E-Flux2 and SPOT: Validated Methods for Inferring Intracellular Metabolic Flux Distributions from Transcriptomic Data

Abstract

BackgroundSeveral methods have been developed to predict system-wide and condition-specific intracellular metabolic fluxes by integrating transcriptomic data with genome-scale metabolic models. While powerful in many settings, existing methods have several shortcomings, and it is unclear which method has the best accuracy in general because of limited validation against experimentally measured intracellular fluxes.ResultsWe present a general optimization strategy for inferring intracellular metabolic flux distributions from transcriptomic data coupled with genome-scale metabolic reconstructions. It consists of two different template models called DC (determined carbon source model) and AC (all possible carbon sources model) and two different new methods called E-Flux2 (E-Flux method combined with minimization of l2 norm) and SPOT (Simplified Pearson cOrrelation with Transcriptomic data), which can be chosen and combined depending on the availability of knowledge on carbon source or objective function. This enables us to simulate a broad range of experimental conditions. We examined E. coli and S. cerevisiae as representative prokaryotic and eukaryotic microorganisms respectively. The predictive accuracy of our algorithm was validated by calculating the uncentered Pearson correlation between predicted fluxes and measured fluxes. To this end, we compiled 20 experimental conditions (11 in E. coli and 9 in S. cerevisiae), of transcriptome measurements coupled with corresponding central carbon metabolism intracellular flux measurements determined by 13C metabolic flux analysis (13C-MFA), which is the largest dataset assembled to date for the purpose of validating inference methods for predicting intracellular fluxes. In both organisms, our method achieves an average correlation coefficient ranging from 0.59 to 0.87, outperforming a representative sample of competing methods. Easy-to-use implementations of E-Flux2 and SPOT are available as part of the open-source package MOST (http://most.ccib.rutgers.edu/).ConclusionOur method represents a significant advance over existing methods for inferring intracellular metabolic flux from transcriptomic data. It not only achieves higher accuracy, but it also combines into a single method a number of other desirable characteristics including applicability to a wide range of experimental conditions, production of a unique solution, fast running time, and the availability of a user-friendly implementation.

Enhanced flux prediction by integrating relative expression and relative metabolite abundance into thermodynamically consistent metabolic models

“…To effectively compare the predicted in silico fluxes from REMI with the corresponding 13 C-determined in vivo intracellular fluxes, the following two metrics were used: the uncentered Pearson correlation coefficient (Equation 18), and the average percentage error in predicted fluxes (Equations (18)- (20)). The uncentered Pearson correlation is a good metric for the flux comparison, as fluxes are usually not centered, and it has been used for comparing two flux vectors [23]. The average percentage error has been used in the GX-FBA method [12] to compare two fluxes.…”

“…of E.coli by Kim et al[23] and were originally obtained from two independent studies done by Ishii et al(8 datasets) [20] and Holm et al(3 datasets) [21] were used for the evaluation of the REMI methodology.The three datasets from Holm et al [20] comprise genome-wide transcriptomic data together with fluxomic data (21 measured fluxes) collected from three experimental conditions: wildtype E. coli, cells overexpressing NADH oxidase (NOX), and cells overexpressing the soluble F1-ATPase (ATPase). The eight datasets from Ishii et al [20] include genome-wide transcriptomic, fluxomic (31 measured fluxes), and metabolomic (42 metabolites) data obtained under eight different experimental conditions: wildtype E. coli cells cultured at different growth rates of 0.2, 0.6, and 0.7 per hour along with single-gene knockout mutants of the glycolysis and pentose phosphate pathway (pgm, pgi, gapC, zwf, and rpe).…”

Bioenergetics of Monoterpenoid Essential Oil Biosynthesis in Nonphotosynthetic Glandular Trichomes

“…To the best of our knowledge, fermentative reactions have not been included in any metabolic reconstruction for plants; therefore, we added the relevant reactions manually to menpiGT_2015 (version 6). In addition, transcriptome data obtained with isolated peppermint GTs at secretory stage (Ahkami et al, 2015) were incorporated using a modified SPOT algorithm (Kim et al, 2016; Figure 1B, were disabled and, therefore, became values to be predicted by the model based on gene expression patterns. Allowed outputs were monoterpenes, sesquiterpenes, polymethoxylated flavones, ethanol, lactate, cellulose, amino acids, water, and carbon dioxide (Supplemental Table S3).…”

“…The SPOT algorithm was implemented according to Kim et al (2016), with the addition of Boolean constraints to prevent reversible reactions from being counted twice in the objective function (referred to as Boolean constrained SPOT). Prior to running Boolean constrained SPOT, boundary fluxes were modified to disable the experimentally derived biomass export reaction and to allow the optional import and export of small molecules.…”

Gene-centric constraint of metabolic models

“…When looking to predict a specific flux distribution, however, we need to take further steps to ensure a unique distribution. In general there will be a space of fluxes that correspond to the maximal objective value, and the recent paper published by Kim et al outlines an extension to the original E-Flux algorithm [29]. The E-Flux2 approach consists first of creating a model in which all a priori information on cellular uptake rates and ATP maintenance flux have been removed, before performing a two-stage optimization.…”