Such similarity information
need not include continuous evolutionary distances, but could be as simple as assigning similarity values based on general taxonomic group. Our simulations showed that, to some extent, the choice of q did effect the agreement between naïve and similarity-based diversity calculations. Generally speaking, for small positive q values it appears that there was greater #Selleckchem Elafibranor randurls[1|1|,|CHEM1|]# agreement between naïve and similarity-based diversity calculations. These differences were statistically significant when the difference in proportion of agreement between two q was ~ 0.15 (based on Z test for two population proportions). Turning to the impacts of tree typology and sample relative abundance distributions, our results showed that the percent agreement between the naïve and similarity-based diversity calculations decreased slightly with increasing skewed abundance distributions (Figure 5C) and increasing tree imbalance (Figure 5D). This finding is significant because, while tree shape changes greatly between different sized trees , skewed abundance distributions [66, 67] and higher tree imbalances [25, 65] are likely better representations of the majority of true environmental communities than perfectly balanced abundance distributions and phylogenies would be.
In contrast, the percent of agreement increased slightly with increasing sample size (Figure 5A) and the use of non-ultrametric trees (Figure 5B), which are also likely good representations of the majority selleck products of true environmental microbial communities that may include thousands of OTUs e.g.,  and may produce undated non-ultrametric trees. Since Loperamide these simulations of
phylogenetic trees with characteristics that resemble those of real datasets showed both slight increases and decreases in the percent agreement between the naïve and similarity-based diversity calculations, the percent agreement between naïve and similarity-based diversity calculations for real datasets is probably approximately 50%. Figure 5 Agreement between naïve and similarity-based diversity profiles for different simulated communities. (A) For different numbers of OTUs sampled from the total pool of 2048, (B) for ultrametric (grey) and non-ultrametric trees (white), (C) for communities with different Fisher’s alpha diversity values, (D) for communities with different tree imbalances. For panels (B), (C), &(D) sampled communities sized was 256; (A), (B), &(C) tree imbalance was 9.54; (A), (B), &(D) community abundance distribution was logseries with a Fisher’s Alpha of 1. Proportion of agreement is based on 100 simulations. “black square symbol” (q = 0), “red circle symbol” (q = 1.1) “blue triangle symbol” (q = 3.1), “magenta triangle symbol” (q = 5.1). Conclusions This study explored whether similarity-based diversity profiles can aid our interpretation of microbial diversity.