Figure 5

Number of proteins in PCC 7942 with an unusual phylogenetic relationship. For each protein coded in the genome of PCC 7942 we did blast searches against NCBI nr database to identify top scoring best hits. Then, for each protein having at least 4 hits we reconstructed a Neighbor-Joining tree (100 bootstrap replications). We then identified whether the nearest neighbor in the tree belonged to a cyanobacteria or not. When it was not possible to use the tree to identify the nearest neighbor (because the immediate ancestral node of PCC 7942 protein did not share a unique co-descendent OTU; or the bootstrap value was lower than 50; or because there were not at least four homologs in the database) we used the BLAST best hit criteria to decide whether or not the query protein has an unusual phylogenetic relationship. As shown, most proteins do not have its nearest homolog outside cyanobacteria (1177 + 1035 = 2212); however 253 proteins have an unusual phylogenetic relationship (10 of them to protists endowed with plastids). *Number of proteins in which the phylum identified was the same with BLAST and with Neighbor-Joining.