Limits...
Projection to latent pathways (PLP): a constrained projection to latent variables (PLS) method for elementary flux modes discrimination.

Ferreira AR, Dias JM, Teixeira AP, Carinhas N, Portela RM, Isidro IA, von Stosch M, Oliveira R - BMC Syst Biol (2011)

Bottom Line: PLP has two concomitant goals: (1) maximisation of correlation between EFM weighting factors and measured envirome data and (2) minimisation of redundancy by eliminating EFM with low correlation with the envirome.Overall, our results demonstrate that PLP slightly outperforms PLS in terms of predictive power.This offers a significant advantage over PLS since its abstract structure cannot be associated with the underlying biological structure.

View Article: PubMed Central - HTML - PubMed

Affiliation: REQUIMTE, Systems Biology & Engineering Group, DQ/FCT, Universidade Nova de Lisboa, Campus Caparica, Portugal.

ABSTRACT

Background: Elementary flux modes (EFM) are unique and non-decomposable sets of metabolic reactions able to operate coherently in steady-state. A metabolic network has in general a very high number of EFM reflecting the typical functional redundancy of biological systems. However, most of these EFM are either thermodynamically unfeasible or inactive at pre-set environmental conditions.

Results: Here we present a new algorithm that discriminates the "active" set of EFM on the basis of dynamic envirome data. The algorithm merges together two well-known methods: projection to latent structures (PLS) and EFM analysis, and is therefore termed projection to latent pathways (PLP). PLP has two concomitant goals: (1) maximisation of correlation between EFM weighting factors and measured envirome data and (2) minimisation of redundancy by eliminating EFM with low correlation with the envirome.

Conclusions: Overall, our results demonstrate that PLP slightly outperforms PLS in terms of predictive power. But more importantly, PLP is able to discriminate the subset of EFM with highest correlation with the envirome, thus providing in-depth knowledge of how the environment controls core cellular functions. This offers a significant advantage over PLS since its abstract structure cannot be associated with the underlying biological structure.

Show MeSH

Related in: MedlinePlus

Predicted metabolic fluxes by PLP. Predicted against measured fluxes computed by the PLP model for the BHK data set. Blue circles and red triangles represent the calibration and validation data points, respectively.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3750108&req=5

Figure 8: Predicted metabolic fluxes by PLP. Predicted against measured fluxes computed by the PLP model for the BHK data set. Blue circles and red triangles represent the calibration and validation data points, respectively.

Mentions: To test the predictive power, PLS and PLP models were calibrated with the calibration data set composed by 50% of data points and then simulated on the validation data set composed by the remaining 50% measured points. The PLS model with 18 latent variables explained 90.1% of R variance in the calibration dataset but only 76.8% of the validation dataset. The quality of the results can be visually inspected in Figure 7. The degradation of accuracy in the validation dataset is rational given that the model is requested to predict data points, which may lay outside of the domain of experience defined by the calibration data set. As for PLP it is a very interesting result to verify that despite explaining a lower variance in the calibration data set (83.2% against 90.1% for PLP and PLS respectively), the accuracy of the validation data set was higher than that of PLS (81.9% against 76.8% for PLP and PLS respectively). Moreover, the variance of the validation data set is almost equal to that of the calibration data set, denoting a more consistent model, with higher predictive power than the PLS one (Figure 8).


Projection to latent pathways (PLP): a constrained projection to latent variables (PLS) method for elementary flux modes discrimination.

Ferreira AR, Dias JM, Teixeira AP, Carinhas N, Portela RM, Isidro IA, von Stosch M, Oliveira R - BMC Syst Biol (2011)

Predicted metabolic fluxes by PLP. Predicted against measured fluxes computed by the PLP model for the BHK data set. Blue circles and red triangles represent the calibration and validation data points, respectively.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3750108&req=5

Figure 8: Predicted metabolic fluxes by PLP. Predicted against measured fluxes computed by the PLP model for the BHK data set. Blue circles and red triangles represent the calibration and validation data points, respectively.
Mentions: To test the predictive power, PLS and PLP models were calibrated with the calibration data set composed by 50% of data points and then simulated on the validation data set composed by the remaining 50% measured points. The PLS model with 18 latent variables explained 90.1% of R variance in the calibration dataset but only 76.8% of the validation dataset. The quality of the results can be visually inspected in Figure 7. The degradation of accuracy in the validation dataset is rational given that the model is requested to predict data points, which may lay outside of the domain of experience defined by the calibration data set. As for PLP it is a very interesting result to verify that despite explaining a lower variance in the calibration data set (83.2% against 90.1% for PLP and PLS respectively), the accuracy of the validation data set was higher than that of PLS (81.9% against 76.8% for PLP and PLS respectively). Moreover, the variance of the validation data set is almost equal to that of the calibration data set, denoting a more consistent model, with higher predictive power than the PLS one (Figure 8).

Bottom Line: PLP has two concomitant goals: (1) maximisation of correlation between EFM weighting factors and measured envirome data and (2) minimisation of redundancy by eliminating EFM with low correlation with the envirome.Overall, our results demonstrate that PLP slightly outperforms PLS in terms of predictive power.This offers a significant advantage over PLS since its abstract structure cannot be associated with the underlying biological structure.

View Article: PubMed Central - HTML - PubMed

Affiliation: REQUIMTE, Systems Biology & Engineering Group, DQ/FCT, Universidade Nova de Lisboa, Campus Caparica, Portugal.

ABSTRACT

Background: Elementary flux modes (EFM) are unique and non-decomposable sets of metabolic reactions able to operate coherently in steady-state. A metabolic network has in general a very high number of EFM reflecting the typical functional redundancy of biological systems. However, most of these EFM are either thermodynamically unfeasible or inactive at pre-set environmental conditions.

Results: Here we present a new algorithm that discriminates the "active" set of EFM on the basis of dynamic envirome data. The algorithm merges together two well-known methods: projection to latent structures (PLS) and EFM analysis, and is therefore termed projection to latent pathways (PLP). PLP has two concomitant goals: (1) maximisation of correlation between EFM weighting factors and measured envirome data and (2) minimisation of redundancy by eliminating EFM with low correlation with the envirome.

Conclusions: Overall, our results demonstrate that PLP slightly outperforms PLS in terms of predictive power. But more importantly, PLP is able to discriminate the subset of EFM with highest correlation with the envirome, thus providing in-depth knowledge of how the environment controls core cellular functions. This offers a significant advantage over PLS since its abstract structure cannot be associated with the underlying biological structure.

Show MeSH
Related in: MedlinePlus