Biology subjects

Bifarin, O. O.

Publications and source records attributed to Bifarin, O. O..

4 recordsLinked to original sources

bioRxiv2023

Automated machine learning and explainable AI (AutoML-XAI) for metabolomics: improving cancer diagnostics

MotivationMetabolomics 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. ResultsWe tested our approach on two datasets: renal cell carcinoma (RCC) urine metabolomics and ovarian cancer (OC) serum metabolomics. AutoML, using auto-sklearn, surpassed standalone ML algorithms such as SVM and random forest in differentiating between RCC and healthy controls, as well as OC patients and those with other gynecological cancers (Non-OC). Autosklearn employed a mix of algorithms and ensemble techniques, yielding a superior performance (AUC of 0.97 for RCC and 0.85 for OC). 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. Availabilityhttps://github.com/obifarin/automl-xai-metabolomics Contact: facundo.fernandez@chemistry.gatech.edu

bioinformatics↗

bioRxiv2023

Serum Lipidome Profiling Reveals a Distinct Signature of Ovarian Cancer in Korean Women

Distinguishing ovarian cancer (OC) from other gynecological malignancies remains a critical unmet medical need with significant implications for patient survival. However, non-specific symptoms along with our lack of understanding of OC pathogenesis hinder its diagnosis, preventing many women from receiving appropriate medical assistance. Accumulating evidence suggests a link between OC and deregulated lipid metabolism. Most studies, however, are limited by small sample size, particularly for early-stage cases. Furthermore, racial/ethnic differences in OC survival and incidence have been reported, yet most of the studies consist largely of non-Hispanic white women or women with European ancestry. Studies of more diverse racial/ethnic populations are needed to make OC diagnosis and prevention more inclusive. Here, we profiled the serum lipidome of 208 OC, including 93 patients with early-stage OC, and 117 non-OC (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 approaches. Results show that lipidome alterations unique to OC 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 OC in comparison to other gynecological diseases. Machine learning methods selected a panel of 17 lipids that discriminated OC from non-OC cases with an AUC of 0.85 for an independent test set. This study provides a systemic analysis of lipidome alterations in human OC, specifically in Korean women, emphasizing the potential of circulating lipids in distinguishing OC from non-OC conditions.

cancer biology↗

bioRxiv2023

Machine Learning Reveals Lipidome Remodeling Dynamics in a Mouse Model of Ovarian Cancer

Ovarian cancer (OC) is one of the deadliest cancers affecting the female reproductive system. It may present little or no symptoms at the early stages, and typically unspecific symptoms at later stages. High-grade serous ovarian cancer (HGSC) is the subtype responsible for most ovarian cancer deaths. However, very little is known about the metabolic course of this disease, particularly in its early stages. In this longitudinal study, we examined the temporal course of serum lipidome changes using a robust HGSC mouse model and machine learning data analysis. Early progression of HGSC was marked by increased levels of phosphatidylcholines and phosphatidylethanolamines. In contrast, later stages featured more diverse lipids alterations, including fatty acids and their derivatives, triglycerides, ceramides, hexosylceramides, sphingomyelins, lysophosphatidylcholines, and phosphatidylinositols. These alterations underscored unique perturbations in cell membrane stability, proliferation, and survival during cancer development and progression, offering potential targets for early detection and prognosis of human ovarian cancer. TeaserTime-resolved lipidome remodeling in an ovarian cancer model is studied through lipidomics and machine learning.

cancer biology↗

bioRxiv2022

Interpretable machine learning with tree-based Shapley additive explanations: application to metabolomics datasets for binary classification

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 models 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 a tree-based model was interpreted using Tree SHAP. The results show that SHAP has a more explanation depth than PLS-DAs VIP, making it a powerful method for rationalizing machine learning predictions from metabolomics studies.

bioinformatics↗