How to delete 16s ribosomal sequence from a genbank or fasta file?

Question

[benzhang]

Hello, all! I want to delete 16s sequence from bacteria genome in either gbk or fasta format. I dob't have any access to commercial software. Are there any open source programs or command line tools that can achieve this? Thanks!

Best, Ben

[GenoMax]

You could open your fasta file and locate the 16S sequence start (by find) and then delete the range you need with any text editor on *nix/OS X. If you are working on windows then try ApE or SnapGene Viewer.

Answer 1

[Brian Bushnell]

Using the (free) BBMap package:

If you know the 16S sequences (or those of a close relative), you can align them to the genome to produce a sam file. For example:

mapPacBio.sh ref=genome.fasta in=16S.fasta out=mapped.sam ambig=all maxindel=20

Then you can run BBMask:

bbmask.sh in=genome.fasta out=masked.fasta sam=mapped.sam masklowentropy=false maskrepeats=false

This will mask the sequences covered by the mapped 16S sequences. You can alternatively do it using BBDuk's kmer-based masking mode:

bbduk.sh in=genome.fasta out=masked.fasta ref=16S.fasta k=200 hdist=1 kmask=N

...but I'd suggest the alignment method unless you encounter problems with it.

[GenoMax]

And then delete the masked region?

This answer has left me wondering if @benzhang wanted to delete as in remove or mask the sequence so it is no longer considered. Guess we will find out.

[benzhang]

I'm trying to map bacterial genomes to human microbiome data without considering 16s sequence. Would removing and masking achieve the same result? Thanks!

[GenoMax]

In a way yes. But masking would preserve the overall coordinates so that may be the better option to use.

[benzhang]

I see, thank you so much!

[benzhang]

Thanks, Brian! Would it work if I find the 16s sequence from the gbk file, and align them to the fasta file, and then mask the aligned region?

[Brian Bushnell]

Yes, that would work. But I'm not familiar with gbk format. The sequence would need to be transformed to fasta first, which is fairly straightforward.

[benzhang]

Got it, Thanks, Brian!