[EMBOSS] Transeq question, frame phases

[David Mathog]

> Now let's do that with the reverse complement strand:
> 
> $ transeq asis:TTTTTTTT -filter -frame 1
> >asis_1
> FFX
> $ transeq asis:TTTTTTTT -filter -frame 2
> >asis_2
> FFX
> $ transeq asis:TTTTTTTT -filter -frame 3
> >asis_3
> FF

That is the problem.  Let me try to explain more clearly what the issue is.


AAAAAAAA               Forward
TTTTTTTT                          Reverse
                         abc      cba  <--- codons in diagram
^a^^b^^c^    phase 1   1 KKX    4 XFF    EXPECTED
x^a^^b^^c^   phase 2   2 KKX    5 XFF    EXPECTED
xx^a^^b^     phase 3   3 KK     6  FF    EXPECTED


^a^^b^^c^    phase 1   1 KKX    4  FF    OBSERVED
x^a^^b^^c^   phase 2   2 KKX    5 FFX    OBSERVED
xx^a^^b^     phase 3   3 KK     6 FFX    OBSERVED

Assume an extra codon L to the left of a.
                         abc      baL  <--- codons in diagram
^a^^b^^c^    phase 1   1 KKX    4 FF     EXPLAINED?
^^a^^b^^c^   phase 2   2 KKX    5 FFX    EXPLAINED?
L^^a^^b^     phase 3   3 KK     6 FFX    EXPLAINED?


That is, if the meaning of the + phases is to define the three codons
a,b,c as shown in the diagram, such that the forward translation is as
shown, then the reverse translation should be as shown above in
expected.  That is, it is the translation of the exact same set of
codons done individually, but for the - strand reverse complement the
codon first, and then invert the resulting translated sequence.  That
way the X, where it occurs is attached to the same partial codon "c". 
What I think is happening in transeq is that it is starting with the
first full codon in the frame on the given strand.  In effect that
shifts the translated codons as shown in the "EXPLAINED?" section.

If partial codons were not translated then these would all be equivalent.

Regards,

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech