[gmx-users] pdb2gmx yielding (a couple of) -nan values for coordinates of built hydrogen atoms

[Christopher Neale]

In most cases, I get what I expect when I execute the following command:

echo -e "6\n1\n" | pdb2gmx -f ${b}e.gro -o ${b}et.gro -p tmp.top -i ${b}_posre.itp -ter -ignh

However, I have one input .gro file for which the output .gro file has -nan for all coordinates for HB1 and HB2 atoms. I built the input .gro file using molefacture in VMD, and maybe it is not perfect, but I was under the impression that pdb2gmx didn't care about coordinates, so I am stumped about what could be the problem.

Got same result with 4.6.1 and 4.6.5. Also a second run through pdbtgmx (giving the output .gro from first round as input .gro to second round) gave the same results.

Has anybody else run into this? If not, I'll file a redmine.

Thank you,
Chris.