Machine learning (ML) models are used in clinical metabolomics studies most notably for biomarker discoveries, to identify metabolites that discriminate between a case and control group. To improve understanding of the underlying biomedical problem and to bolster confidence in these discoveries, model interpretability is germane. In metabolomics, partial least square discriminant analysis (PLS-DA) and its variants are widely used, partly due to the model’s interpretability with the Variable Influence in Projection (VIP) scores, a global interpretable method. Herein, Tree-based Shapley Additive explanations (SHAP), an interpretable ML method grounded in game theory, was used to explain ML models with local explanation properties. In this study, ML experiments (binary classification) were conducted for three published metabolomics datasets using PLS-DA, random forests, gradient boosting, and extreme gradient boosting (XGBoost). Using one of the datasets, PLS-DA model was explained using VIP scores, while one of the best-performing models, a random forest model, was interpreted using Tree SHAP. The results show that SHAP has a more explanation depth than PLS-DA’s VIP, making it a powerful method for rationalizing machine learning predictions from metabolomics studies.

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

Urine metabolomics profiling has potential for non-invasive RCC staging, in addition to providing metabolic insights into disease progression. In this study, we utilized liquid chromatography-mass spectrometry (LC-MS), nuclear magnetic resonance (NMR), and machine learning (ML) for the discovery of urine metabolites associated with RCC progression. Two machine learning questions were posed in the study: Binary classification into early RCC (stage I and II) and advanced RCC stages (stage III and IV), and RCC tumor size estimation through regression analysis. A total of 82 RCC patients with known tumor size and metabolomic measurements were used for the regression task, and 70 RCC patients with complete tumor-nodes-metastasis (TNM) staging information were used for the classification tasks under ten-fold cross-validation conditions. A voting ensemble regression model consisting of elastic net, ridge, and support vector regressor predicted RCC tumor size with a R² value of 0.58. A voting classifier model consisting of random forest, support vector machines, logistic regression, and adaptive boosting yielded an AUC of 0.96 and an accuracy of 87%. Some identified metabolites associated with renal cell carcinoma progression included 4-guanidinobutanoic acid, 7-aminomethyl-7-carbaguanine, 3-hydroxyanthranilic acid, lysyl-glycine, glycine, citrate, and pyruvate. Overall, we identified a urine metabolic phenotype associated with renal cell carcinoma stage, exploring the promise of a urine-based metabolomic assay for staging this disease.

Background: Distinguishing ovarian cancer from other gynecological malignancies is crucial for patient survival yet hindered by non-specific symptoms and limited understanding of ovarian cancer pathogenesis. Accumulating evidence suggests a link between ovarian cancer and deregulated lipid metabolism. Most studies have small sample sizes, especially for early-stage cases, and lack racial/ethnic diversity, necessitating more inclusive research for improved ovarian cancer diagnosis and prevention. Methods: Here, we profiled the serum lipidome of 208 ovarian cancer, including 93 early-stage patients with ovarian cancer and 117 nonovarian cancer (other gynecological malignancies) patients of Korean descent. Serum samples were analyzed with a high-coverage liquid chromatography high-resolution mass spectrometry platform, and lipidome alterations were investigated via statistical and machine learning (ML) approaches. Results: We found that lipidome alterations unique to ovarian cancer were present in Korean women as early as when the cancer is localized, and those changes increase in magnitude as the diseases progresses. Analysis of relative lipid abundances revealed specific patterns for various lipid classes, with most classes showing decreased abundance in ovarian cancer in comparison with other gynecological diseases. ML methods selected a panel of 17 lipids that discriminated ovarian cancer from nonovarian cancer cases with an AUC value of 0.85 for an independent test set. Conclusions: This study provides a systemic analysis of lipidome alterations in human ovarian cancer, specifically in Korean women. Impact: Here, we show the potential of circulating lipids in distinguishing ovarian cancer from nonovarian cancer conditions.

We present a comprehensive map of the metabolomics research landscape, synthesizing insights from over 80,000 publications. Using PubMedBERT, we transformed abstracts into 768-dimensional embeddings that capture the nuanced thematic structure of the field. Dimensionality reduction with t-SNE revealed distinct clusters corresponding to key domains, such as analytical chemistry, plant biology, pharmacology, and clinical diagnostics. In addition, a neural topic modeling pipeline refined with GPT-4o mini reclassified the corpus into 20 distinct topics─ranging from "Plant Stress Response Mechanisms" and "NMR Spectroscopy Innovations" to "COVID-19 Metabolomic and Immune Responses." Temporal analyses further highlight trends including the rise of deep learning methods post-2015 and a continued focus on biomarker discovery. Integration of metadata such as publication statistics and sample sizes provides additional context to these evolving research dynamics. An interactive web application (https://metascape.streamlit.app/) enables the dynamic exploration of these insights. Overall, this study offers a robust framework for literature synthesis that empowers researchers, clinicians, and policymakers to identify emerging research trajectories and address critical challenges in metabolomics while also sharing our perspectives on key trends shaping the field.