Hi @dana. I believe this is because we remove “-”, “N” and “?” characters from the insertion before reporting it. So in this case it looks like the 739bp insertion is largely due to missing data.
I don’t think we have an easy way for augur align to produce an alignment which does not remove insertions, but you could try running the following and then examining the alignment file itself to see exactly what the insertion in strain “18925” is
You might also be interested in nextalign which does reference alignments (and translations). We are using this for large SARS-CoV-2 alignments. It reports insertions in a similar way. See here for details: