Premium
Plant stress biomarkers from biosimulations: the Transcriptome‐To‐Metabolome ™ ( TTM ™ ) technology – effects of drought stress on rice
Author(s) -
Phelix C. F.,
Feltus F. A.
Publication year - 2015
Publication title -
plant biology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.871
H-Index - 87
eISSN - 1438-8677
pISSN - 1435-8603
DOI - 10.1111/plb.12221
Subject(s) - metabolome , transcriptome , biology , metabolomics , computational biology , lipidomics , phenomics , biomarker discovery , proteomics , bioinformatics , genetics , genomics , gene , gene expression , genome
Measuring biomarkers from plant tissue samples is challenging and expensive when the desire is to integrate transcriptomics, fluxomics, metabolomics, lipidomics, proteomics, physiomics and phenomics. We present a computational biology method where only the transcriptome needs to be measured and is used to derive a set of parameters for deterministic kinetic models of metabolic pathways. The technology is called Transcriptome‐To‐Metabolome ™ ( TTM ™ ) biosimulations, currently under commercial development, but available for non‐commercial use by researchers. The simulated results on metabolites of 30 primary and secondary metabolic pathways in rice ( Oryza sativa ) were used as the biomarkers to predict whether the transcriptome was from a plant that had been under drought conditions. The rice transcriptomes were accessed from public archives and each individual plant was simulated. This unique quality of the TTM ™ technology allows standard analyses on biomarker assessments, i.e . sensitivity, specificity, positive and negative predictive values, accuracy, receiver operator characteristics ( ROC ) curve and area under the ROC curve ( AUC ). Two validation methods were also used, the holdout and 10‐fold cross validations. Initially 17 metabolites were identified as candidate biomarkers based on either statistical significance on binary phenotype when compared with control samples or recognition from the literature. The top three biomarkers based on AUC were gibberellic acid 12 (0.89), trehalose (0.80) and sn1‐palmitate‐sn2‐oleic‐phosphatidylglycerol (0.70). Neither heat map analyses of transcriptomes nor all 300 metabolites clustered the stressed and control groups effectively. The TTM ™ technology allows the emergent properties of the integrated system to generate unique and useful ‘Omics’ information.