A common problem is that of base overcalls that will result in perhaps one reading having and extra base, and all the others being padded by the alignment routines:
Read1 ACC*AG Read2 ACC*AG Read3 ACC*AG Read4 ACCCAG Read5 ACC*AG Consensus ACC*AG
In this first case we see that Read4 has an extra C, probably
due to an overcall. Check that the trace for Read4 shows an overcall.
It is a good idea to check good quality traces for both strands as well as the
trace with the apparent problem. Also note that enabling "Show reading
quality" (Settings menu) will show the reading quality as grey scales.
We now need to remove the column. It would appear that this could be done by
removing the * from each of Readings 1, 2, 3 and 5, and removing the
C from Reading 4. However this will only make edits to those five
readings. As we're trying to remove an entire column from the contig, we need
to shift to the left by a single base the position of any readings to the
right. Naturally this is not the ideal method.
By placing the editing cursor in the consensus (on the second A) we can press Delete to remove the entire column. This automatically makes sure that everything is consistent. If we are editing at 100% consensus cutoff then this consensus base will be a '-' instead of a '*'. For this to work we need to make sure that we have "Allow del dash in cons" enabled in the Edit Modes menu. See section Editing Modes.