Eur Heart J. 2025 May 6:ehaf176. doi: 10.1093/eurheartj/ehaf176. Online ahead of print.

NO ABSTRACT

PMID:40326349 | DOI:10.1093/eurheartj/ehaf176

J Pharmacokinet Pharmacodyn. 2025 May 5;52(3):29. doi: 10.1007/s10928-025-09976-5.

ABSTRACT

Two recent papers offer contrasting perspectives on integrating Quantitative Systems Pharmacology (QSP) and Artificial Intelligence/Machine Learning (AI/ML): one views QSP as the primary driver using AI/ML to enhance computational tasks, while the other argues that AI/ML should provide an alternative mechanistic framework. Rather than perpetuate this tension, we used Large Language Models (LLMs) to examine both papers in two tests-one comparing their core arguments and another probing which methodology LLM should take precedence. Repeating each test multiple times with an identical and neutral prompt, the LLM revealed that each perspective suits specific stages of the drug development pipeline. QSP offers mechanistic rigor and regulatory clarity, and AI/ML excels in high-dimensional data analysis and exploratory modeling. A hybrid approach might best serve researchers and decision-makers, especially when harmonizing data-driven insights with mechanistic integrity. This exercise also highlights the potential of LLMs as promising tools for synthesizing complex information, offering an arguably less biased viewpoint that can trigger deeper discussion from the broader community seeking to align QSP and AI/ML in model-informed drug development (MIDD). By combining our human expertise with AI-driven analyses, we hope to further discuss with the scientific community how QSP and AI/ML-and the synergy between them-can drive innovation in therapeutic discovery and optimization.

PMID:40325298 | DOI:10.1007/s10928-025-09976-5

NPJ Digit Med. 2025 May 6;8(1):250. doi: 10.1038/s41746-025-01673-4.

ABSTRACT

Clinical trials are an essential component of drug development for new cancer treatments, yet the information required to determine a patient's eligibility for enrollment is scattered in large amounts of unstructured text. Genomic biomarkers are especially important in precision medicine and targeted therapies, making them essential for matching patients to appropriate trials. Large language models (LLMs) offer a promising solution for extracting this information from clinical trial study descriptions (e.g., brief summary, eligibility criteria), aiding in identifying suitable patient matches in downstream applications. In this study, we explore various strategies for extracting genetic biomarkers from oncology trials. Therefore, our focus is on structuring unstructured clinical trial data, not processing individual patient records. Our results show that open-source language models, when applied out-of-the-box, effectively capture complex logical expressions and structure genomic biomarkers, outperforming closed-source models such as GPT-4. Furthermore, fine-tuning these open-source models with additional data significantly enhances their performance.

PMID:40325165 | DOI:10.1038/s41746-025-01673-4

Genes Dev. 2025 May 5. doi: 10.1101/gad.352512.124. Online ahead of print.

ABSTRACT

Despite the general detriment of aneuploidy to cellular fitness, >90% of solid tumors carry an imbalanced karyotype. This existing paradox and the molecular responses to aneuploidy remain poorly understood. Here, we explore these cellular stresses and unique vulnerabilities of aneuploidy in human mammary epithelial cells (HMECs) enriched for breast cancer-associated copy number alterations (CNAs). To uncover the genetic dependencies specific to aneuploid cells, we conducted a comprehensive, genome-wide CRISPR knockout screen in isogenic aneuploid and diploid HMEC lines. Our study reveals that aneuploid HMECs exhibit an increased reliance on pyrimidine biosynthesis and mitochondrial oxidative phosphorylation genes and demonstrate heightened fitness advantages upon loss of tumor suppressor genes. Using an integrative multiomic analysis, we confirmed nucleotide pool insufficiency as a key contributor to widespread cellular dysfunction in aneuploid HMECs with net copy number gain. Although diploid cells can switch seamlessly between pyrimidine synthesis and salvage, cells with increased chromosomal content exhibit p53 activation and S-phase arrest when relying on salvage alone, alongside increased sensitivity to DNA-damaging chemotherapeutics. This work advances our understanding of the consequences of aneuploidy and uncovers potential avenues for patient stratification and therapeutic intervention based on tumor ploidy.

PMID:40324880 | DOI:10.1101/gad.352512.124

Biotechnol Adv. 2025 May 3:108592. doi: 10.1016/j.biotechadv.2025.108592. Online ahead of print.

ABSTRACT

Inflammation is the body's response to infection, trauma or injury and is activated in a coordinated fashion to ensure the restoration of tissue homeostasis and healthy physiology. This process requires communication between stromal cells resident to the tissue compartment and infiltrating immune cells which is dysregulated in disease. Clinical innovations in patient diagnosis and stratification include measures of inflammatory activation that support the assessment of patient prognosis and response to therapy. We propose that (i) the recent advances in fast, dynamic monitoring of inflammatory markers (e.g., cytokines) and (ii) data-dependent theoretical and computational modeling of inflammatory marker dynamics will enable the quantification of the inflammatory response, identification of optimal, disease-specific biomarkers and the design of personalized interventions to improve patient outcomes - multidisciplinary efforts in which biomedical engineers may potentially contribute. To illustrate these ideas, we describe the actions of cytokines, acute phase proteins and hormones in the inflammatory response and discuss their role in local wounds, COVID-19, cancer, autoimmune diseases, neurodegenerative diseases and aging, with a central focus on cardiac surgery. We also discuss the challenges and opportunities involved in tracking and modulating inflammation in clinical settings.

PMID:40324661 | DOI:10.1016/j.biotechadv.2025.108592

PLoS One. 2025 May 5;20(5):e0322498. doi: 10.1371/journal.pone.0322498. eCollection 2025.

ABSTRACT

Embeddings are semantically meaningful representations of words in a vector space, commonly used to enhance downstream machine learning applications. Traditional biomedical embedding techniques often replace all synonymous words representing biological or medical concepts with a unique token, ensuring consistent representation and improving embedding quality. However, the potential impact of replacing non-biomedical concept synonyms has received less attention. Embedding approaches often employ concept replacement to replace concepts that span multiple words, such as non-small-cell lung carcinoma, with a single concept identifier (e.g., D002289). Also, all synonyms of each concept are merged into the same identifier. Here, we additionally leveraged WordNet to identify and replace sets of non-biomedical synonyms with their most common representatives. This combined approach aimed to reduce embedding noise from non-biomedical terms while preserving the integrity of biomedical concept representations. We applied this method to 1,055 biomedical concept sets representing molecular signatures or medical categories and assessed the mean pairwise distance of embeddings with and without non-biomedical synonym replacement. A smaller mean pairwise distance was interpreted as greater intra-cluster coherence and higher embedding quality. Embeddings were generated using the Word2Vec algorithm applied to a corpus of 10 million PubMed abstracts. Our results demonstrate that the addition of non-biomedical synonym replacement reduced the mean intra-cluster distance by an average of 8%, suggesting that this complementary approach enhances embedding quality. Future work will assess its applicability to other embedding techniques and downstream tasks. Python code implementing this method is provided under an open-source license.

PMID:40324016 | DOI:10.1371/journal.pone.0322498

PLoS One. 2025 May 5;20(5):e0322308. doi: 10.1371/journal.pone.0322308. eCollection 2025.

ABSTRACT

Riboswitches are structured elements predominantly found in the 5'-untranslated region of many bacterial mRNA. These noncoding RNA regions play a vital role in bacterial metabolism and overall function. Each riboswitch binds to a specific small molecule and causes conformational changes in the mRNA leading to regulation of transcription or translation. In this work, we have synthesized SK4, a novel nucleoside analog that binds to the guanine riboswitch mRNA of the xanthine phosphoribosyl transferase gene in Bacillus subtilis and terminates transcription of the riboswitch mRNA to a greater extent than the native ligand guanine. Molecular dynamics simulations of SK4 with riboswitch mRNA reveal an overall stable complex with additional bonding interactions in comparison to guanine. Our work with SK4 demonstrates that specific genes in bacteria can be effectively controlled by ligand analogs, providing an alternative mechanism to regulate the function of bacteria.

PMID:40323922 | DOI:10.1371/journal.pone.0322308

Aging (Albany NY). 2025 May 3;17. doi: 10.18632/aging.206243. Online ahead of print.

ABSTRACT

Apolipoprotein E (APOE) modifies human aging; specifically, the ε2 and ε4 alleles are among the strongest genetic predictors of longevity and Alzheimer's disease (AD) risk, respectively. However, detailed mechanisms for their influence on aging remain unclear. In the present study, we analyzed multi-omic association patterns across APOE genotypes, sex, and biological age (BA) axes in 2,229 community dwelling individuals. Our analysis, supported by validation in an independent cohort, identified diacylglycerols as the top APOE-associated plasma metabolites. However, despite the known opposing aging effects of the allele variants, both ε2- and ε4-carriers showed higher diacylglycerols compared to ε3-homozygotes. 'Omics association patterns of ε2-carriers and increased biological age were also counter-intuitively similar, displaying significantly increased associations between insulin resistance markers and energy-generating pathway metabolites. These results demonstrate the context-dependence of the influence of APOE, with ε2 potentially strengthening insulin resistance-like pathways in the decades prior to imparting its longevity benefits. Additionally, they provide an atlas of APOE-related 'omic associations and support the involvement of bioenergetic pathways in mediating the impact of APOE on aging.

PMID:40323280 | DOI:10.18632/aging.206243

J Biosci. 2025;50:31.

ABSTRACT

The switch in the predominantly expressed transcript isoform of the same gene has been identified as a significant factor in the progression of various types of cancer. These switches can impact the gain or loss of different 3'UTRs, which are hotspots for the binding of microRNAs (miRNAs) and RNA-binding proteins (RBPs). In this study, we found that in cancer-specific dominant expressing transcripts, the binding of miRNA and RBP is disrupted, suggesting that transcript switching could play a part in modulating post-transcriptional gene expression during the progression and development of cancer. Our spatial correlation analysis demonstrated that changes in miRNA and RBP binding, triggered by transcript switching, could interrupt their interplay. Additionally, statistical analysis revealed that local folding energy (LFE) is a key factor in changing miRNA and RBP interactions due to isoform switching. Overall, this study revealed that changes in cancerspecific transcripts could influence miRNA-RBP interactions due to alternations in the local RNA structure of the transcript caused by isoform switching, thereby leading to the dysregulation of crucial genes involved in the evolution and progression of cancer.

PMID:40323003

Nucleic Acids Res. 2025 May 5:gkaf387. doi: 10.1093/nar/gkaf387. Online ahead of print.

ABSTRACT

The WashU Epigenome Browser (https://epigenomegateway.wustl.edu/) is a web-based tool for exploring genomic data and providing visualization, investigation, and analysis of epigenomic datasets. Since its 2018 update, the redesigned user interface and newly developed features have enhanced how investigators interact with both the Browser and the extensive genomic data it hosts. The rapid evolution of the JavaScript ecosystem has presented new challenges and opportunities in maintaining and developing the WashU Epigenome Browser. In this update, we present a completely rewritten codebase. This new codebase minimizes the use of external libraries whenever possible, resulting in a significantly smaller code bundle size after production compilation. The reduced code size improves loading efficiency and boosts the Browser's performance, with improved scripting, graphics rendering, and painting performance. Lowering external dependencies also allows for faster and more straightforward installation. Additionally, the update includes a redesign of the user interface to further enhance user experience and features a new modular design in the codebase that enables the Browser to be exported as stand-alone modules for use in other web applications. Several novel track types for long-read methylation data and single-cell methylation data visualization have been added, and we continue to update and expand the data hubs we host for major consortia. We constructed the first data hub to systematically compare genomic data mapped to different genome assemblies, focusing on comparisons between hg38 and the first human T2T genome, chm13, using our new comparative genomics track function. The WashU Epigenome Browser also serves as a foundation for other genomics platforms, such as the WashU Virus Genome Browser, developed for SARS-COV-2 research, the WashU Comparative Epigenome Browser, and the WashU Repeat Browser.

PMID:40322916 | DOI:10.1093/nar/gkaf387

J Exp Bot. 2025 May 4:eraf187. doi: 10.1093/jxb/eraf187. Online ahead of print.

ABSTRACT

Tomato (Solanum lycopersicum L.) serves as a major food source and a model crop for understanding plant responses to stress. Abiotic and biotic stresses, exacerbated by climate change, threaten global tomato production. Stress hormones, including abscisic acid (ABA), ethylene (ET), jasmonates (JAs), and salicylic acid (SA), orchestrate intricate signaling pathways that mediate plant immunity and metabolism. This review synthesizes the roles of these hormones in tomato stress responses. We discuss the biosynthesis and signalling cascades of these stress hormones, and focus on the cellular and metabolic reprogramming they cause and the crosstalk that occurs between them. Increased understanding of these molecular events and interactions provides insights to improve tomato resilience and productivity under environmental challenges.

PMID:40322793 | DOI:10.1093/jxb/eraf187

Zookeys. 2025 Apr 24;1236:19-39. doi: 10.3897/zookeys.1236.143336. eCollection 2025.

ABSTRACT

Exploring genetic diversity is essential for precise species delimitation, especially within taxonomically complex groups like passerine birds. Traditional morphological methods often fail to resolve species boundaries; however, DNA barcoding, particularly through the mitochondrial cytochrome c oxidase subunit I (COI) gene, provides a powerful complementary method for accurate species identification. This study establishes a comprehensive DNA barcode library for Iranian passerine birds, analyzing 546 COI sequences from 94 species across 23 families and 53 genera. There is a pronounced barcode gap, with average intraspecific divergence at 0.41% and interspecific divergence at 18.6%. Notable intraspecific variation emerged in the Persian nuthatch (Sittatephronota) and the Lesser whitethroat (Currucacurruca), while the European goldfinch (Cardueliscarduelis) and the grey-crowned goldfinch (Cardueliscaniceps) showed limited genetic differentiation despite marked morphological distinctions. Phylogenetic analysis revealed significant east-west genetic splits in C.curruca and S.tephronota, reflecting Iran's geographic and zoogeographic boundaries. These findings demonstrate the effectiveness of DNA barcoding in elucidating biogeographic patterns, emphasizing Iran's key role as an ornithological crossroads for avian biodiversity. Moreover, our results suggest that much of the genetic variation in the COI gene arises from synonymous mutations, highlighting the role of purifying selection in shaping mtDNA diversity across species.

PMID:40322611 | PMC:PMC12046340 | DOI:10.3897/zookeys.1236.143336

One Health. 2025 Apr 12;20:101035. doi: 10.1016/j.onehlt.2025.101035. eCollection 2025 Jun.

ABSTRACT

BACKGROUND AND OBJECTIVE: To assess the association of pulmonary tuberculosis (PTB) recurrence with fine particulate matter (PM2.5) and residential greenness using a population-based retrospective study design.

METHODS: All incident PTB patients, registered in Tuberculosis Information Management System (TBIMS) from 2015 to 2019 in Quzhou City, China, were included. The data on PM2.5 exposure was extracted from the China High Air Pollutants dataset and the level of greenness was estimated using the Normalized Difference Vegetation Index (NDVI) values around the patient's residence. The Cox proportional hazards models were used to quantify the risk of PTB recurrence.

RESULTS: 6732 Eligible PTB incident patients were included in the study with a mean age of 56.86 years and a median follow-up time of 750 days. Recurrence was observed in 554 patients (8.2 %). Exposure to NDVI was observed to be negatively associated with PTB recurrence (HR: 0.86, 95 % CI: 0.75-0.98 per 0.1-unit increase). The strength of the association between higher PM2.5 and the risk of PTB recurrence was greater than that of lower PM2.5 concentrations in both low and high NDVI groups (HR:6.62 and 4.35, p-interaction <0.001).

CONCLUSIONS: Our findings suggest that higher PM2.5 exposure might increase the risk of PTB recurrence, while residential greenness might have a protective effect. Like other chronic respiratory diseases, prevention and control of PTB will also benefit from comprehensive environmental management.

PMID:40321627 | PMC:PMC12047573 | DOI:10.1016/j.onehlt.2025.101035

ACS Omega. 2025 Apr 15;10(16):16339-16354. doi: 10.1021/acsomega.4c10653. eCollection 2025 Apr 29.

ABSTRACT

INTRODUCTION: Primary Sjogren's Syndrome (pSS) is a chronic inflammatory autoimmune disease that manifests as dry mouth and eyes. Luteolin can repress immuno-inflammation and improve the function of exocrine glands.

METHODS: Bibliometrics was used to visualize pSS-related key indicators. The efficacy of traditional Chinese medicines (TCMs) in pSS treatment was analyzed with the internal database containing the clinical records of pSS. Using the network pharmacology technology to identify involved pathways. Additionally, molecular docking and cell experiments were performed to screen and verify the therapeutic effect of luteolin on pSS.

RESULTS: Key indicators that were selected according to the bibliometrics were worse in pSS and had certain compatibility and correlation with laboratory and immunoinflammatory indicators. After treatment, pSS patients showed improvements in the above indicators. The results of risk analyses revealed that TCMs were protective factors for laboratory indicators and key indicators. The main effective TCMs for pSS treatment and TNF pathways were identified with network pharmacology. Cell experiments validated that luteolin indeed improved the secretion dysfunction and inflammation of modeled human submandibular gland epithelial cells through the TNF/NF-κB pathway.

CONCLUSIONS: TCMs may effectively improve transcription factors and immuno-inflammatory markers in pSS patients. Moreover, we hypothesized and verified the potential mechanism of action of luteolin in HSG cells.

PMID:40321503 | PMC:PMC12044447 | DOI:10.1021/acsomega.4c10653

Ann N Y Acad Sci. 2025 May 5. doi: 10.1111/nyas.15352. Online ahead of print.

ABSTRACT

The human gut microbiome is a complex ecosystem that plays a vital role in maintaining health and contributing to the pathogenesis of various diseases. This review proposes a transformative approach that involves engineering artificial microbial consortia-precisely designed communities of microorganisms-for personalized modulation of the gut microbiome and targeted therapeutic interventions. By integrating synthetic biology, systems biology, and advanced culturing techniques, tailored microbial consortia can be developed to perform specific functions within the gut, including the production of therapeutic molecules, modulation of immune responses, and competition against pathogenic bacteria. In vitro and in vivo studies indicate that these engineered consortia can effectively restore microbial balance and enhance host resilience. This personalized approach holds immense potential to revolutionize healthcare by addressing the root causes of diseases such as metabolic disorders, inflammatory conditions, and gastrointestinal infections through precise manipulation of the gut microbiome. Future research should focus on rigorous clinical trials to evaluate the safety, efficacy, and long-term impacts of these engineered consortia in diverse human populations, paving the way for innovative microbial therapies that promote overall health and well-being.

PMID:40320966 | DOI:10.1111/nyas.15352

Clin Exp Allergy. 2025 May 4. doi: 10.1111/cea.70071. Online ahead of print.

NO ABSTRACT

PMID:40320688 | DOI:10.1111/cea.70071

Immunology. 2025 May 4. doi: 10.1111/imm.13940. Online ahead of print.

ABSTRACT

Our previous findings demonstrated that naïve B cells elicit suppressive CD4+ regulatory T (Treg) cells, named as Treg-of-B cells. However, the capability of antigen-specific B cells in that process remains unclear. Using ovalbumin (OVA) as a model antigen, the present study showed that B cells from OVA-immunised mice decreased that ability. Instead, OVA-activated OVA-specific (OB1) B cells induced effector-like T-of-OB1 cells without regulatory function. Phenotypically, Treg-of-B cells reduced the production of interferon (IFN)-γ, interleukin (IL)-17 and IL-2 and expressed CD62L, PD1 and endothelial cell adhesion molecule 1 (PECAM1). Functionally, adoptive transfer of Treg-of-B cells significantly attenuated Th1 cell-mediated delayed-type hypersensitivity (DTH) responses and inhibited IFN-γ-producing Th1 cells, while T-of-OB1 cells did not. Mechanistically, activated antigen-specific B cells increased the expression of costimulatory molecules and promoted higher T cell activation, contributing to effector T cell phenotype. Conversely, Treg-of-B cells exhibited lower T cell activation, possibly mediated through the expression of PECAM1, Dusp2, Dusp5, Ptpn7, Ptpn22 and Ms4a4b. These findings suggest that non-antigen-specific B cells elicit CD4+ Treg cells, potentially via attenuating T cell activation, whereas that capacity is absent in antigen-specific B cells. This distinction underscores the critical role of B cell antigen specificity in immune regulation and inflammation.

PMID:40320632 | DOI:10.1111/imm.13940

Sci Rep. 2025 May 4;15(1):15612. doi: 10.1038/s41598-025-99628-y.

ABSTRACT

Dengue is an important tropical disease with considerable global impact. Despite this, there remains an urgent need for reliable biomarkers to predict disease severity, as well as effective antiviral drugs and targeted treatments. In this study, we conducted a comprehensive profiling of 41 plasma mediators in patients with asymptomatic dengue (AD) and symptomatic dengue (SD), which includes mild dengue fever (DF) and severe dengue hemorrhagic fever (DHF). Our findings revealed that the levels of nearly all measured mediators were consistently lower in AD compared to SD patients, suggesting a potential protective cytokine response signature. Time-course cytokine analysis in SD shown significantly elevated levels of pro-inflammatory cytokines and chemokines associated with inflammation and viral clearance upon the acute phase, while various growth factors were elevated during the convalescence. Notably, we identified elevated IL-15 levels in DHF patients three days before fever subsidence, highlighting its potential as an early prognostic biomarker for severe disease outcomes. Furthermore, prolonged high levels of IL-8 and IP-10 in DHF during the critical period may contribute to dengue immunopathogenesis. This study advances the understanding of cytokine dynamics in the natural course of human dengue infection, providing valuable insights for the development of targeted treatments and prognostic biomarkers.

PMID:40320430 | DOI:10.1038/s41598-025-99628-y

Acta Biomater. 2025 May 2:S1742-7061(25)00328-9. doi: 10.1016/j.actbio.2025.05.007. Online ahead of print.

ABSTRACT

Scaffold-guided bone regeneration (SGBR) offers a promising solution for treating large-volume bone defects. However, its efficacy in compromised healing environments, such as those associated with metabolic conditions like Type 2 Diabetes (T2D), remains poorly understood. This study evaluates the potential of 3D-printed polycaprolactone (PCL) scaffolds for large-volume bone regeneration in preclinical models simulating T2D-induced metabolic challenges. Our results reveal that scaffolds alone are insufficient to overcome the metabolic barriers to effective bone regeneration. Metabolomic analysis of regenerating tissue identified significant disruptions in key metabolic pathways involved in energy production and amino acid synthesis in T2D rats compared to controls. Notably, aconitic acid, ornithine, and glycine levels were elevated in non-diabetic conditions, whereas phosphoenolpyruvate was markedly increased under T2D conditions. Secondary harmonic generation (SHG) imaging further demonstrated impaired collagen organization within T2D regenerating tissue, correlating with disrupted collagen synthesis critical for bone matrix formation. In vitro, the exogenous supplementation of alpha-ketoglutarate (α-KG)-a crucial citric acid cycle intermediate-enhanced mineralized tissue formation in human adipose-derived mesenchymal stem cells (hAdMSCs) from T2D donors, achieving levels superior to non-T2D cells. These findings underscore the metabolic underpinnings of impaired bone regeneration in T2D. Optimized 3D printed scaffolds alone do not counterbalance the impaired regeneration in T2D. Here we highlight a therapeutic potential of metabolic supplementation to optimize SGBR outcomes. This study provides a critical foundation for advancing translational research and developing regenerative therapies tailored to high-risk metabolic disease populations. STATEMENT OF SIGNIFICANCE: Scaffold-guided bone regeneration (SGBR) holds great promise for addressing large bone defects, but its efficacy in metabolically challenged conditions like Type 2 Diabetes (T2D) remains limited. This study uses a metabolomics-driven approach to reveal how metabolic dysregulation in T2D, including disruptions in energy and amino acid pathways, impairs collagen organization and extracellular matrix (ECM) formation-critical for successful bone healing. By identifying α-ketoglutarate (α-KG) as a potential supplement to restore metabolic balance, this work offers novel insights into enhancing scaffold performance under compromised conditions. These findings provide a foundation for integrating bioactive compounds into scaffold designs, advancing personalized strategies in regenerative medicine, and addressing a critical gap in bone defect treatment for diabetic patients.

PMID:40319991 | DOI:10.1016/j.actbio.2025.05.007

Comput Biol Med. 2025 May 3;192(Pt A):110284. doi: 10.1016/j.compbiomed.2025.110284. Online ahead of print.

ABSTRACT

BACKGROUND: The endometrioid subtype of endometrial cancer is a significant health concern for women, making it crucial to study the factors influencing patient outcomes.

METHOD: This study presents a novel survival analysis pipeline applied to multiomics data, including transcriptome, methylation, and proteome data, extracted from endometrioid samples in the TCGA-UCEC project to identify potential survival biomarkers. A major innovation in our work was the development of a deep learning autoencoder designed to capture the complex non-linear relationships between biological variables and survival outcomes. To achieve this, we defined a new loss function specifically for the autoencoder.

RESULT: The newly defined loss function can lead to extracting more survival information. The output of our pipeline includes 346 features ranked by their survival importance based on SHAP analysis, with a focus on the top 30 features. We analyzed the biological pathways enriched by these omics data and their contributions. As a result, we identified a relationship between Vitamin D, its receptor, and the Galanin receptor pathways with survival in endometrioid cancer.

CONCLUSION: This study introduces an innovative approach to survival analysis using multi-omics data and deep learning, with a greater focus on censored data to extract more survival information. It offers potential biomarkers for improved prognostic evaluation in endometrial cancer and presents pathway associations related to survival. These findings contribute to a better understanding of the progression of endometrial cancer.

PMID:40319755 | DOI:10.1016/j.compbiomed.2025.110284