Predicting interpretability of metabolome models based on behavior, putative identity, and biological relevance of explanatory signals
- 3 October 2006
- journal article
- research article
- Published by Proceedings of the National Academy of Sciences in Proceedings of the National Academy of Sciences of the United States of America
- Vol. 103 (40), 14865-14870
- https://doi.org/10.1073/pnas.0605152103
Abstract
Powerful algorithms are required to deal with the dimensionality of metabolomics data. Although many achieve high classification accuracy, the models they generate have limited value unless it can be demonstrated that they are reproducible and statistically relevant to the biological problem under investigation. Random forest (RF) generates models, without any requirement for dimensionality reduction or feature selection, in which individual variables are ranked for significance and displayed in an explicit manner. In metabolome fingerprinting by mass spectrometry, each metabolite can be represented by signals at several m/z. Exploiting a prior understanding of expected biochemical differences between sample classes, we aimed to develop meaningful metrics relevant to the significance both of the overall RF model and individual, potentially explanatory, signals. Pair-wise comparison of related plant genotypes with strong phenotypic differences demonstrated that robust models are not only reproducible but also logically structured, highlighting correlated m/z derived from just a small number of explanatory metabolites reflecting the biological differences between sample classes. RF models were also generated by using groupings of samples known to be increasingly phenotypically similar. Although classification accuracy was often reasonable, we demonstrated reproducibly in both Arabidopsis and potato a performance threshold based on margin statistics beyond which such models showed little structure indicative of either generalizibility or further biological interpretability. In a multiclass problem using 25 Arabidopsis genotypes, despite the complicating effects of ecotype background and secondary metabolome perturbations common to several mutations, the ranking of metabolome signals by RF provided scope for deeper interpretability.Keywords
This publication has 38 references indexed in Scilit:
- Thousands of samples are needed to generate a robust gene list for predicting outcome in cancerProceedings of the National Academy of Sciences of the United States of America, 2006
- Modelling of classification rules on metabolic patterns including machine learning and expert knowledgeJournal of Biomedical Informatics, 2005
- Potential of metabolomics as a functional genomics toolTrends in Plant Science, 2004
- Supervised machine learning techniques for the classification of metabolic disorders in newbornsBioinformatics, 2004
- Identification of optimal classification functions for biological sample and state discrimination from metabolic profiling dataBioinformatics, 2004
- Classification and identification of Arabidopsis cell wall mutants using Fourier‐Transform InfraRed (FT‐IR) microspectroscopyThe Plant Journal, 2003
- Nontargeted Metabolome Analysis by Use of Fourier Transform Ion Cyclotron Mass SpectrometryOMICS: A Journal of Integrative Biology, 2002
- Genomic Computing. Explanatory Analysis of Plant Expression Profiling Data Using Machine LearningPlant Physiology, 2001
- Metabolic Profiling Allows Comprehensive Phenotyping of Genetically or Environmentally Modified Plant SystemsPlant Cell, 2001
- Boosting the margin: a new explanation for the effectiveness of voting methodsThe Annals of Statistics, 1998