Could be a bias due to the primers used or to the database employed for 16S rRNA reads analysis. Have a look on this paper Article Evaluating Established Methods for Rumen 16S rRNA Amplicon S...

José Francisco Cobo Díaz Thank you so much for your quick response, it's incredibly appreciated! This was an extremely useful paper for us, as it validated our use of V4 primers and the SILVA database, which we were still unsure about. Since we're still not seeing Enterococcus with the optimal primers/database for it, could we be having an extraction issue instead?

Two possibilities excluding PCR issues: 1. The amplicon sequence depth and the orders of magnitude that Enterococcus is found over the total bacterial density. To exemplify this if you have Enterococcus in concentrations of 10 to the power of 3 cells per ml of reconstituted swab sample (thousand viable cells per ml), it could be readily cultured and isolated. Also, while very low, DNA extracted from this kind of bacterial cell concentration could have enough target copies for a positive result. But if you have in your swab sample were 10 to the power of 9 are other bacterial cells which is a feasible number (i.e. hundred times less than the bacterial cells per gram of fecal material) and in those 1 billion cells still containing thousand Enterococcus cells, even if you have 100.000 reads per V4 amplicon, you may not detect even once the Enterococcus if they are in such ratio of 1:1000000. Of course, this is an extreme example, but even if you have closer ratios of 1 Enterococcus cell vs 10.000 of total bacterial cells in the swab, still you can encounter the non detection in amplicons of 30.000 reads but stochastic chance in amplification, library or sequencing production for low abundance bacterial types. 2. The extraction of Enterococcus DNA is somehow more difficult than for other microbiome bacteria, but this issue would not be the problem if you have PCR positive results of Enterococcus presence. I suggest you to look at this relationships between estimated total bacterial amounts vs Enterococcus amounts per surface/swab or reconstituted volume as this may explain why even sequencing. I hope this helps!

I agree with Howard Junca , maybe the concentration of Enterococcus are so low to be detected by universal 16SrRNA primers.

Thank you very much Howard Junca and José Francisco Cobo Díaz for your insight, we really appreciate it! The reason we're questioning our culture-positive swabs is because our Illumina data showed a higher relative abundance of Enterococcus on some culture-negative swabs (even accounting for per-sample sequence counts)... granted this was just a pilot so sample numbers are low, but I guess we're at a crossroads of whether to trust culture or NGS more. We unfortunately do not have PCR data for all swabs/enough cognate swabs for it.