How Do I Translate Different Regions Of My Sequence In Different Frames?

Question

[Gimly_Gloin]

What I want to do is say something like in sequence D translate region 1-100 in frame 1 from 102-150 in frame 3 from 160-180 frame -3

I would typically extract this information from a blastx table and (the idea I had) feed it into EMBOSS's transeq... However transeq seems to mess up which frame does what, and even in the help it says (for the -region function) :

"Note: you should not try to use this option with any other frame than the default, -frame=1"

I wonder if there is any software that can do this? I've looked at bioperl, and haven't seen anything that caught my eye .... ? thoughts?

Answer 1

[Neilfws]

I think you can use the Bioperl methods subseq, translate and revcom to achieve what you want. A brief example using this test sequence, saved as nirs.fa:

#!/usr/bin/perl -w 
use strict;
use Bio::SeqIO;

my $file  = "nirs.fa";
my $seqio = Bio::SeqIO->new(-file => "$file", -format => "fasta");
my $seqi  = $seqio->next_seq;

my $s1 = $seqi->subseq(1,100);
my $s2 = $seqi->subseq(102,150);

# apologies for chaining together these methods...
# frame 1: in Bioperl frames 1,2,3 = 0,1,2
print Bio::Seq->new(-seq => $s1)->translate(-frame => 0)->seq, "\n";
# MKAKNWLYPAIAALPLSLWLGLPHAATKAEPKA
# frame -3: we reverse complement, then frame -3 = Bioperl frame 2
print Bio::Seq->new(-seq => $s2)->revcom->translate(-frame => 2)->seq, "\n";
# RGRTQRQ*GWGWRLS

I too am somewhat confused by EMBOSS transeq (but I think it is doing what you want). Here are the equivalent commands, with results:

transeq -sequence nirs.fa -frame 1 -regions 1:100 -stdout -auto
# >2576786-2578450_1 Ralstonia eutropha H16 chromosome 2, complete sequence
# MKAKNWLYPAIAALPLSLWLGLPHAATKAEPKAX

transeq -sequence nirs.fa -frame -3 -regions 102:150 -stdout -auto             
# >2576786-2578450_6 Ralstonia eutropha H16 chromosome 2, complete sequence
# RGRTQRQ*GWGWRLSV

So transeq seems to be returning one residue more. I'm sure there's a good explanation if we think about it. Maybe the Bioperl method assumes that the last 3 bases are a terminator unless told otherwise.

[Gimly_Gloin]

I get the correct frames from parsing my blastx output by printing out $hsp->hitstring or $hsp->querystring.thats not the issue what I would like is to adjust that information from the blast output so that instead of getting 5 or 6 fragments of the protein I get a single string that is correct for the protein in the several frames blastx has detected. I thought I could do this with transec, but for any given information in the blast table it shows a different result to the transec output (I've also corrected the GFF frame() used in SearchIO to show the blastX frame).Is blast skipping bases?

[Gimly_Gloin]

Also it doesn't allow multiple commands

[Gimly_Gloin]

I think your probably right about transeq though... too much "intuition"

[Neilfws]

I would not use the BLAST output for this task. Since BLAST tells you the frames, why not just go back to the original input sequence? As your question says: "How do I translate different regions of my sequence in different frames."