:40788053 | DOI:10.1002/advs.202507189

Adv Sci (Weinh). 2025 Aug 11:e07044. doi: 10.1002/advs.202507044. Online ahead of print.

ABSTRACT

Peptide drugs have revolutionized modern therapeutics, offering novel treatment avenues for various diseases. Nevertheless, low design efficacy, time consumption, and high cost still hinder peptide drug design and discovery. Here, an efficient approach that integrates deep learning-based protein design with functional screening is presented, enabling the rapid design of biotechnologically important peptides with improved stability and efficacy. 10,000 de novo glucagon-like peptide-1 receptor agonists (GLP-1RAs) are designed, 60 of these satisfied the stability, efficacy, and diversity criteria in the virtual functional screening. In vitro validations reveal a 52% success rate, and in vivo experiments demonstrate that two lead GLP-1RAs (D13 and D41) exhibit extended half-lives, approximately three times longer than that of Semaglutide. In diabetic mouse models, candidate D13 results in significantly lower blood glucose levels than Semaglutide. In the obesity mouse model, D13 induces weight loss efficacy comparable to that of Semaglutide. The AI-driven peptide design pipeline-which integrates protein design, functional screening, and experimental validation-reduces the number of iterations required to find novel peptide candidates. The entire process, from design to screening, can be completed in a single cycle within two weeks.

PMID:40787887 | DOI:10.1002/advs.202507044

Compr Rev Food Sci Food Saf. 2025 Sep;24(5):e70248. doi: 10.1111/1541-4337.70248.

ABSTRACT

Cereal quality significantly influences human health, requiring thorough evaluation of authenticity, nutritional composition, and food safety hazards. Conventional detection methods are often characterized by limitations, including time-consuming intricacy, complexity, and limited sensitivity. Recently, optical imaging and spectroscopy have emerged as rapid, nondestructive, and high-throughput alternatives for assessing cereal quality. The integration of chemometrics and artificial intelligence (AI), particularly deep learning algorithms, is paramount in the processing and analysis of optical data, which is indispensable for extracting key features from large datasets. In this work, the advanced spectroscopy and optical imaging techniques are comprehensively introduced, and their recent progress in applied research is outlined, emphasizing the major innovations and practical applications of these techniques. Besides, the latest developments of these techniques and AI-driven data processing methods in various aspects of cereal quality assessment have been summarized in order to highlight the potential research directions and future trends for practical application.

PMID:40787808 | DOI:10.1111/1541-4337.70248

ACS Omega. 2025 Jul 22;10(30):33662-33674. doi: 10.1021/acsomega.5c04524. eCollection 2025 Aug 5.

ABSTRACT

The interaction between circular RNAs (circRNAs) and RNA-binding proteins (RBPs) plays a crucial role in gene regulation; however, experimental identification is costly and inefficient. Current computational methods often overlook the structural features of circRNAs, thereby limiting prediction accuracy. To address these challenges, we propose GGCRB, a deep learning framework that integrates both sequence and structural features for predicting circRNA-RBP binding sites. Sequence features are captured through five encoding schemes (HFN, ND, NCP, DPCP, and Doc2Vec), followed by convolutional layers for local pattern extraction. Structural features are derived from base-pairing adjacency matrices generated by RNAstructure and modeled using graph convolutional networks and graph attention networks to learn topological dependencies. The fused representations are further processed by bidirectional LSTM and multihead attention modules to capture global interactions. Final predictions are made through pooling and softmax layers. Extensive experiments on 16 benchmark data sets demonstrate that GGCRB significantly outperforms existing models. Ablation studies and motif analyses further confirm its effectiveness, underscoring the importance of integrating structural and sequence information for accurate prediction of circRNA-RBP interactions.

PMID:40787315 | PMC:PMC12332793 | DOI:10.1021/acsomega.5c04524

ACS Omega. 2025 Jul 23;10(30):33031-33044. doi: 10.1021/acsomega.5c01924. eCollection 2025 Aug 5.

ABSTRACT

DNA methylation is an epigenetic modification that plays a crucial role in genome stability and cellular specialization, essential for maintaining normal cellular function and development, also a manifestation indicator of some diseases. Various tools have been proposed for methylation detection, typically leveraging a third-generation sequencing technology called nanopore sequencing, which provides more accurate DNA sequencing data. However, existing tools have their own limitations and advantages in terms of computational resources and information processing, without achieving a good balance. In this situation, we developed EDNTOM (Ensemble Deep Network Tool Of Methylation), a DNA methylation detection tool based on deep learning technology. We employed ensemble learning techniques, integrating predictions from multiple pretrained single models, and introduced an attention weight mechanism to provide accurate and reliable detection, reducing the consumption of computational resources. Results demonstrate that EDNTOM outperforms individual models. Additionally, in cross-species transfer experiments, EDNTOM exhibits strong transfer learning capabilities. We hope this work can provide a more powerful and reliable solution for methylation detection, contributing to the fields of biological science and medicine. The project code is available at https://github.com/ViceMusic/EDNTOM.

PMID:40787313 | PMC:PMC12332607 | DOI:10.1021/acsomega.5c01924