Longitudinal metabolomics data analysis informed bymechanistic models
Li, L.; Hoefsloot, H. C.; Bakker, B. M.; Horner, D.; Rasmussen, M. A.; Smilde, A.; Acar, E.
Show abstract
MotivationMetabolomics measurements are noisy, often characterized by a small sample size and missing entries. While data-driven methods have shown promise in terms of analyzing metabolomics data, e.g., revealing biomarkers of various phenotypes, metabolomics data analysis can significantly benefit from incorporating prior information about metabolic mechanisms. In this paper, we introduce a novel data analysis approach where data-driven methods are guided by prior information through joint analysis of simulated data generated using a human metabolic model and real metabolomics measurements. ResultsWe arrange time-resolved metabolomics measurements of plasma samples collected during a meal challenge test from the COPSAC2000 cohort as a third-order tensor: subjects by metabolites by time samples. Simulated challenge test data generated using a human whole-body metabolic model is also arranged as a third-order tensor: virtual subjects by metabolites by time samples. Real and simulated data sets are coupled in the metabolites mode and jointly analyzed using coupled tensor factorizations to reveal the underlying patterns. Our experiments demonstrate that joint analysis of simulated and real data has a better performance in terms of pattern discovery achieving higher correlations with a BMI (body mass index)-related phenotype compared to the analysis of only real data in males while in females, the performance is comparable. We also demonstrate the advantages of such a joint analysis approach in the presence of incomplete measurements and its limitations in the presence of wrong prior information. AvailabilityThe code for joint analysis of real and simulated metabolomics data sets is released as a GitHub repository. Simulated data can also be accessed using the GitHub repo. Real measurements of plasma samples are not publicly available. Data may be shared by COPSAC through a collaboration agreement. Data access requests should be directed to Morten A. Rasmussen (morten.arendt@dbac.dk).
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- A Generalized Higher-order Correlation Analysis Framework for Multi-Omics Network Inference 95%
- PaIRKAT: A pathway integrated regression-based kernel association test with applications to metabolomics and COPD phenotypes 94%
- RCFGL: Rapid Condition adaptive Fused Graphical Lasso and application to modeling brain region co-expression networks 94%
Similar papers in this journal
- PiDeel: Pathway-informed deep learning model for survivalanalysis and pathological classification of gliomas 94%
- NetSHy: Network Summarization via a Hybrid Approach Leveraging Topological Properties 94%
- Exploring High-Dimensional Biological Data with Sparse Contrastive Principal Component Analysis 94%
Similar papers in this journal
Similar papers in this journal
- Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping 94%
- Molecular Group and Correlation Guided Structural Learning for Multi-Phenotype Prediction 94%
- VBayesMM: Variational Bayesian neural network to prioritize important relationships of high-dimensional microbiome multiomics data 93%
Similar papers in this journal
- ACMTF-R: supervised multi-omics data integration uncovering shared and distinct outcome-associated variation 96%
- Interpretable machine learning with tree-based Shapley additive explanations: application to metabolomics datasets for binary classification 93%
- Features extracted using tensor decomposition reflect the biological features of the temporal patterns of human blood multimodal metabolome 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.