Сравнение методов

	Анализ метаболомики временных рядов	Анализ метаболомики с использованием машинного обучения
Область	Биоинформатика	Биоинформатика
Семейство	Process / pipeline	Process / pipeline
Год появления	2000s–2010s	2000s–2010s (rapid adoption 2015–present)
Автор метода	Developed from general metabolomics workflows; longitudinal extensions pioneered by A. K. Smilde, R. Bino, and colleagues	Convergent development; foundational reviews by Liebal et al. (2020) and earlier multivariate metabolomics work by Trygg, Holmes, and Nicholson
Тип	Quantitative longitudinal omics pipeline	Integrative analytical pipeline
Основополагающий источник	Smilde, A. K., van der Werf, M. J., Bijlsma, S., van der Werff-van der Vat, B. J. C., & Jellema, R. H. (2005). Fusion of mass spectrometry-based metabolomics data. Analytical Chemistry, 77(20), 6729–6736.	Liebal, U. W., Phan, A. N. T., Sudhakar, M., Raman, K., & Blank, L. M. (2020). Machine learning applications for mass spectrometry-based metabolomics. Metabolites, 10(6), 243.
Другие названия	longitudinal metabolomics, dynamic metabolomics, temporal metabolome profiling, kinetic metabolomics	ML-metabolomics, chemoinformatics ML, metabolite profiling with machine learning, ML-driven metabolic profiling
Связанные	6	1
Сводка	Time-series metabolomics analysis profiles small-molecule metabolites from biological samples collected at multiple, ordered time points, enabling researchers to capture the dynamic flux of metabolic pathways in response to stimuli, disease progression, drug treatment, or developmental change. By integrating longitudinal statistical models with standard metabolomics preprocessing, the approach goes beyond a static metabolic snapshot to reveal how, when, and in what sequence metabolic responses unfold.	Machine learning-assisted metabolomics analysis is an integrative bioinformatics pipeline that couples untargeted or targeted metabolite profiling — via mass spectrometry or NMR — with supervised and unsupervised ML algorithms to discover biomarkers, classify phenotypes, and model metabolic states. By handling the extreme dimensionality and collinearity inherent in metabolomics datasets (hundreds to thousands of features, tens to hundreds of samples), ML methods such as random forests, support vector machines, and neural networks extract biologically interpretable patterns that classical univariate statistics routinely miss.