I was wondering if I have multiple 16S regions--some sequences do not overlap--sequenced in the same fastq file. Do I have to analyze each region separately or can I run all the regions at once?
Regions:
27F and 518R (V1,V2,V3)
515F and 907R (V4,V5)
895F and 1100R (V6)
1237 and 1492R (V8,V9)
I did try to run it all at once but when I taxonomically classified the samples a majority (40-60%) of the features were classified only to Domain or were unclassified.
I tried both the Naive-Bayes and BLAST consensus classifiers using the SILVA database as the reference.
Hi @ChaseR,
You'll need to analyze those separately for most QIIME 2 analyses to make sense. The issue is that you can have V8/V9 reads from the same organism as a V4/V5 read, but (outside of taxonomy assignment, or closed-reference OTU picking which isn't yet supported in QIIME 2) QIIME wouldn't know how to relate those sequences to one another. Most of the analyses that you'll run in QIIME are not dependent on taxonomy (e.g., alpha and beta diversity), so for those analyses a V8/V9 read from the same organism as a V4/V5 read would be treated as observations of different organisms.
We plan to have closed-reference OTU picking supported in the 2017.8 release (due out in late August), so you can alternatively wait for that to process these data together.
I figured this is what I needed to do but worth confirming. I think I will analyze just the V4/V5 now and rerun it when the 2017.8 release comes out to compare results.