In depth view of V3 and V5 potentials in Illumina based Bacterial 16S rDNA surveys of highly complex environments.

Ubiquitous distribution of Bacteria in most known earth environments and the reciprocal interaction of their diversity and richness with environmental conditions, has raised the interest of in depth 16S rDNA analysis in sequence based diversity studies. Recent advances in the chemistry of high throughput sequencing techniques provide the foreground for sample multiplexing and screening of millions of very short reads of 16S rDNA polymerase chain reaction (PCR) products per run. The widely utilized in environmental studies V3 hypervariable 16S rDNA region was also the prime selection in recent sequencing attempts using Illumina technology. In the present study based on the provided information and sequence length criteria we have evaluated the V5 hypervariabe region as candidate for screening highly complex environmental samples like agricultural or natural soils. Based on a real dataset of three highly complex soil environments (10-20,000 richness for OTUs at the “species” 16S rDNA sequence distance definition level) in triplicates previously generated with the referred technology, we have extracted full-length 16S rDNA sequences from the RDP database with Mega-BLAST. The extracted sequences were used for performing in silico PCRs using literature derived primer-sets targeting the V3 and V5 regions. Comparisons between the referred hypervariable region related, the full length 16S rDNA and the original datasets were performed using OTU and taxonomy based approaches. Hierarchical clustering sample groupings, sample topologies according to multivariate analyses and correlations of the “real world” environmental variables with sequence derived ecological distances were consistent throughout all virtual and also the real dataset. Next to that, diversity indices followed the same trends among different samples for all datasets. In the taxonomy-based approach the V5 virtual dataset performed slightly better than the V3 and close to the full-length 16S rDNA values in terms of classification depth. Class level sequence classification showed over-representation of taxa related to more frequently studied environments for both the V3 and V5 datasets with V3 being more affected by the study frequency factor. Overall, the combination of Illumina sequencing and the V5 16S rDNA demonstrated robustness against the frequently selected V3, in very short read based microbial ecology studies of environments as complex as agricultural or natural soils.