I want to model ONLY magnesium ion (from protein B) on an already
crystallized structure (protein A). So the sequence of my target and
template is exactly the same. My alignment file is
Can you send me (not the list) your structure files too, please? I can't
reproduce your problem (and thus fix the bug) without them. It is
definitely a Modeller bug though (not something wrong with your inputs):
Modeller should never crash with a segfault, regardless of what you feed it.
I think the surface area calculation is getting confused since you have
no amino acid resides in one structure (but I'll need to check with your
inputs). As a workaround, you can use the BLK residue '.' rather than
'$'. Since MG is just a point anyway, your modeling results should be
almost identical.