NPJ Systems Biology and Applications最新文献

{"title":"A novel approach to quantify out-of-distribution uncertainty in Neural and Universal Differential Equations.","authors":"Stefano Giampiccolo, Giovanni Iacca, Luca Marchetti","doi":"10.1038/s41540-026-00749-5","DOIUrl":"10.1038/s41540-026-00749-5","url":null,"abstract":"<p><p>Dynamical systems play a central role across the quantitative sciences, offering a powerful mathematical framework to describe, analyze, and predict the evolution of complex processes over time. In systems biology, dynamical systems provide a foundation for modeling and predicting the intricate behaviors of biological systems. Recent advances in data-driven approaches, such as Neural Ordinary Differential Equations (NODEs) and Universal Differential Equations (UDEs), have enabled the development of models that are either fully or partially data-driven. Integrating data-driven components into dynamical systems amplifies the challenge of generalization beyond training data, highlighting the need for robust methods to quantify uncertainty in out-of-distribution (OOD) scenarios-i.e., conditions not encountered during training. In this work, we investigate the reliability of uncertainty quantification (UQ) based on ensembles of models in the reconstruction of dynamical systems. We show that standard ensembles (i.e., models trained independently with different random initializations) risk producing overconfident predictions in previously unseen scenarios, as the models in the ensemble tend to exhibit similar behaviors. To address this issue, we propose a novel ensemble construction method for NODEs and UDEs that fosters diversity in the reconstructed vector field across models within specific regions of the state space, while maintaining explicit control over the fit on the training set. We first evaluate our method on numerical test cases derived from three models commonly used as benchmarks for data-driven reconstruction of dynamical systems: the Lotka-Volterra model, the damped oscillator, and the Lorenz system. We then apply the method to a biologically motivated model of cell apoptosis, considering more realistic conditions such as partial observability of the system outputs and noise in the training dataset. Overall, our results show that the proposed method improves the reliability of UQ in previously unseen scenarios compared with standard ensembles, especially where the latter exhibit overconfidence.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13494032/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148043030","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Untargeted metabolomics reveals organism specific biomarkers of carbapenem resistance in Klebsiella pneumoniae and Escherichia coli.","authors":"Nicholas Bartelo, Yaxin Li, Ying Hao, Vijay Soni, Jamie Marino, He S Yang, Michael J Satlin, Fei Wang, Jan Krumsiek, Zhen Zhao","doi":"10.1038/s41540-026-00747-7","DOIUrl":"10.1038/s41540-026-00747-7","url":null,"abstract":"<p><p>Antimicrobial resistance (AMR) is one of the greatest global concerns due to the increase in the rate of AMR infections and the lack of development of antimicrobial agents to combat AMR. The development of resistance to carbapenems among common pathogenic bacteria, including Klebsiella pneumoniae and Escherichia coli, is particularly concerning because carbapenems are relied upon for the treatment of Gram-negative infections. Metabolomics offers an approach to identify biomarkers associated with metabolic mechanisms of carbapenem resistance. We analyzed 512 annotated metabolites from a total of 297 bacterial isolates spanning eight organisms. We discovered that the metabolome has greater variation between organisms compared with the resistance phenotype. As a result, we conducted organism-specific statistical analyses to discriminate carbapenem-resistant from carbapenem-sensitive isolates, and supervised learning models resulted in test set mean and standard deviation of area under the receiver operating characteristic curve of 0.822 ± 0.092 and 0.670 ± 0.110, and area under the precision-recall curve of 0.973 ± 0.016 and 0.653 ± 0.121, for Klebsiella pneumoniae and Escherichia coli, respectively. Feature selection using lasso logistic regression identified four metabolite biomarkers of carbapenem resistance in Klebsiella pneumoniae as 6-hydroxyphenazine-1-carboxamide (6-OH-Phz-1-Cam), N-Lactoyl tyrosine (N-Lac tyrosine), flavin adenine dinucleotide (FAD), and amalorin, and seven metabolite biomarkers in Escherichia coli as ibha#30, Cyclo(Leu-Lys), Asukamycin A-II, LPA 16:0, sphinganine, phytosphingosine, and PA 14:0/4:1;O. These unique metabolic signatures provide a vital foundation for exploring the emergence of AMR, warranting follow-up studies to clarify their role in carbapenem resistance and inform improved diagnostics and treatments for AMR.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13490478/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148033514","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Decoding cellular population dynamics through mechanistic modelling and statistical data analysis.","authors":"Nissrin Alachkar, Nicholas Kwasi-Do Ohene Opoku, Nicholas A M Monk, Kevin Thurley","doi":"10.1038/s41540-026-00751-x","DOIUrl":"10.1038/s41540-026-00751-x","url":null,"abstract":"<p><p>Cell-cell communication underlies key processes in development, immunity, and disease, yet capturing its mechanistic complexity remains challenging. While advances in single-cell omics have revealed new insights into cell-type diversity, mathematical modelling has become essential for deriving mechanistic understanding of their communication networks. Here, we overview established modelling approaches and highlight the need for frameworks that move beyond steady-state assumptions and single-step processes, better reflecting the nature of cell-cell communication.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":"12 1","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13201771/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148016620","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Molecular maps of diseases from omics data and network embeddings.","authors":"Dewei Hu, Anna-Lisa Schaap-Johansen, Julia Villarroel, Clara Ekebjærg, Simon Rasmussen, Daniel Hvidberg Hansen, Rasmus Wernersson, Lars Juhl Jensen","doi":"10.1038/s41540-026-00746-8","DOIUrl":"10.1038/s41540-026-00746-8","url":null,"abstract":"<p><p>Identifying disease-relevant proteins and pathways remains a fundamental challenge in understanding disease mechanisms and supporting therapeutic development. While omics analyses can provide valuable insights, they typically consider each gene/protein separately rather than at the level of biological systems. This can be addressed by combining the omics data with protein networks. We integrate disease-specific omics data with a universal functional association network from STRING, which we represent using node2vec embedding. This way, we constructed disease maps for seven diseases spanning inflammatory, oncological, neurological, and vascular diseases based on genetics, transcriptomics, somatic mutation, and proteomics data. Compared to omics analysis alone, the use of a simple linear model on top of network embedding enabled us to identify 2-4 times as many known disease-relevant proteins at the same specificity. Clustering of the resulting disease maps revealed both functional modules shared by many diseases, such as inflammatory pathways and cancer hallmarks, and disease-specific modules, such as keratinization in atopic dermatitis and extracellular matrix remodeling in aortic aneurysm. Together, these results highlight the value of protein network embedding when analyzing omics data to understand diseases.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13470450/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148006160","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cell trajectory inference based on schrödinger problem and a mechanistic model of stochastic gene expression.","authors":"Clémence Fournié, Elias Ventre, Ulysse Herbach, Aymeric Baradat, Olivier Gandrillon, Fabien Crauste","doi":"10.1038/s41540-026-00743-x","DOIUrl":"10.1038/s41540-026-00743-x","url":null,"abstract":"<p><p>Cellular differentiation is the biological process that leads a cell to opt for a particular cellular identity. Recently, single-cell RNA sequencing has enabled the simultaneous measurement of gene expression levels at specific times for a large number of individual cells and a large number of genes. Repeating such measurements at different time points gives them access to the temporal variation, or transport, of a distribution on a gene expression space. The whole temporal trajectory of distributions thus characterizes the differentiation process at the population level, but trajectories of individual cells are still out of reach since most measurement techniques are destructive. The optimal transport theory that has been used so far to infer cellular differentiation trajectories from time-stamped single-cell RNA-seq data involves solving the so-called Schrödinger problem in its most common version. This implies assuming that cells move, in the gene expression space, by diffusion. Yet, real gene dynamics are much more complex. In the present work, we assume that mRNA dynamics are characterized by brief and important production of RNA, with long periods of inactivity in between, and consider the so-called Bursty model of gene dynamics. We use this model to define a reference process for the Schrödinger problem. By comparing the solutions of the Schrödinger problems with a Diffusive and a Bursty reference process, under different conditions, we show that the Bursty model provides a better approximation of the underlying gene dynamics than the standard Diffusive process when inferring cell trajectories.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13463083/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147988385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Gene overexpression reduces inhibitory metabolites to enhance CHO cell growth and IgG1 production.","authors":"Duc Hoang, Bingyu Kuang, Zhao Wang, George Liang, Seongkyu Yoon","doi":"10.1038/s41540-026-00707-1","DOIUrl":"10.1038/s41540-026-00707-1","url":null,"abstract":"&lt;p&gt;&lt;p&gt;Controlling the generation of toxic by-products in mammalian bioprocess to maximize therapeutic protein production and glycosylation patterns is a challenge. Intracellular metabolism is often not well-regulated and known to secrete toxic intermediate by-products which hampers cellular performance and negatively impacts critical quality attributes (CQA) of cells. Previous studies have identified trigonelline (TRI), n-acetyl putrescine (NAP), aconitic acid (AA), and cytidine monophosphate (CMP) generated through CHO cell metabolism and verified their negative impacts on growth and antibody production. In this approach, a genetic engineering strategy was developed to control downstream accumulation of inhibitory metabolites. The study successfully identified four different metabolic genes in CHO cells, including Cat (nicotinate and nicotinamide metabolism) to control the generation of TRI, Got1 and Hoga1 (proline metabolism) to control the generation of NAP, Got1 (TCA cycle) to control the generation of AA, and Slc35a1 (n-glycan biosynthesis) to control the generation of CMP. Each target gene-of-interest (GOI) was cloned from CHO genomic library, inserted into linearized vector plasmid, and subsequently transfected into cells. CQA of the bioprocess realized 22-30% increase in peak cell density, 16-22% increase overall IVCD, with an improving growth rate during cellular expansion phase when comparing engineered cells against control cells. The study also conducted a follow-up quadruple transfection study where all four GOIs were co-transfected into cells at ¼ of the total DNA concentration per GOI. An increase in cellular performance was also realized, as increases in peak VCD (17% increase), cumulative IVCD (17% increase), and growth rate were achieved. Both studies also found higher IgG1 antibody synthesis when cell metabolism was better regulated, as the studies measured 4% to 40% titer increase across all engineered cells when compared against control cells. The study also measured higher levels of G1F and G2F glycans with decreased level of G0F across all transfected cells, further indicating improvement in bioprocess, as cells were able to produce a higher fraction of semi-complex and complex versus simple glycoforms. Further investigation revealed that Cat and Slc35a1 exhibited comparable expression levels in the MG condition to their single-gene conditions (within 1% and 10% difference, respectively), corresponding to modest titer improvements closest to the control. These findings suggest that when all four genes are co-expressed, Cat and Got1 may act as rate-limiting factors influencing both cellular phenotypes and titer production. In both studies, the concentrations of downstream metabolic inhibitors were measured to be significantly decreased when comparing engineered cells against control cells, further demonstrating that overexpression of genes to re-allocate metabolic fluxes away from synthesizing toxic by-products can significa","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13572490/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147982496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Interpretable multitask model for clinical pathology image prediction and interpretation.","authors":"Qitao Chen, Zhe Wang, Xia Lin, Yuying Shi, Botao Xu, Jie Chai, Tao Zhang, Cheng Wang","doi":"10.1038/s41540-026-00744-w","DOIUrl":"10.1038/s41540-026-00744-w","url":null,"abstract":"<p><p>Deep learning (DL)-based pathological image modelling and analysis approaches offer transformative potential for early cancer diagnostics, yet limited sample sizes and a lack of interpretability often hinder efficient clinical translation. Here, we present the interpretable Multi-Task Digital Pathology Model (iMDPath), an end-to-end, highly explainable multi-task deep learning framework that simultaneously addresses these challenges by integrating data augmentation, diagnostic prediction, and visualization of pathological image features. The iMDPath comprises three modules: Augmentation (iMDPath-Aug), Prediction (iMDPath-Pred), and Visualization (iMDPath-Vis). iMDPath-Aug incorporates a vector-quantized variational autoencoder (VQ-VAE) for enhanced data augmentation, capturing essential pathological features from limited datasets. A Swin Transformer-Based (Swin-B) predictor in the iMDPath-Pred module leverages the augmented data to achieve better performance than patch-level and foundation-model-based encoders such as InceptionV3 and Phikon across six diverse cancer pathology datasets, including gastric, breast, lung, and colorectal cancer. Finally, iMDPath-Vis, a novel visualization module combining the full gradient (FullGrad) and occlusion sensitivity analysis, provides pathologists with actionable insights by highlighting the specific tissue regions driving model predictions. Overall, iMDPath not only surpasses existing methods in diagnostic accuracy, sensitivity, and generalization across these datasets, but also offers a transparent and interpretable AI solution for precision oncology, paving the way for more reliable and efficient clinical decision-making.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13454177/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147974246","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Control of genes by self-organizing multicellular interaction networks.","authors":"Kyle R Allison","doi":"10.1038/s41540-026-00745-9","DOIUrl":"10.1038/s41540-026-00745-9","url":null,"abstract":"<p><p>Multicellular self-organization drives development in biological organisms, yet a comprehensive theory is lacking as basic properties of cells can complicate common approaches. Framing such properties by dynamic graphs led to new theoretical propositions for multicellular self-organization in Escherichia coli. Here, corresponding ideas are developed from biologically-general first principles. The resulting perspective could aid both experimental and computational approaches to multicellular biology as well as efforts to control and engineer it.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13448638/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147974259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Supervised machine learning identifies impaired mitochondrial quality control in β-cells with development of type 2 diabetes.","authors":"Mirza Muhammad Fahd Qadir, Charles Dana, Madeleine Pittigher, Paul Mauvais-Jarvis, Nazmul Haque, Theodore Dos Santos, Patrick E MacDonald, Kyle J Gaulton, Franck Mauvais-Jarvis","doi":"10.1038/s41540-026-00742-y","DOIUrl":"10.1038/s41540-026-00742-y","url":null,"abstract":"<p><p>Defining molecular pathways driving β-cell failure in type 2 diabetes (T2D) is challenging given donor heterogeneity. We developed an interpretable machine learning framework coupling sparse rule-based classification, pathway constrained modeling, and mitochondrial fitness stratification, applied to single-cell RNAseq from 52 human islet donors. A 50-gene classifier predicted T2D at single-cell resolution, outperforming ensemble models, with donor-level scores correlating with HbA1c. We identified a resilient non-diabetic (ND) β-cell subtype with preserved β-cell identity, while T2D β-cell subtypes showed cellular stress and suppressed oxidative phosphorylation. Mitophagy emerged as the dominant cellular pathway, with PINK1, BNIP3, and FUNDC1 as predictors. At the donor level, PINK1 expression decreased with T2D score and correlated with sex‑specific mitophagy patterns. We developed a mitochondrial fitness index (MFI, R² = 0.934) integrating mitophagy, proteostasis, biogenesis, and respiration, identifying PINK1, SQSTM1, PRKN, and BNIP3 as top T2D contributors. Interpretable machine learning revealed mitophagy as central to β-cell metabolic fitness.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13447972/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147974199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Emergence of Sprouty role from the signaling interplay of DUSP1 and DUSP4 interactions in NSCLC pathogenesis.","authors":"Manasi Tukrul, Pooja Gulhane, Shailza Singh","doi":"10.1038/s41540-026-00740-0","DOIUrl":"10.1038/s41540-026-00740-0","url":null,"abstract":"<p><p>Non-small cell lung cancer (NSCLC), the predominant subtype of lung cancer, remains a major global health challenge due to its complex etiology and resistance to therapy. Aberrant signaling through receptor tyrosine kinases (RTKs) and mitogen-activated protein kinases (MAPKs) drives tumor progression and immune modulation, underscoring the need for systems-level investigation. In this study, we employed mathematical modeling and experimental validation to dissect the dynamic interplay of EGFR, FGFR, and TNFR signaling in NSCLC cells and macrophages. Our flux-based sensitivity analysis revealed persistent pathway biases, highlighting DUSP1, DUSP4, and SPRY2 as critical high-flux regulators within the MAPK and PI3K/AKT cascades. Crosstalk analysis demonstrated that these proteins form an interconnected signaling axis influencing cytokine production (IL6, IL10, IL12, and TNF-α) and macrophage-mediated responses in the tumor microenvironment. Functional assays confirmed that silencing SPRY2 suppressed DUSP1 and DUSP4 expression, impaired cell migration, and induced G2/M arrest, establishing a coordinated regulatory network essential for NSCLC progression. Together, our findings integrate computational modeling with molecular validation to identify DUSP1, DUSP4, and SPRY2 as pivotal drivers of oncogenic signaling and immune regulation. These proteins represent promising therapeutic targets whose inhibition may disrupt tumor-promoting feedback loops and attenuate NSCLC growth and survival.</p>","PeriodicalId":19345,"journal":{"name":"NPJ Systems Biology and Applications","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2026-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13424125/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147949027","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}