Hello! I am trying to compare my sequences with SRA data, which involves analyzing sequencing runs with different hypervariable regions. Do you know of any literature or examples where people analyzed different hypervariable regions? I have been trouble finding anything aside from huge meta-analyses on human microbiomes which is not really what I am going for. Even suggestions for search terms would help since I'm not really sure what to call this type of analysis and most papers don't advertise their methods in the title. Thanks so much, Qiime2 community! 
1 Like
Hi @JadeS,
Literature wise, you might look at some of Cathy Lozupone’s early meta analysis work.
I think the best general advice is to denoise and then use closed reference OTU clustering against the database of choice. This will make your annotation across hypervariable regions somewhat comparable and in my experience gives better results than ASVs with a fragment insertion tree. You may want to work collapsed to a lower level, because there are occasionally still mapping isssues, and of course, primer has a technical effect (see Figure 1 in this paper). I think it’s still an issue in the field!
Best,
Justine
3 Likes
Thanks so much for your response! Those are great resources to check out. Yes, I have been thinking of OTU clustering rather than using ASVs. Since my hypervariable regions overlap (V4 and V4-V5) I am wondering whether trimming and merging could be better than using a reference to cluster? Do you have any thoughts on that?
Hi @JadeS,
I haven’t played with that as much. I think the challenge with trimming and merging may come in read lengths and primers. Theoretically you should be able to just trim. Practically I think you may need to spend some time looking at the data. I would definitely try trimming both to the same length with your forward primer. And then, how you handle merging and trimming, Im not sure.
Best,
Justine