Skip to main content
Advertisement

< Back to Article

Figure 1.

Overview of the reconstruction method.

Genomes of target species are subjected to BLAST, GTG and InterProScan analyses to evaluate enzyme probabilities. Plausible gapless metabolic networks are assembled based on integrated enzyme evidence. Finally reconstructed models are converted into SBML and generic stoichiometric matrix formats.

More »

Figure 1 Expand

Figure 2.

Phylogeny of the 49 fungal species according to [42].

Fungal taxonomic class indicated by colors.

More »

Figure 2 Expand

Table 1.

Prediction performance of CoReCo phases I and II, naive Bayesian classifier and individual BLAST and GTG classifiers expressed in terms of area under the ROC curve (AUC).

More »

Table 1 Expand

Figure 3.

True and false positive rates for S. cerevisae consensus model enzyme prediction in the evolutionary distant species setting with the CoReCo phase II reconstructed model (cyan), CoReCo phase I model (black), naive Bayesian classifier (red), and BLAST (blue) and GTG (green) classifiers.

More »

Figure 3 Expand

Figure 4.

Reconstructed models summarized in terms of number of reactions (NumReactions), number of gapfilling reactions (Gapfills) and fraction of gapfilling reactions to all reactions in the reconstructed model (GapfillRatio).

Density plots shown in diagonals. Species colored according to taxonomic class (see Fig. 2). In top left, pink and blue ellipses denote Saccharomycotina and Eurotiomycetes, respectively. In bottom right, purple and cyan ellipses denote Sordariomycetes and Agaricomycetes, respectively. Data for E. cuniculi not shown.

More »

Figure 4 Expand

Figure 5.

Yields of biomass components in fungal models.

Reaction directionality constrained. CO2 assimilation reaction unconstrained. Column-wise normalized yields reported: (white), (light green), (green), (dark green). Fungal taxonomic class indicated by row colors in left margin.

More »

Figure 5 Expand

Figure 6.

For each S. cerevisiae gene knockout, all genes involved in complexes with the deleted gene that do not participate in other complexes are also deleted.

For the four complexes shown, YEL039C knockout will result also in the deletion of YDL178W and reactions associated with it because no other YDL178W containing complex remains. In contrast, YEL071W is not deleted, because it participates in a still functional complex with YJR048W.

More »

Figure 6 Expand

Table 2.

Growth prediction accuracy in knockout models with constrained reaction directionality.

More »

Table 2 Expand

Figure 7.

Bayesian network model for computing enzyme probabilities containing three node groups: hypothetical ancestral species (yellow), present species (green), and BLAST and GTG evidence nodes (blue).

More »

Figure 7 Expand

Figure 8.

A. Conditional probability distributions for BLAST (left) and B. GTG (right) sequence evidence.

Blue solid line: , red dashed line: .

More »

Figure 8 Expand