31 March 98
by N.Guex
Tutorial :
superimposition
In this example, we will compare the hemoglobin of a marine blood worm (pdb entry 1hbg) with the leghemoglobin of the yellow lupine root nodules (pdb entry 1gdi).
- First of all, open the pdb file 1HBG (it is included
in the tutorial package).
- Hide all the sidechains by option-clicking (right
mouse button for PC users) on the sidechain column of the
control panel, and colour the whole molecule in white by
option-clicking in one of the colour boxes of the control
panel . Then, scroll at the bottom of the control panel
and colour the heme in yellow.
- Repeat the same procedure with the file 1gdi, but this time, colour the protein in red, and the heme in violet. Now hit the "=" key of the numerical keyboard to automatically recenter the view (right mouse button for PC users).
- Now select the "Magic fit" option of the "Tool" menu and accept all default parameters. The two proteins should now be superimposed. Here again, recenter and expand the view by hitting the "=" key (right mouse button for PC users).
- Now bring the align window to front, and move the
cursor onto the Trp15 of 1gdi. As you can see, it will
blink from blue to yellow in the display window, allowing
you to immediately know where it lays in the protein. In
addition, if you look at the bottom left of the align
window, you will see a RMS distance. This is the rms
deviation between backbone atoms of 1gdi:Trp15 and
1hbg:Lys15.
The value is high (4.108), because the alignment you see in the window is just the raw sequence of each protein The Trp15 does not correspond to the Lys15. What you will have to do now is generate the structural alignment.
- Click on the name 1gdi on the left of the align
window to bring this protein to front (the control Panel
should reflect the change). Now use the option "Generate
Structural Alignment" of the menu "Tools", and go back
onto the Trp15 of 1GDI. It should now be aligned with the
Ala18 of 1hbg, and its rmsd should be much lower (2.745).
- While you're at this stage, choose the option "Colour
by RMS" from the "Tool Menu" and look at the alignment
window. Amino acids will appear in dark blue if they have
a close correspondent in the reference structure, and
appear brighter (from blue to yellow) when their rmsd
increases, up to be red if they have no correspondent in
the reference structure.
The result is that loops inexistent in the reference are clearly visible in red in the main window (as you can see in the small icon at the top of the page).
- Click on the little text icon of the Align window. A
text window containing the structural alignment that will
look like the one below, depending on the options you
have set in the alignment preferences should appear.
1HBG 1 GLSAAQRQV IAATWKDIAG ADNGAGVGKK CLIKFLSAHP
1GDI 1 GALTESQAAL VKSSWEEFN- -ANIPKHTHR FFILVLEIAP
.*. .* . . ..* .. * . . .* * *
1HBG 40 QMAAVFG--F SGASDP---- GVAALGAKVL AQIGV-AVSH
1GDI 39 AAKDLFSFLK GTSEVPQNNP ELQAHAGKVF KLVYEAAIQL
.*. . . * . * ..**. . *.
1HBG 73 LGDEGKMVAQ MKAVGVRHKG Y-GNKHIKAQ YFEPLGASLL
1GDI 79 EVTG--VVVT D-ATLKNLGS VHVSKGVADA HFPVVKEAIL
.* * . .* . * . ..*
1HBG 112 SAMEHRIGGK MNAAAKDAWA AAYADISGAL ISGLQS-
1GDI 116 KTIKEVVGAK WSEELNSAWT IAYDELAIVI KKEMDDAA
.. .*.* . . **. ** ... . ..
- Now Select the "Save Alignment item of the File menu.
This will save your preview as a text file. Note that it
is not necessary to look at the preview, you can also
directly save the structural alignment file.
- The Magic Fit does a crude superimpostion. You can
achieve a much better one by using the "iterative Magic
fit" option of the Tool menu. Try it!
- Other means to achieve a better superimposition are
to use the "Select AA matching Ref. structure" item of
the "Select" menu, providing your RMS deviation cutoff
value (for example 2.5) and then using the "Fit Molecule
Auto" item of the "Tool" menu. Alternatively, you can
refine the superimposition by manually selecting parts
(for example Helixes) and then using the "Fit Molecule
Auto".