Figure 5

Analysis on in silico simulated sequencing experiments across different levels of coverage and taxon distribution. Sim1 (blue) contains ten tier-2 PGDB genomes in approximately equal proportion. Sim2 (red) has one taxon overrepresented by 20-fold. Tier-2 taxa were selected on the basis of approximately equal genome size and gene content (a). Predicted pathway recovery as a percentage of the total pathways predicted from the full genome (b). Specificity (triangles) and sensitivity (squares) classification performance of predicted pathways using the pathways predicted on the full genomes as the gold standard (c). Interpolating lines were drawn via a natural spline.