Metal ions got stripped after MM sampling

[nancyzmn]

I'm trying to optimize a Zn(2+) ion coordinating with four residues (2 CYM, 1 HID and 1 ligand). After the initial MM sampling, the xyz files generated from the MM trajectory (stored in 2_sampling/1_cycle_1/train) no longer have the Zn(2+) ion.

[ethan-r-curtis]

The problem is that your prmtop file specifies the Zn atom as having an atomic number of -1. ff-opt determines atomic symbols for the xyz files from the prmtop atomic numbers. By default, atoms with an atomic number of -1 are treated as dummy atoms (like off-center charges for 4-point water models) and are omitted from the xyz.

I'm going to leave this issue open while considering what the desired user experience should be.