We presume such a similar environment is more likely to homogenize microbial communities, rather than promote individual differences. Nevertheless, this shared number of OTUs appears relatively low compared to the number of shared OTUs (21 OTUs, at 97% sequence #Bucladesine research buy randurls[1|1|,|CHEM1|]# identity cut-off) among populations of zebrafish from radically different environmental conditions, coming either from natural populations in India or from artificial environments in two separate laboratories in the USA [17]. For now, this difference
in shared OTUs between our study and the study focusing on zebrafish is difficult to interpret due to methodological variation e.g. pooled versus individual samples, V1-V2 versus V3 16S rRNA region, [17]. It will be interesting to investigate if these differences in shared OTUs membership are environmentally determined (e.g., a largely different food preference and habitat) or are species specific (e.g., the unusual Atlantic cod immune system which might affect its host-microbe interactions [12, 13]). Community diversity estimates based on 454 amplicon data are influenced by methodological factors such as fragment length, PCR bias and choice of 16S rRNA gene region. Specifically, shorter amplicon lengths (e.g. < 400 bp) may result in relatively higher
diversity estimates compared to longer fragments [26] and arguably provide a better assessment https://www.selleckchem.com/products/ipi-145-ink1197.html of community structure [27]. In contrast, species richness estimate based on analyses of the 16 s rRNA V3 region appears to slightly underestimate diversity relative to the full-length gene [28]. Such methodological issues make it difficult to compare community diversity across different studies [29], although metrics that use both richness and relative abundance (i.e. Shannon and Inverse Simpson indices) appear robust [30], in particular considering our extensive sequencing depth [31]. Interestingly, these metrics fluctuate several orders of magnitude among our different specimens, and show large individual variation in community composition and diversity. The most diverse individuals appear to have a comparable
community OSBPL9 complexity relative to those found in humans [7, 32]. A variety of properties, such as shared OTU membership, shared phylogeny, persistence or connectivity can be used to define microbial cores [33]. Here we investigated a core microbiota based on shared membership. Definitions for such a core have been proposed ranging from a lineage present in more than half the population [3] to an abundant lineage shared among all individuals [8]. We argue that the utility of such concept depends on the specificity with which it describes a biological phenomenon and favor the idea that a lineage should be reliably identified among all individuals in order to belong to a core microbiota, hence with a detection probability of at least 99%.