RDP: confirming V3-V4 region

[bio_informatics]

Hi there,
I'm working on RDP training set for genera of our interests. For that I download sequences from link .
Next, I want to confirm if these sequences really have V3V4 in them, as data I receive are for V3-V4 region. I check this paper and go for E. coli used in paper: J01695.2 .Next I get 16s region from it using online rRNA region. Further I get V3V4 region from this paper.

Until now data, sequences are all in hand. I BLAST them against this V3V4 region I get no Hit.
When I BLAST this v3v4 region on NCBI database I get hits from E. coli only.

How can I verify if the sequences I've have V3V4 in them? My understanding is/was that E. coli region is available in other genera as well. Or that is how the paper uses it.

I'm unable to go ahead with BLAST approach. BLAST would have provided me quantitative evidence as to which sequences are good enough to go ahead or otherwise.

Has anyone come across similar situation? How did you solve it?
Looking for some pointers.

[thermophile]

What are you trying to do? Create a custom database? For what purpose? It's hard to answer your question without knowing why you are trying to do something.

[bio_informatics]

What are you trying to do? Create a custom database? For what purpose? It's hard to answer your question without knowing why you are trying to do something.

Hi,

Thanks for your reply.
I'm trying to confirm if the downloaded sequences I've contain V3-V4 region in them of 16S rRNA gene.
Create custom database to verify if my V3-V4 region is available in the sequences. If not I'm going to get rid of those sequences in custom RDP training set.

I got over this: I used dc-megablast and set perc_identity to 20

I've follow up question on BLAST results:-

1) I want to limit number of hits against subject. As in Seq1 in query to have 4 hits in database and after that no more.

Based upon my need I'm going for:

blastn -query barcode.fasta -db complete_barcodes.txt -task dc-megablast -out full_blast_p20_max_target_seqs_4.txt -outfmt "6" -perc_identity 20 -max_target_seqs 4

I've been through:

https://www.biostars.org/p/111491/

https://www.biostars.org/p/131329/

Thank you again for your reply.

[thermophile]

I wouldn't use blast for that. Align the sequences to a reference alignment (I like SILVA), then trim based on whatever primers you use to define v3v4

Or if you really want to use blast, convert your downloaded RDP sequences into a blast db and query your v3v4 sequences against that database.

[bio_informatics]

Thank you for your reply. I am unable to below work flow:

I wouldn't use blast for that. Align the sequences to a reference alignment (I like SILVA), then trim based on whatever primers you use to define v3v4

1- Would I get any quantitative evidence?

Or if you really want to use blast, convert your downloaded RDP sequences into a blast db and query your v3v4 sequences against that database.

Yes, this is exactly what I'm doing. BLAST v3-v4 region against my sequences.

[Brian Bushnell]

I got over this: I used dc-megablast and set perc_identity to 20

Um, what? I've never used dc-megablast, but at that setting you would expect everything to match everything, no?

[bio_informatics]

Um, what? I've never used dc-megablast, but at that setting you would expect everything to match everything, no?

Hi brian,
Yes, I am looking for whatever matches.