In the score-shaded multiple alignment frame you can edit the position of gaps in the alignment (profile) if you have selected Profile Edit mode (Edit menu). This method should be used with caution and only for good reason since it may cause deviation from a theoretically optimal alignment. Nevertheless, it will frequently be the case that you will have specific knowledge about the functional domains of the proteins being aligned that will permit you to make better choices for certain gap positions. It may also be desirable to modify the profile to match a published alignment, but beware that published alignments frequently have poorly justified segments that were suboptimally computer-aligned or unwisely modified by hand. This is practice I urge to you to undertake cautiously, and with this warning I hereby wash my hands of responsility.
To move gaps in an alignment, use the mouse to highlight one or more contiguous regions of the alignment. Use click then shift-click to select a block (currently, only contiguous blocks can be selected). To move a block of gaps left in the profile, type "Shift-<" (less-than sign). To move a block of gaps right in the profile, type "Shift->". The block may contain both gap characters and amino acid residues - only the gaps will be moved of course. If you attempt to move the selection past the end of the alignment, nothing will happen. You may move gaps in all the sequences at once or any contiguous subset of them by making the appropriate selection. To insert a column of all gaps just left of the selected column, type "Cntrl-Shift-G". To remove a column that is all gaps, select the column and type "Cntrl-Shift-X". Using a combination of these, you can make your profile match any previous alignment (in principle - it may be tedious). When you move, delete, or insert one or more gaps, the displayed score quality will be recomputed, so you get immediate visual feedback about the score quality for the change. The overall alignment score isn't recomputed (should be fixed). Try moving a block of gaps back and forth to get a sense of this, or display two identical alignment frames and manipulate one to compare with the other.
Profile editing is especially useful if you want to construct a gold standard profile to be used in determining the optimal alignment of a new sequence to your gold standard. If you attempt to do this, it is wise to take considerable care in constructing your gold standard, selecting sequences for it with care and making gap changes with circumspection. If you construct such a profile and provide a paragraph or two describing its rationale and why it is useful, please send it to me as an email attachment. After review, I may choose to include it in the librarary of gold standard profiles that I am currently building for release with Bonsai.
1) within reason, include as many sequences as possible. A good profile will have 10 or more sequences. More than 50 is likely to be superfluous but I am just guessing here.
2) edit the included sequences to spread out their relatedness, particularly avoiding 2 or more very similar sequences and avoiding sequences that are unusually divergent compared to the others. Bonsai always weights the sequences according to their relatedness, but it is most effective to do some of this effort up front in the form of sound sequence choices.
3) if any of the included sequences are purely predicted (not experimentally verified by cDNA sequence or some other method) be especially cautious about including them. If a purely predicted sequence has an unusual feature (a large insertion etc) always exclude it, since experience shows that this is usually a prediction error.
4) be cautious in moving gaps around after multiple alignment. A detailed knowledge of the functional constraints on specific residues in your chosen protein family should generally be your basis for moving gaps to improve alignment fidelity to structure. Do not move gaps just to make the profile "pretty". Bonsai places gaps where they are for a reason (though it may be a small score difference and hard to discern).
5) when moving gaps, keep a sharp eye on the sequence alignment quality - even if you think you are moving a gap for good reasons, if the local score quality dramatically degrades think twice!
6) if you plan to let others use your gold standard profile, annotate it carefully. At a minimum, provide an overview of the family that the profile represents, how you made it, and why it is useful. Include literature references if available.