Building of the multiple alignments
We will make multiple alignments with the help of different programs.
Here're our identifiers:
TRPB_ECOLI, TRPB_BACSU, TRPB_MYCTU, TRPB_SALTY, TRPB_CHLTR
Firstly we will make a multiple alignment with muscle program.
The second way is the work with UniProt site.We search sequences by their identifiers.
Then we choose them by pinning tick and press Align.
Finally we press Download, choose Uncompressed and FASTA format.
And the last but not the least way of alignment is Jalview: Web Service - Alignment - Tcoffee with Defaults.
So, we have gotten our multiple alignments using three ways. Then we will import them to the Jalview to make three groups.
After importing our alignments we formed groups and aligned them.
You can see groups in Jalview here
The first difference
Pay your attention to the second column
Similarities: GG*S*
Differences: Asn70 in fasta and uniprot versus Gly68 in muscle,
Arg70 in fasta and uniprot versus Gly68 in muscle.
Speaking about this column, fasta and uniprot-alignments are more similar.
One very local example: [TCoffe comparing to the muscle]
Asn70 aligned not to 96 residue, but to the 98 (another numbers).
| fasta (TCoffee) |
muscle |
uniprot |
G(72)G(73)A(74)K(75)
D(66)G(67)P(68)R(69)
-(..)N(70)T(71)T(72)
G(95)S(96)A(97)R(98)
-(..)R(70)T(71)T(72)
|
G(72)G(73)A(74)K(75)
D(66)G(67)P(68)R(69)
A(67)G(68)T(69)N(70)
G(95)S(96)A(97)R(98)
A(67)G(68)T(69)R(70)
|
G(72)G(73)A(74)K(75)
D(66)G(67)P(68)R(69)
T(69)N(70)T(71)T(72)
G(95)S(96)A(97)R(98)
T(69)R(70)T(71)T(72)
|
|
|
The second difference
Look at the first column, we can see the bias of Arg386 in muscle.
It aligned to another letters, (because of gaps); you can see it from the picture.
The first column in fasta (TCoffee) and uniprot are similar.
One very localexample: [TCoffe comparing to the muscle]
Arg386 aligned not to 390 residue but to the 392 (another numbers)
| fasta (TCoffee) |
muscle |
uniprot |
V(391)L(392)E(393)E(394)
R(386)-(...)-(...)N(387)
I(390)L(391)K(392)A(393)
W(415)F(416)G(417)L(418)
I(390)L(391)K(392)A(393)
|
V(391)L(392)E(393)E(394)
-(...)-(...)-(...)-(...)
I(390)L(391)-(...)-(...)
W(415)F(416)-(...)-(...)
I(390)L(391)-(...)-(...)
|
V(391)L(392)E(393)E(394)
R(386)N(387)R(388)G(389)
I(390)L(391)K(392)A(393)
W(415)F(416)G(417)L(418)
I(390)L(391)K(392)A(393)
|
|
|
The third difference
If we look at the third column, we can note that Ala67
moved by muscle to another position in alignment
(it aligned to residues with another coordinates).
Fasta and uniprot-alignments are similar
One very local example: [TCoffe comparing to the muscle]
Ala67 aligned not to 92 residue but to 95; another numbers
| fasta (TCoffee) |
muscle |
uniprot |
V(67)T(68)E(69)Y(70)
F(61)A(62)R(63)A(64)
I(65)T(66)A(67)G(68)
L(90)S(91)Q(92)H(93)
I(65)T(66)A(67)G(68)
|
V(67)T(68)E(69)Y(70)
F(61)A(62)R(63)A(64)
I(65)T(66)-(..)-(..)
L(90)S(91)Q(92)H(93)
I(65)T(66)-(..)-(..)
|
V(67)T(68)E(69)Y(70)
F(61)A(62)R(63)A(64)
I(65)T(66)A(67)G(68)
L(90)S(91)Q(92)H(93)
I(65)T(66)A(67)G(68)
|
|
|
Comment:
The beginning of alignment which was built with muscle is shifted to two residues
to the left of the other two (in the befinning). The shift "disappears" by gaps between 75 (Lys) and 76 (Ile)
residues of the first sequence of the second alignment. After that we can note "cluster of conformity" in all three alignments.
I suggest that the alignment which was gotten with muscle is not the best
(speaking about shifting in the beginning, about gaps in the middle of alignment, which are not being noted in others).
I think, that the best variant of multiple alignment is UniProt's (we work with alignment on the site)
(we don't see wide indels in the beginning as we see in TCoffe with Defaults based on Jalview; but this variant is very good too).
So, the best variants are: TCoffee (with Defaults on Jalview interface) and UniProt
TCoffee has an advantage - it's more general, not tied to a particular database