Supplementary Figure 1

This figure shows the phylogenetic profiles of protein interactions associated with both RNA polymerase and DNA polymerase complexes. By looking at the co-conservation of interacting proteins across multiple species, it can clearly be seen that the composition of these complexes has diverged, with modules being lost or replaced by significantly divergent functional homologs. Analyses of this kind are important when trying to use protein-protein interaction data to rationalize the targets of new anti-microbial compounds.

Supplementary Figure 2

This figure displays as a heatmap, the conservation of proteins present in the interaction network across 148 genomes, including Bacteria, Archea and Eukarya, as judged by BLAST homology. Proteins are clustered together based on the similarity of their conservation profiles. Many of the proteins are broadly conserved across many different phylogenetic groups suggesting that these proteins may be good targets for anti-microbial compounds.

Supplementary Figure 3

This figure shows the network properties of the essential conserved set of 209 proteins. Parts a and b are included for direct comparison with the same analyses for the entire dataset present in Fig. 3a and Fig.3b.

Supplementary Figure 4

This figure shows the results of more detailed network analysis. Part a shows the removal of the 20 most highly connected protein ‘hubs’ (>15 connections each) from the interaction network and can be compared directly with part b which is the network before attack (same as figure 3c). The graphs presented in parts c and d show how the number of connections (interactions) a protein has correlates with the conservation of the protein across different species.

Supplementary Figure 5

This figure shows the conservation of essential and non-essential proteins and their interactions from the interaction dataset reported in this study. Analyses of protein occurrence across all three domains of life indicated that interacting proteins were more likely to be co-conserved than control sets of randomly selected E. coli proteins. This suggests that the protein-protein interactions are themselves conserved between the sets of homologous proteins.

Supplementary Figure 6

This figure is analogous to that presented in Figure 3 parts c and d, except that gene co-occurrence is based on COGs genomes. The network nodes are shaded to reflect the increasing number of COGs genomes (>= 1 to >= 40) genes are co-conserved in.

Supplementary Table 1

This table shows the complete protein-protein interaction dataset reported in this study, along with the method the interaction was observed by. Validated interactions are shown in bold typeface. Detection of the bait protein is shown in italics.

Supplementary Table 2

This table details the name, annotation and experimental status of each of the 1000 genes targeted in this study. The 209 genes present on the “essential-conserved” list are shown in bold typeface.

Supplementary Table 3

This table reports protein-protein interactions uncovered in this study which have previously been reported in the Database of Interacting Proteins (DIP), the Biomolecular Interaction Network Database (BIND), STRING database, or Prolinks database. Approximately 15 % of the validated interactions reported here have previously been identified, demonstrating the power of this technique in uncovering novel interactions.

Supplementary Table 4

This table lists the most highly connected 20 ‘hubs’ present in the interaction network (Figs 3a/3c), which are conserved in >= 125 different genome sequences. Networks that exhibit scale-free properties are expected to be susceptible to the removal of these ‘hub’ proteins when compared to removal of random proteins.

Supplementary Table 5

This table lists (in bait-prey format) the protein-protein interactions of proteins which are both conserved in >= 125 genomes. The numbers of genomes in which both proteins co-occur are broken down into Bacteria, Archea and Eukarya kingdoms. Strikingly, aside from a few outlying pairs, these protein-protein interactions form a single interconnected component (Figure 3d).

Supplementary Table 6

This table lists protein-protein interactions (in bait-prey format), where the corresponding genes are separated by less than 500 base pairs (in E. coli or another species). The known or predicted E. coli operon is shown for those interacting pairs close together in E. coli.

Supplementary Table 7

This table lists the COGs genomes used in this study in their respective groupings.