[gmx-users] pair-distance distribution function

[Jon Ellis]

Daniela, thank you for your help.

I tried using g_rdf to calculate the vector lengths of all atom-pair 
distances in a protein. I dumped a single frame from my trajectory using 
trjconv, then ran
    g_rdf -f protein.gro -o rdf -n protein.ndx -bin 0.001
This gave me what appeared to be a pair distance distribution, except 
that it seemed to truncate the list at r=4.085 nm, and the P(r) value at 
this r was non-zero. I know from other analyses that the protein should 
have longer distances than 4 nm. I tried different bin sizes, and 
different endq and startq values, but the result remained the same.

Is there something I'm missing, or can I not do pair-distance 
distribution functions using g_rdf?

Cheers
Jon

>hi jon,
>
>g_rdf counts the radial distribution function of an index group around 
>another index group.
>
>daniela
>
>
>Jon wrote:
>>/ Hello
>/>/ 
>/>/ Is there a program in Gromacs, or a standalone program, to calculate the 
>/>/  pair-distance distribution function P(r), aka the vector length 
>/>/ distribution, for a protein?
>/>/ 
>/>/ The algorithm is P(r)=Dn(r)/nDr, where Dn(r) is the number of atom pairs 
>/>/ whose separation is between r-Dr/2 and r+Dr/2, and n is the total number 
>/>/ of atoms in the system.
>/>/ 
>/>/ I would appreciate any help.
>/>/ 
>/>/ Cheers
>/>/ Jon
>/>/ /
>