Figure 2

A visualization of the F -measure for finding highly conserved virulence genes in pathogens, with and without a cereal host. The set of 174 fungal proteomes from the JGI MycoCosm tree [33] is shown with the subsets of pathogenic fungi and cereal-infecting fungi. The genera of cereal-infecting fungi are colored according to fungal lifestyle (biotrophs, hemibiotrophs, necrotrophs). The F-measure is the harmonic mean of precision and recall and ranges from 0 to 1. For a query pathogen protein of interest, a F-measure of 0 means that the protein is only found in the query pathogen genome. A perfect pathogen F-measure of 1 means that this protein is found in all pathogen genomes and at the same time only in pathogenic fungi, thus indicating a protein highly relevant to pathogenicity. In the middle of the scale sit pathogen proteins which can be found mostly in pathogens and at the same time in some non-pathogens. We also introduce the more specific cereal F-measure to identify proteins which are highly conserved and exclusive to cereal-infecting fungi to identify host-specific infection patterns.