Unclassified 16SrRNA using green gene taxonomy

[Uzobiogene]

Hello Fellow NGS users,
I recently sequenced my micro biome (Five body sites) using the Illumina 16SrRNA V4 metagenomics platform. The paired reads were analyzed on base space with sample information showing 100% quality filtering. From the classification statistics, one of the samples had over 50% as unclassified at Kingdom level, phylum, class, order, family, genus and species levels. What could be responsible for this and is there a way of extracting the sequence reads that were reported as unclassified.

[henry.wood]

Have you looked at the distribution of read lengths in the badly behaved sample. I was having this issue a while back and found that many more reads in poor samples were the wrong length. If I'm expecting a 240bp PCR product, then getting rid of any read which is more than 5bp away from 240bp massively reduces the number of unclassified reads. I suspect that it is some kind of PCR artefact, caused by a poor initial DNA sample.
You could play around with read lengths, seeing what kind of filtering works best for you.

[thermophile]

The basespace applications are just a very basic first pass. I'd suggest your run your sequences through mothur (my preferred software for amplicons) or qiime to really see what's going on.