A comparative analysis of taxonomic and phylogenetic based community measures as applied to marine Prochlorococcus marinus communities

There are two approaches to measuring the divergence between biological communities. There are the traditional taxonomic based measures which examine communities based on the similarity of their observed species, and there are the new phylogenetic based measures, which identify community similarity through the use of a phylogenetic tree that relates all members of the community. When examining microbial communities, both of these approaches have been applied. In order to carry out the traditional taxonomic-based community analysis, microbes are clustered into operational taxonomic units (OTUs) based on the percent similarity between their respective copies of selected marker genes. The DNA sequence used to cluster microbes is typically the 16S gene or the 16S/23S intergenic transcribed spacer region (ITS). The clustering is then based on a pre-selected percent similarity. This analysis was undertaken to explore relationships between microbial communities and environmental factors based on the type of community measure used. In order to do this, taxonomic (Bray-Curtis Index) and phylogenetic (UniFrac) analyses were performed on a series of Prochlorococcus marinus communities, collected from the Atlantic and Pacific Oceans. The results of the OTU analysis in this work differed from a previously published analysis of these data due to large differences in the number of OTUs observed and the relative abundances of each, as well as observed environmental variables such as depth and dispersal time. Both analyses identified light (at 90% OTU similarity) and nitrate (at 97% OTU similarity) as having a significant relationship to community composition. As well both analyses concluded that phosphate (at 90% and 97% OTU similarity) and light (at 99.5% OTU similarity) were not significantly influencing community composition. Unweighted UniFrac analysis, which estimates diversity based on the presence and absence of sequences in different communities, identified depth and geographic distance as significantly influencing the community structure of P. marinus, consistent with the results identified at the 99.5% OTU level by this work. Based on these results it can be shown that taxonomic-based analyses at 99.5% agreed with the results from the unweighted UniFrac distances. Weighted UniFrac, which considers the relative abundance of sequences in addition to their presence or absence, did not match any taxonomic based results for the P. marinus communities, nor did it identify any significant relationships between the observed environmental variables and community composition. This suggests that relative abundances provide additional information not captured by taxonomic community measures. This analysis identified depth and geographic distance as having significant relationships with P. marinus communities. The significant role of depth in structuring communities of P. marinus is consistent with previous literature. The large geographic distance covered by the samples observed in this data cover a range of environmental influences.