Supplementary Figure 3: Testing for bias in effector detection.
From: Genomic analysis of 38 Legionella species identifies large and diverse effector repertoires
(a) Comparison of the number of putative effectors and ORFs with Icm/Dot C-terminal translocation signal per species. The x axis represents the percentage of the total set of ORFs of interest that was detected in each species (for example, the bottom bar represents the percentage of putative effectors in L. parisiensis out of the total number of putative effectors in all the genomes combined). The species tree is presented on the left. (b) Correlation between the number of putative effectors and the number of ORFs with C-terminal Icm/Dot translocation signal. A strong (R2 = 0.762) and significant correlation (P value = 1.004 × 10−13, Pearson correlation test) was found between the two. Both the number of effectors and the number of ORFs with translocation signal per species did not significantly deviate from the normal distribution (Shapiro-Wilk normality test P values = 0.43 and 0.68, respectively). (c) Correlation between the evolutionary distance from L. pneumophila and the ratio of putative effectors to ORFs with Icm/Dot C-terminal translocation signal. No significant correlation was detected (P value = 0.14, Spearman correlation test), suggesting that there is probably no significant preferential effector detection in species that are evolutionarily close to L. pneumophila. Spearman correlation was used because both distributions deviated from normality (P value = 0.004 and 0.01, respectively, Shapiro-Wilk test).
