Pathway scores of the relevant pathways for the periodontal species. Plotting the pathway scores of the relevant pathways (from Table 6), the differences of A. actinomycetemcomitans (black) compared to F. nucleatum (red), P. gingivalis (green), and T. denticola (blue) become apparent. In contrast, the scores for F. nucleatum, P. gingivalis, and T. denticola are very similar.

Overview of the approach. The three major steps of our approach are: metabolic reconstruction of completely sequenced genomes resulting in pathway profiles; pathway selection resulting in lists of pathways ranked by relevance; and cross-checking of the resulting pathway rankings by classification in order to estimate their significance (Figure S1 in Additional data file 2).

Classification quality for the classification of archaea into methanogens and non-methanogens using the nearest neighbor classifier. The classification based on the four most relevant pathways yields a perfect separation of methanogenic archaea and non-methanogenic archaea for all attribute subset selection methods used (green, ReliefF; yellow, SVMAttributeEval; blue, wrapper (naïve Bayes)). Classification based on all pathways (marked by a horizontal line) and based on randomly picked pathways (red) show lower classification quality.

Classification quality for the phenotype member of red or orange cluster. Left: classification of all genomes (266) into genomes that are members and non-members of the 'red/orange' cluster using the nearest neighbor classifier (IB1). Right: classification of oral genomes (15) into genomes that are members and non-members of the 'red/orange' cluster related using the nearest neighbor classifier (IB1). Compared to classification based on all pathways (marked by a horizontal line) and based on randomly picked pathways (red), the classification based on the most relevant pathways yields better separation of the cluster members and other species in both genome datasets.

Classification quality for the phenotype periodontal disease causing. Left: classification of all genomes (266) into genomes related and not related to periodontal disease using the nearest neighbor classifier (IB1). Right: classification of oral genomes (15) into genomes related and not related to periodontal disease using the nearest neighbor classifier (IB1). Compared to classification based on all pathways (marked by a horizontal line) and based on randomly picked pathways (red), the classification based on the most relevant pathways yields better separation of periodontal species and other species in both genome datasets.

Cross-checking for the phenotype methanogenesis. The classification quality diagrams for nearest neighbor classifier (IB1) and the naïve Bayes classifier show that the identified most relevant pathways are well suited to distinguish methanogens and non-methanogens (sensitivity × selectivity = 1.0). According to the cross-check, the most relevant pathways identified by pathway selection are considered as significant. Apart from using ReliefF top-ranking pathways (green) for the classification with IB1, the maximum classification quality is already reached for the (up to) five most relevant pathways (these pathways are listed in Table 1).

Classification quality for the classification of archaea into methanogens and non-methanogens using the nearest neighbor classifier while omitting the pathway of methane synthesis. Omitting the pathway of methane synthesis (methane1) in our analyses, the classification based on the most relevant pathways still reaches perfect separation of methanogenic archaea and non-methanogenic archaea for all attribute subset selection methods used (green, ReliefF; yellow, SVMAttributeEval; blue, wrapper (naïve Bayes)). Classification based on all pathways (marked by a horizontal line) and based on randomly picked pathways (red) show lower classification quality.

Degradation of histidine. The pathways glutamate fermentation (fnc1) (red) and degradation of histidine to L-glutamate (histidine2) (black) describe (amino acid) degradations producing ammonia as an end product. Due to the reversibility of all its reactions, this also includes the pathway of biosynthesis of 5-formimino-THF (blue), which - inversely followed - describes the degradation of 5-formimino-THF to glutamate (c2). All three pathways are interconnected and can be interpreted as complete degradation of L-histidine to acetate and ammonia (NH₃). Thereby, three moles of ammonia per mole of histidine are produced (green or turquoise boxes, respectively). As histidine2 includes an alternative route from L-histidine to glutamate (dashed line), one mole of ammonia is either produced by the conversion of N-carbamoyl-L-glutamate to L-glutamate or by the conversion of N-formimino-tetrahydrofolate to 5,10-methenyl-tetrahydrofolate.

Estimating the significance of pathway rankings provided by pathway selection. For phenotypes that are weakly associated with the presence or absence of specific metabolic pathways, the classification quality should be within the same range for classification based on randomly picked pathways (red), all pathways (marked by a horizontal line), and pathways highly ranked in attribute subset selection (green, ReliefF; yellow, SVMAttributeEval; blue, wrapper (naïve Bayes)). As an example, the right diagram shows the classification quality for the phenotype 'habitat: soil' (depending on the number of top-ranking pathways used for classification). In this case, the top-ranking pathways provided by attribute subset selection are considered as not significant for the phenotype. The left diagram shows the classification quality values for the phenotype 'obligate intracellular'. Using the most relevant pathways for classification results in higher classification quality compared to using all pathways or randomly picked pathways. Furthermore, the quality values lie above 0.6. In this case, the most relevant pathways derived by attribute subset selection are considered as significant.