Figure 5

Sensitivity within the top 1% samples (left) and top 5% samples (right) on the sm-enzyme gold standard set for ECK12 as sorted by each genome context method, varying the list of organisms used for computing these scores. Parameter E is set to 10^-4. Solid lines correspond to subsets of organisms chosen by the clustering method, and dashed lines correspond to a random selection of organism subsets. The full reference list corresponds to the right most point in each figure (number of organisms equal to 460).