N with a conserved genomic neighbourhood in all species examined. Seven additional orthologs were unique to the relaxed core gene sets of the cyanobacterial phylum (allowed to be missing from 1-3 of the finished genomes) but are currently poorly characterized. Genome sizes and number of duplicated genes within the major Simvastatin acid ammonium clade, which comprises species of varying complexity and choice of habitat (including marine and freshwater, unicellular and filamentous and heterocystous species) evolve at a higher evolutionary rate than the clade of marine picocyanobacteria. Variation in cyanobacterial genome sizes is the result of a mix of gains of losses in the former clade and of one single reduction event in the latter. It is also deduced that the common ancestor of extant cyanobacteria had a genome size of 4.5 Mbp and contained between 1678 and 3291 proteins, 4-6 of which are unique to cyanobacteria today. The largest genomes of cyanobacteria contain the highest number of paralogs and there is a strong correlation between genome and gene family expansions within the more complex clade. The latter is likely to result in increased adaptive potential but may also lead to a an accumulation of non-coding genomic regions as the organisms shift their habitat and/or life-style, rendering duplicated genes redundant and subject PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/27646945 to increased mutation rates. A clear example of this phenomenon is the obligate symbiont `Nostoc azollae’ 0708 in which 64 of in-paralogs are pseudogenes. A number of orthologous proteins underpin traits of certain cyanobacterial groups. These include a set of signal transduction proteins and a glycosyl hydrolase specific to plant symbionts. Conversely, the lack of certain conserved orthologs or domains in a few cyanobacteria can be correlated with the phenotype they exhibit as seen in Cylindrospermopsis raciborskii with terminal heterocysts only, and the loss of functional genes involved in pigment biosynthesis in `Nostoc azollae’ 0708.MethodsDatasetConclusions With the dataset used, based on 58 cyanobacterial genomes, the stringent core genetic repertoire of cyanobacteria now represents 404 orthologs. A relaxed core gene set, based on 39 of these genomes sequenced to completion,All available cyanobacterial genomes as of June 24th 2010 were downloaded from NCBIs genome database FTP server. The dataset included 58 genomes in total, 39 of which were sequenced to completion. Sequences for all protein coding open reading frames (ORFs) andLarsson et al. BMC Evolutionary Biology 2011, 11:187 http://www.biomedcentral.com/1471-2148/11/Page 18 ofpredicted pseudogenes, the latter acquired from the IMG database [68], were included in the analysis.Orthologous protein groups and annotationsAll ORFs in the dataset were subjected to orthologous protein grouping using OrthoMCL v.2.0.1 [69,70]. Protein sequences for each group were aligned with MUSCLE v.3.5.1 [71], and Hidden Markov Models (HMMs) [72] were built from the alignments using hmmbuild v.3.0 [73]. Pfam and Uniprot databases (26 August 2010) were searched with PubMed ID:https://www.ncbi.nlm.nih.gov/pubmed/28119436 the HMMs as queries using hmmsearch v.3.0 [73] and annotations were assigned to orthologous groups for hits with full and domain-specific e-values < 0.01 and a bias/score ratio < 10. COG functional groups were assigned to orthologous groups using RPS-BLAST against the CDD database (23 August 2010) with all sequences of a group as query. For each group, the most predominant best hit among orthologous sequences (at e-value cut-off.