Figure 1.
Distribution of bacterial similarity to the human proteome.
All bacteria in set 1 were used in creating this histogram. Each bar represents the number of bacteria having between x% and of their 5-mers not found in the human proteome, where x is the number below the bar.
Table 1.
Pathogenic bacteria used in this study (set 2).
Table 2.
Nonpathogenic bacteria used in this study (set 3).
Table 3.
The most similar and dissimilar bacteria to the human proteome.
Figure 2.
Pathogenic versus nonpathogenic bacteria.
Relative similarity to the human proteome of the bacteria in set 1 (all bacteria sequenced to date), set 2 (20 pathogenic bacteria), and set 3 (14 nonpathogenic bacteria). Each point indicates that, on average, y% of the 5-mers in the proteomes in that set were found x times in the human proteome. Because only rare 5-mers are of interest in this study, bacterial 5-mers that were found more than ten times in the human proteome are not represented. The length in one direction of the error bar associated with each point represents the standard deviation of the measurements that were averaged to calculate that point.
Figure 3.
Similarity to the human proteome of surface-accessible proteins from pathogens and nonpathogens.
The similarity to the human proteome is shown for proteins (A) from Gram-positive bacteria that are predicted to localize to the cell wall, and (B) from Gram-negative bacteria that are predicted to localize to the outer membrane. Bacterial 5-mers that were found more than ten times in the human proteome are not represented. The length in one direction of the error bar associated with each point represents the standard deviation of the measurements that were averaged to calculate that point.
Figure 4.
Acute pathogens versus chronic pathogens.
Relative similarity to the human proteome of the bacteria in set 2 a (bacteria that cause acute infections), set 2 b (bacteria that cause chronic infections), set 4a (one randomly-generated bacterial proteome corresponding to each proteome in set 2 a ), and set 4b (one randomly-generated bacterial proteome corresponding to each proteome in set 2 b ). Bacterial 5-mers that were found more than ten times in the human proteome are not represented. The length in one direction of the error bar associated with each point represents the standard deviation of the measurements that were averaged to calculate that point.
Figure 5.
Relationship between proteome size and percentage of 5-mers absent from the human proteome.
The best-fit line (in green) was calculated using least squares and had an value of 0.062.
Figure 6.
Relationship between G-C content and percentage of 5-mers absent from the human proteome.
As the plot exhibits two distinct regions, two best-fit lines were calculated. The green best-fit line was calculated using points with G-C contents less than 0.52 and had an value less than 0.01, whereas the blue best-fit line was calculated using points with G-C contents greater than or equal to 0.52 and had an
value of 0.74.