I find mistakes in the output file of gmx hbond -hbn when I run a simulation where I want to know the hydrogen bonds between 2 kinds of moleculars, H2O and glycerol. The command I use to analyze the hydrogen bonds is: _echo 0 1 | gmx hbond -f frame10.gro -s md_1.tpr -n traj0.ndx -a 20 -r 0.31 -hbn wg0.ndx_ Here is part of the index content of [ hbonds_water_glycerol ] in wg0.ndx: - 44533 44534 25652 - 44533 **44534** 29978 By comparing the indexes and the ones in the frame10.gro file, it is easy to know that the order of the indexes above corresponds the order of donor-Hydrogen atom-acceptor. Firstly, although it may occurs that the same donor-Hydrogen atom forms 2 different hydrogen bonds with 2 acceptors (sometimes 3 in the file) due to the geometric criterion, I find too many of them. Secondly by calculating the angles of Hydrogen atom-donor-acceptor which demand <20°, I find the second line above should be: 44533 **44535** 29978(see the figure attached). I use gromacs2018.4 to run the md and analyze the H-bonds, and similar issues can be found using gromacs5.0.7. The relevant files and the figure showing the angles I calculate manually are attached. [Angles](/uploads/cfe0b640ba2794b2095820daba4af90d/Angles.jpg) [frame10.gro](/uploads/5a856904c11d31ee626f6b42ac8633a4/frame10.gro) [md_1.tpr](/uploads/9b7527beaacc99838d6a5376fb189de4/md_1.tpr) [traj0.ndx](/uploads/413d16729cb7d25344cc2a5658830e16/traj0.ndx) [wg0.ndx](/uploads/b5bbe6292ca3011374ca7d87da5e427f/wg0.ndx)