Limits...
Detection and interpretation of metabolite-transcript coresponses using combined profiling data.

Redestig H, Costa IG - Bioinformatics (2011)

Bottom Line: Our aim was to predict pathway comemberships between metabolites and genes based on their coresponses to applied stress.Therefore, we propose a supervised method that applies pathway information to summarize similarity statistics to a consensus statistic that is more informative than any of the single measures.Using four combined profiling datasets, we show that comembership between metabolites and genes can be predicted for numerous KEGG pathways; this opens opportunities for the detection of transcriptionally regulated pathways and novel metabolically related genes.

View Article: PubMed Central - PubMed

Affiliation: RIKEN Plant Science Center, Yokohama, Japan. henning@psc.riken.jp

ABSTRACT

Motivation: Studying the interplay between gene expression and metabolite levels can yield important information on the physiology of stress responses and adaptation strategies. Performing transcriptomics and metabolomics in parallel during time-series experiments represents a systematic way to gain such information. Several combined profiling datasets have been added to the public domain and they form a valuable resource for hypothesis generating studies. Unfortunately, detecting coresponses between transcript levels and metabolite abundances is non-trivial: they cannot be assumed to overlap directly with underlying biochemical pathways and they may be subject to time delays and obscured by considerable noise.

Results: Our aim was to predict pathway comemberships between metabolites and genes based on their coresponses to applied stress. We found that in the presence of strong noise and time-shifted responses, a hidden Markov model-based similarity outperforms the simpler Pearson correlation but performs comparably or worse in their absence. Therefore, we propose a supervised method that applies pathway information to summarize similarity statistics to a consensus statistic that is more informative than any of the single measures. Using four combined profiling datasets, we show that comembership between metabolites and genes can be predicted for numerous KEGG pathways; this opens opportunities for the detection of transcriptionally regulated pathways and novel metabolically related genes.

Availability: A command-line software tool is available at http://www.cin.ufpe.br/~igcf/Metabolites.

Contact: henning@psc.riken.jp; igcf@cin.ufpe.br

Show MeSH

Related in: MedlinePlus

AUC for the classification of cognate pairs of TT, MM and MT (maximum three steps in the metabolic pathway) from non-cognate pairs (minimum of eight steps). The box shows the expected distribution of the AUC under H0 as estimated by randomization and the dot indicates the observed AUC. While Pearson correlation shows higher coresponses between MM pairs in CO2 and sulfur; and with TT pairs in CO2, none of the considered that MT coresponse measures are informative for performing this classification.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3117345&req=5

Figure 5: AUC for the classification of cognate pairs of TT, MM and MT (maximum three steps in the metabolic pathway) from non-cognate pairs (minimum of eight steps). The box shows the expected distribution of the AUC under H0 as estimated by randomization and the dot indicates the observed AUC. While Pearson correlation shows higher coresponses between MM pairs in CO2 and sulfur; and with TT pairs in CO2, none of the considered that MT coresponse measures are informative for performing this classification.

Mentions: A good measure for coresponses should give higher values for truly associated than for unrelated MT pairs. A direct way to determine whether a coresponse measure lives up to this expectation is to compare the values for close neighbors in the underlying metabolic pathways with values for distant pairs. To this end, we generated an undirected aggregated bipartite graph for all A.thaliana-specific KEGG pathways by connecting metabolites to enzymes that catalyze their formation or breakdown. Similar to the methodology of Walther et al. (2010), we used this graph to classify TT, MM and MT pairs at a maximum distance of three steps as cognate pairs and metabolites at a minimum distance of eight steps as non-cognates. We then evaluate the performance of the different coresponse measures for discriminating cognate from non-cognate pairs by calculating the area under the ROC curve (AUC) (Fig. 5). Notably, with all AUC values close to 0.5 as expected from a random classifier for all MT pairs, none appeared informative for this classification, indicating that expression/abundance profiles contain little information about immediate associations in the metabolic reaction graph.Fig. 5.


Detection and interpretation of metabolite-transcript coresponses using combined profiling data.

Redestig H, Costa IG - Bioinformatics (2011)

AUC for the classification of cognate pairs of TT, MM and MT (maximum three steps in the metabolic pathway) from non-cognate pairs (minimum of eight steps). The box shows the expected distribution of the AUC under H0 as estimated by randomization and the dot indicates the observed AUC. While Pearson correlation shows higher coresponses between MM pairs in CO2 and sulfur; and with TT pairs in CO2, none of the considered that MT coresponse measures are informative for performing this classification.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3117345&req=5

Figure 5: AUC for the classification of cognate pairs of TT, MM and MT (maximum three steps in the metabolic pathway) from non-cognate pairs (minimum of eight steps). The box shows the expected distribution of the AUC under H0 as estimated by randomization and the dot indicates the observed AUC. While Pearson correlation shows higher coresponses between MM pairs in CO2 and sulfur; and with TT pairs in CO2, none of the considered that MT coresponse measures are informative for performing this classification.
Mentions: A good measure for coresponses should give higher values for truly associated than for unrelated MT pairs. A direct way to determine whether a coresponse measure lives up to this expectation is to compare the values for close neighbors in the underlying metabolic pathways with values for distant pairs. To this end, we generated an undirected aggregated bipartite graph for all A.thaliana-specific KEGG pathways by connecting metabolites to enzymes that catalyze their formation or breakdown. Similar to the methodology of Walther et al. (2010), we used this graph to classify TT, MM and MT pairs at a maximum distance of three steps as cognate pairs and metabolites at a minimum distance of eight steps as non-cognates. We then evaluate the performance of the different coresponse measures for discriminating cognate from non-cognate pairs by calculating the area under the ROC curve (AUC) (Fig. 5). Notably, with all AUC values close to 0.5 as expected from a random classifier for all MT pairs, none appeared informative for this classification, indicating that expression/abundance profiles contain little information about immediate associations in the metabolic reaction graph.Fig. 5.

Bottom Line: Our aim was to predict pathway comemberships between metabolites and genes based on their coresponses to applied stress.Therefore, we propose a supervised method that applies pathway information to summarize similarity statistics to a consensus statistic that is more informative than any of the single measures.Using four combined profiling datasets, we show that comembership between metabolites and genes can be predicted for numerous KEGG pathways; this opens opportunities for the detection of transcriptionally regulated pathways and novel metabolically related genes.

View Article: PubMed Central - PubMed

Affiliation: RIKEN Plant Science Center, Yokohama, Japan. henning@psc.riken.jp

ABSTRACT

Motivation: Studying the interplay between gene expression and metabolite levels can yield important information on the physiology of stress responses and adaptation strategies. Performing transcriptomics and metabolomics in parallel during time-series experiments represents a systematic way to gain such information. Several combined profiling datasets have been added to the public domain and they form a valuable resource for hypothesis generating studies. Unfortunately, detecting coresponses between transcript levels and metabolite abundances is non-trivial: they cannot be assumed to overlap directly with underlying biochemical pathways and they may be subject to time delays and obscured by considerable noise.

Results: Our aim was to predict pathway comemberships between metabolites and genes based on their coresponses to applied stress. We found that in the presence of strong noise and time-shifted responses, a hidden Markov model-based similarity outperforms the simpler Pearson correlation but performs comparably or worse in their absence. Therefore, we propose a supervised method that applies pathway information to summarize similarity statistics to a consensus statistic that is more informative than any of the single measures. Using four combined profiling datasets, we show that comembership between metabolites and genes can be predicted for numerous KEGG pathways; this opens opportunities for the detection of transcriptionally regulated pathways and novel metabolically related genes.

Availability: A command-line software tool is available at http://www.cin.ufpe.br/~igcf/Metabolites.

Contact: henning@psc.riken.jp; igcf@cin.ufpe.br

Show MeSH
Related in: MedlinePlus