Metabolomics and chemometrics: The next-generation analytical toolkit for the evaluation of food quality and authenticity

Trends in Food Science & Technology, Год журнала: 2024, Номер 147, С. 104481 - 104481

Опубликована: Апрель 7, 2024

The advances in NMR and mass spectrometry metabolomics allows a comprehensive profiling of foods, potentially covering geographical origin, authenticity, quality integrity issues. However, mining specific effects within the corresponding datasets is challenging due to presence set interacting factors that finally determine signatures. This review provides an overview different approaches used food then focusing on chemometric for data interpretation. In particular, interpretation hierarchically presented, starting from unsupervised (PCA, hierarchical clusters) supervised multivariate statistics like OPLS AMOPLS multiblock ANOVA discriminant approaches. Finally, machine learning Artificial Neural Networks are discussed as novel emerging tool support Tailored advisable, rather than unique solutions, with naively provide qualitative recognition patterns, modelling markers identification. Nonetheless, approach able interpretate complex

Язык: Английский

---

Artificial intelligence in metabolomics: a current review

TrAC Trends in Analytical Chemistry, Год журнала: 2024, Номер 178, С. 117852 - 117852

Опубликована: Июль 3, 2024

Язык: Английский

---

Integrating Molecular Perspectives: Strategies for Comprehensive Multi-Omics Integrative Data Analysis and Machine Learning Applications in Transcriptomics, Proteomics, and Metabolomics

Biology, Год журнала: 2024, Номер 13(11), С. 848 - 848

Опубликована: Окт. 22, 2024

With the advent of high-throughput technologies, field omics has made significant strides in characterizing biological systems at various levels complexity. Transcriptomics, proteomics, and metabolomics are three most widely used each providing unique insights into different layers a system. However, analyzing data set separately may not provide comprehensive understanding subject under study. Therefore, integrating multi-omics become increasingly important bioinformatics research. In this article, we review strategies for transcriptomics, data, including co-expression analysis, metabolite-gene networks, constraint-based models, pathway enrichment interactome analysis. We discuss combined integration approaches, correlation-based strategies, machine learning techniques that utilize one or more types data. By presenting these methods, aim to researchers with better how integrate gain view system, facilitating identification complex patterns interactions might be missed by single-omics analyses.

Язык: Английский

---

Artificial intelligence in the management of metabolic disorders: a comprehensive review

Journal of Endocrinological Investigation, Год журнала: 2025, Номер unknown

Опубликована: Фев. 19, 2025

Язык: Английский

---

Explainable Artificial Intelligence Paves the Way in Precision Diagnostics and Biomarker Discovery for the Subclass of Diabetic Retinopathy in Type 2 Diabetics

Metabolites, Год журнала: 2023, Номер 13(12), С. 1204 - 1204

Опубликована: Дек. 18, 2023

Diabetic retinopathy (DR), a common ocular microvascular complication of diabetes, contributes significantly to diabetes-related vision loss. This study addresses the imperative need for early diagnosis DR and precise treatment strategies based on explainable artificial intelligence (XAI) framework. The integrated clinical, biochemical, metabolomic biomarkers associated with following classes: non-DR (NDR), non-proliferative diabetic (NPDR), proliferative (PDR) in type 2 diabetes (T2D) patients. To create machine learning (ML) models, 10% data was divided into validation sets 90% discovery sets. dataset used hyperparameter optimization feature selection stages, while measure performance models. A 10-fold cross-validation technique evaluate ML Biomarker performed using minimum redundancy maximum relevance (mRMR), Boruta, boosting (EBM). predictive proposed framework compares results eXtreme Gradient Boosting (XGBoost), natural gradient probabilistic prediction (NGBoost), EBM models determining subclass. hyperparameters were optimized Bayesian optimization. Combining XGBoost, optimal model achieved (91.25 ± 1.88) % accuracy, (89.33 1.80) precision, (91.24 1.67) recall, (89.37 1.52) F1-Score, (97.00 0.25) area under ROC curve (AUROC). According explanation, six most important course tryptophan (Trp), phosphatidylcholine diacyl C42:2 (PC.aa.C42.2), butyrylcarnitine (C4), tyrosine (Tyr), hexadecanoyl carnitine (C16) total dimethylarginine (DMA). identified may provide better understanding progression DR, paving way more cost-effective diagnostic strategies.

Язык: Английский

---

Automated Machine Learning and Explainable AI (AutoML-XAI) for Metabolomics: Improving Cancer Diagnostics

Journal of the American Society for Mass Spectrometry, Год журнала: 2024, Номер 35(6), С. 1089 - 1100

Опубликована: Май 1, 2024

Metabolomics generates complex data necessitating advanced computational methods for generating biological insight. While machine learning (ML) is promising, the challenges of selecting best algorithms and tuning hyperparameters, particularly nonexperts, remain. Automated (AutoML) can streamline this process; however, issue interpretability could persist. This research introduces a unified pipeline that combines AutoML with explainable AI (XAI) techniques to optimize metabolomics analysis. We tested our approach on two sets: renal cell carcinoma (RCC) urine ovarian cancer (OC) serum metabolomics. AutoML, using Auto-sklearn, surpassed standalone ML like SVM k-Nearest Neighbors in differentiating between RCC healthy controls, as well OC patients those other gynecological cancers. The effectiveness Auto-sklearn highlighted by its AUC scores 0.97 0.85 OC, obtained from unseen test sets. Importantly, most metrics considered, demonstrated better classification performance, leveraging mix ensemble techniques. Shapley Additive Explanations (SHAP) provided global ranking feature importance, identifying dibutylamine ganglioside GM(d34:1) top discriminative metabolites respectively. Waterfall plots offered local explanations illustrating influence each metabolite individual predictions. Dependence spotlighted interactions, such connection hippuric acid one derivatives RCC, GM3(d34:1) GM3(18:1_16:0) hinting at potential mechanistic relationships. Through decision plots, detailed error analysis was conducted, contrasting importance correctly versus incorrectly classified samples. In essence, emphasizes harmonizing XAI, facilitating both simplified application improved science.

Язык: Английский

---

Machine learning for the advancement of genome-scale metabolic modeling

Biotechnology Advances, Год журнала: 2024, Номер 74, С. 108400 - 108400

Опубликована: Июнь 28, 2024

Язык: Английский

---

A Comprehensive Machine Learning Approach for COVID-19 Target Discovery in the Small-Molecule Metabolome

Metabolites, Год журнала: 2025, Номер 15(1), С. 44 - 44

Опубликована: Янв. 11, 2025

Background/Objectives: Respiratory viruses, including Influenza, RSV, and COVID-19, cause various respiratory infections. Distinguishing these viruses relies on diagnostic methods such as PCR testing. Challenges stem from overlapping symptoms the emergence of new strains. Advanced diagnostics are crucial for accurate detection effective management. This study leveraged nasopharyngeal metabolome data to predict virus scenarios control vs. Influenza A, all COVID-19 A/RSV. Method: We proposed a stacking-based ensemble technique, integrating top three best-performing ML models initial results enhance prediction accuracy by leveraging strengths multiple base learners. Key techniques feature ranking, standard scaling, SMOTE were used address class imbalances, thus enhancing model robustness. SHAP analysis identified metabolites influencing positive predictions, thereby providing valuable insights into markers. Results: Our approach not only outperformed existing but also revealed dominant features predicting Lysophosphatidylcholine acyl C18:2, Kynurenine, Phenylalanine, Valine, Tyrosine, Aspartic Acid (Asp). Conclusions: demonstrates effectiveness scenarios. The enhances accuracy, provides key markers, offers robust framework managing

Язык: Английский

---

Metabolomics and (craft) beers – recent advances

Food Research International, Год журнала: 2025, Номер 205, С. 116010 - 116010

Опубликована: Фев. 14, 2025

Язык: Английский

---

Predictive model using systemic inflammation markers to assess neoadjuvant chemotherapy efficacy in breast cancer

Frontiers in Oncology, Год журнала: 2025, Номер 15

Опубликована: Март 24, 2025

Background Pathological complete response (pCR) is an important indicator for evaluating the efficacy of neoadjuvant chemotherapy (NAC) in breast cancer. The role systemic inflammation markers predicting pCR and long-term prognosis cancer patients undergoing NAC remains controversial. purpose this study was to explore potential predictive prognostic value (NLR, PLR, LMR, NMR) clinicopathological characteristics receiving construct a prediction model based on these indicators. Methods A retrospective analysis conducted 209 who received at Nanjing Drum Tower Hospital between January 2010 March 2020. Independent sample t-tests, chi-square tests, logistic regression models were used evaluate correlation data, markers, pCR. Receiver operating characteristic (ROC) curves utilized determine optimal cut-off values NLR, LMR. Survival performed using Kaplan-Meier method log-rank test. constructed machine learning algorithms. Results Among patients, 29 achieved During median follow-up 68 months, 74 experienced local or distant metastasis, 56 died. Univariate showed that lymph node status, HER-2 LMR associated with ROC curve revealed 1.525, 113.620, 6.225, respectively. Multivariate indicated independent factors demonstrated prognosis. Machine algorithm identified random forest (RF) as Conclusion This had significant patients. RF provides simple cost-effective tool prediction, offering strong support clinical decision-making treatment aiding optimization individualized strategies.

Язык: Английский

---