ADL: protein-protein docking

Yang Ye leafyoung at
Wed Oct 24 23:10:30 PDT 2007

Hi, Zheng Cai

What you did is alright.

The useful information to me is that you have big ligand. Since you only opt for rigid-body docking, you are required to edit ligand.pdbqt to include all atoms into ROOT/ENDROOT. Don't leave any BRANCH/ENDBRANCH section.

About torsdof, it is good to change in ligand.pdbqt. Generally, TORSDOF = NDIHE - bond linking to Hydrogen. So I reached my statement that torsdof could be ignored since ndihe line has been ignored. But, definitely, you shall be put a right number for it.

Yang Ye

----- Original Message ----
From: Zheng Cai <caizheng78 at>
To: autodock at; Yang Ye <leafyoung at>
Sent: Thursday, October 25, 2007 7:59:14 AM
Subject: Re: ADL: protein-protein docking

Hi Ye, Thank you so much for your quick reply.

I think the first problem will be solved following your instructions. But for the 2nd one, I checked the output file of my first trial docking, i found the following.

TORSDOF record detected: number of torsional degress of freedom has been set to 600.

*** No Rotatable Bonds detected in Small Molecule. ***

>>> No torsions detected, so skipping "nonbonds", "weedbonds" and "torNorVec" <<<

Number of atoms in ligand:  1130

Number of vibrational degrees of freedom of ligand:  3384

Number of torsional degrees of freedom = 600
Estimated loss of torsional free energy upon binding = +164.6400 kcal/mol

and this energy was used in the next steps even for clustering, and i found also


DPF> ndihe 0                              # number of active torsions
./autodock4: WARNING!  The "ndihe" command is no longer supported.  The number of torsions in the PDBQT file(s) is the number that will be used (
i.e. 0)

so do you really think this torsdof  line will be ignored by AD4?  or only the ndihe line will be ignored by AD4? 

I am confused, because both "No torsions detected" and "Number of torsional degrees of freedom = 600", could you tell me what it means? Does this enerygy really matter?

For my case, TORSDOF 600 was defined in the last line of the 
ligand.pdbqt file. I checked the AD4 manual, and find that the only step I skiped for ligand preparation is the second step: Ligand->Tortion Tree->Detect Root, i skipped this step because I found it was extremely slow for this big Ligand. in the 3rd step, i simply chose "Make all active bonds non-rotable". Then in the 4th step, Ligand->torsion Tree-> set number of Tortions, I typed number 0 in, and hit enter. Then I Ligand->output->Save as PDBQT..

Do you think I can skip the  "Detect Root"?  or I have to wait until the "Detect Root" finished? Can I simply edit last line of the .pdbqt file to 



Thank you so much for your time! and look forward for your reply.



On 10/24/07, Yang Ye <leafyoung at> wrote:
1. You can work around this by having an understanding of the pdbqt file and manual editing. When atoms are placed in one block with ROOT/ENDROOT or BRANCH/ENDBRANCH, they become relative fixed to each other. First, generate two pdbqt files for each part of the ligand. Identify their "bonding" parts and paste two files together.

2. This line will be ignored by AD4. I am curious why there are 600 TORS for the ligand. Please check your ligand and procedure of preparation.

Yang Ye

----- Original Message ----

From: Zheng Cai <caizheng78 at>
To: autodock at
Sent: Thursday, October 25, 2007 6:07:00 AM
Subject: ADL: protein-protein docking

Hi autodockers!

I am trying to use autodock4 to do a antibody-antigen docking. The antigen
protein( 2 chains) was simply considered  as Macromolecular. And the
antibody Fab (also 2 chains) was considered as ligand. I met several

troubles as following when I run the program. I wonder if any of you had the
same problem before and could kindly help me out?

1. two chain ligand.  it seems autodock4 can only deal with 1-chain ligand,
if i load the two-chain coordinate in, it will simply output the first chain

it read, while the second chain was lost. (fortunately, there is no trouble
for a two-chain macromolecule)

2. for the torsdof parameter in the .dpf file, I saw the value below

torsdof  600 0.274000          # torsional degrees of freedom and


for my understanding, this parameter should be set as 0 in the
protein-protein rigid docking case, am i correct? if it is, can I manually
edit this line in the .dpf file to:

torsdof  0 
0.274000          # torsional degrees of freedom and coefficient

Any input will be greatly appreciated!

All my best regards!

Zheng Cai, Ph.D.
Department of Pathology and Laboratory Medicine

University of Pennsylvania School of Medicine
252 John Morgan Building
36th Street and Hamilton Walk
Philadelphia, PA 19104
(lab-tel) 215-898-2870
--- ADL: AutoDock List  --- ---
--- ADL: AutoDock List  --- ---

More information about the autodock mailing list