[gmx-users] Re: HB lifetime in g_hbond

[Patricia Soto Becerra]

>> I have a question regarding the calculation of hydrogen bond lifetimes
>> in g_hbond:
>>
>> Line 1094 of gmx_hbond.c v 1.16.2.11 2006/04/02 (which is equivalent to
>> line 1121 of gmx_hbond.c v 1.28 2006/08/30) reads:
>>
>> t  = hb->time[i] - hb->time[0] - 0.5*dt;
>>
>> leading to the time series 1, 3, 5,.. that does not correspond to my
>> sampling frequency of 2ps. This affects the calculation of x1 in the
>> immediate next line. My question is: Why is t defined in this way?
>>
> The program computes an histogram and centers the points in between time
> points. Note that this flag gives you useless lifetimes, see J. Phys.
> Chem. B 110 pp. 4393-4398 (2006). You want to use the lifetime that
> comes from the -ac option (look at line with Forward)
>

Yes, I had indeed noticed that the value reported by this flag was far 
from my own calculations so I wanted to make sure I was understanding what 
g_hbond was doing.  Thanks. (of course, I already read your paper in JPC 
B.)

Patricia