Hi 

you can use :

http://www.ccp14.ac.uk/solution/bond_length/

and 

http://webapps.ccpn.ac.uk/acpype/

Thanks for your answer. I am looking for a solution without using additional programs.

I would make an index file by selecting the desired atoms and then using gmx distance. i believe you are doing the same. So, yes, this is the most probable approach.

Thanks.

 And particularly for distributions, use your .xvg file in gmx analyze and calculate distributions.

What is the Unit of distribution

Actually this is an issue with the approach.

As you know that gmx distance calculates the distance between two entities. Now, the way these entities are provided is not totally correct.

You are interested in calculating the distance between, C and O.

And you have provided the index for [ Catoms ] and [ Oatoms ]. This would work if you would have single methanol molecule.

Even if you provide single index and do "ctr+d" it will calculate the distance (if you have even numbers).

Answer lies here;

mol 1 : C1 O2

mol 2 : C3 O4

.

.

mol n : C2n-1 O2n

your index file for [Catoms ] is

1 3 5 7 9 .. 2n-1

If you give this index for gmx distance, it will calculate, distance between 1-3, 5-7 ...[2n-3 - 2n-1], which is incorrect and you don't want this.

Therefore you have to make another index entry [ bond_CO ] and write

1 2 3 4 5 6 7 8...

This will give you the distance between

1-2, 3-4, 5-6, ...[2n-1- 2n]. This is the desired distances that you want.

Now if you have two atoms in the molecule then making the index of entire molecule will have same order, hence, it will work fine. But for three and higher atoms in a molecules, you have to write index.

If you make the index file, ensure that index of correct atoms are written in the index.

Since, I find making a correct index complex, I wrote a code to write index of individual atoms in the above mentioned manner. This clear out all the doubts and correct distances are computed.

Same goes with angle calculations.

Hope it helps.